A University of Sussex DPhil thesis
Available online via Sussex Research Online:
http://sro.sussex.ac.uk/
This thesis is protected by copyright which belongs to the author.
This thesis cannot be reproduced or quoted extensively from without first
obtaining permission in writing from the Author
The content must not be changed in any way or sold commercially in any
format or medium without the formal permission of the Author
When referring to this work, full bibliographic details including the
author, title, awarding institution and date of the thesis must be given
Please visit Sussex Research Online for more information and further details
Psychophysiological Indices of
Recognition Memory
Becky Heaver
Thesis submitted for the degree of Doctor of Philosophy
University of Sussex, December 2011
i
I hereby declare that this thesis has not been and will not be submitted in whole
or in part to another University for the award of any other degree.
Signature: ………………………………………………
Date: ……………………………………………………
ii
Acknowledgements
I owe my deepest gratitude to my DPhil supervisor, Dr Sam Hutton, without
whom this thesis simply would not have been possible, and who has made
available his support in a number of ways, and encouraged, guided and
supported me from green masters student to completed doctoral candidate. I
would also like to thank him for, amongst other things, his hardcore binary reprogramming of a parallel cable, a number of interesting side projects, and never
giving me cause to attend a ‘How to Manage Your Supervisor’ workshop. I
would like to thank Francesca Citron, Kathrin Mikan and Steve Hamilton for their
tuition and assistance in ERP techniques. Special thanks go to Jacob Readman,
Emily Lewis and Rakshita Roplekar, University of Sussex, for helping with data
collection during Experiments 1, 3, 6 and 7, and to all my participants, many of
whom were friends who helped out of the goodness of their hearts, and some of
whom managed to stay awake for the duration of entire experiments. I also
thank Prof Alan Garnham for asking me questions that really made me think; Prof
Brennis Lucero-Wagoner who truly touched me by looking through ‘the detritus of
40 years of academic life’ just to find me a technical report in my first year when
she didn’t know me; Prof Stuart Steinhauer for being so approachable when I
was star-struck; everybody at the PIL lab, Stirling University, for giving me
confidence in my ERP data, and pointing me in the right direction to develop my
skills further; Pennie Ingram for everything she does behind the scenes; Prof
Angie Hart for working so hard to understand me, accommodate me and help me
grow; Sarah Hendrickx and Victoria Franklin for showing me who I am at the
iii
point when I probably needed it the most; Terry Buckley for his assistance with
my Latin references. I am indebted to the friends, family and colleagues who
have supported me, especially Anne and Malcolm, Trish and Mitch, Ben, Riley
and Toby, Sue, Anneka, Nikki, Clare, Gary, Rebecca, Munira, JJQ, Graham,
Ringo, Sam C, Sarah RH, Martha, Toni, Chrissy, Kim, Dan, Alice, Jo, Prozac,
alcohol… and of course my cats Dizzy, Smoo and Lennie for their constant
unconditional love, support and amusement. Because it has been essential to
maintaining my motivation, I would like to thank ‘past me’ for making copious
notes whilst reading things I knew ‘future me’ was going to need whilst writing up.
My ultimate thanks goes to Alex for going to the shop to get wine more than his
fair share of the time, taking over the daily cooking and chores, putting up with
me no matter what, and giving me space whilst still being close. Without these
people I could not have achieved what I have, and I offer my regards and thanks
to everybody else who supported me in any respect during the completion of my
thesis.
iv
Publications
The following journal articles and conference presentations have been adapted from
experimental work detailed in this thesis:
Heaver, B., & Hutton, S.B. (2011). Keeping an eye on the truth? Pupil-size changes
associated with recognition memory. Memory, 19(4), 398-405.
doi:10.1080/09658211.2011.575788
Heaver, B., & Hutton, S.B. (2010). Keeping an eye on the truth: Pupil-size,
recognition memory and malingering. International Journal of
Psychophysiology, 77(3), 306. doi:10.1016/j.ijpsycho.2010.06.206
Heaver, B., & Hutton, S.B. (2010). Keeping an eye on the truth: Pupil-size,
recognition memory and malingering. Poster presented at The 15th World
Congress of Psychophysiology of the International Organization of
Psychophysiology (I.O.P.), Budapest, Hungary, 1-4 September, 2010.
Heaver, B., & Hutton, S.B. (2009). Keeping an eye on the truth. Oral presentation at
the Annual Conference of the British Psychological Society, Brighton, UK, 1-3
April, 2009.
Heaver, B., & Hutton, S.B. (2008). You can’t hide your lying eyes: Pupillary
responses during recognition memory. Poster presented at the Cognitive
Section Conference of the British Psychological Society, Southampton, UK, 9
September, 2008.
v
Thesis Summary
UNIVERSITY OF SUSSEX
BECKY HEAVER
THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
PSYCHOPHYSIOLOGICAL INDICES OF RECOGNITION MEMORY
SUMMARY
It has recently been found that during recognition memory tests participants’ pupils
dilate more when they view old items compared to novel items. This thesis sought to
replicate this novel ‘‘Pupil Old/New Effect’’ (PONE) and to determine its relationship
to implicit and explicit mnemonic processes, the veracity of participants’ responses,
and the analogous Event-Related Potential (ERP) old/new effect. Across 9
experiments, pupil-size was measured with a video-based eye-tracker during a
variety of recognition tasks, and, in the case of Experiment 8, with concurrent
Electroencephalography (EEG). The main findings of this thesis are that:
the PONE occurs in a standard explicit test of recognition memory but not in
“implicit” tests of either perceptual fluency or artificial grammar learning;
the PONE is present even when participants are asked to give false behavioural
answers in a malingering task, or are asked not to respond at all;
vi
the PONE is present when attention is divided both at learning and during
recognition;
the PONE is accompanied by a posterior ERP old/new effect;
the PONE does not occur when participants are asked to read previously
encountered words without making a recognition decision;
the PONE does not occur if participants preload an “old/new” response;
the PONE is not enhanced by repetition during learning.
These findings are discussed in the context of current models of recognition memory
and other psychophysiological indices of mnemonic processes. It is argued that
together these findings suggest that the increase in pupil-size which occurs when
participants encounter previously studied items is not under conscious control and
may reflect primarily recollective processes associated with recognition memory.
vii
Contents
Acknowledgements
Publications
Thesis Summary
List of Figures
1.
Introduction – Pupil-Size and Cognitive Function
1.1.
ii
iv
v
xi
1
The Pupil
1.1.1. Anatomy & Physiology
1.1.1.1. Pharmacology
1.1.1.2. Pathology
1.1.2. Pupil-Size Change
1.1.2.1. The Light Reflex
1.1.2.2. The Darkness Reflex
1.1.2.3. The Lid-Closure Reflex
1.1.2.4. The Accommodation Response
1.1.2.5. Pupillary Hippus
1.1.2.6. Iris Colour
1.1.2.7. Role of the Locus Coeruleus in Stimulus-Evoked Dilations
1.2.
Pupillometry Research Literature
1.2.1. Eckhard H. Hess
1.2.1.1. Criticisms of Hess and Early Work
1.2.2. Arousal
1.2.2.1. Fatigue & Sleepiness
1.2.3. “Cognitive Effort” – Kahneman
1.2.3.1. Signal Detection & Discrimination
1.2.3.2. Working Memory
1.2.3.3. Visual Search
1.2.3.4. Effort and the Red Pupillary Reflex
1.2.3.5. Language & Comprehension
1.2.3.6. Imagery
1.2.3.7. Lie Detection
1.2.3.8. Auditory Stimuli
1.2.4. Attention
1.2.4.1. Locus Coeruleus
1.2.4.2. Blinks
1.2.4.3. Decision Making & Uncertainty
1.2.5. Physical Effort
1.2.6. Pupil-Size and Concurrent Psychophysiological Measures
1.2.7. Memory
1.3.
Recognition Memory
1.3.1. Models of Recognition Memory
1.3.1.1. Single-Process Models
1.3.1.2. Dual-Process Models
1.3.1.3. Evaluating Models
1.3.2. Psychophysiological Correlates of Recognition Memory Processes
1.3.2.1. Event-Related Potentials
1.3.2.2. PET and fMRI
1.3.2.3. Pupil-Size
1.4.
Summary
5
5
7
9
10
12
13
14
14
15
16
17
20
22
24
28
32
33
39
40
41
42
42
43
43
44
44
46
47
48
48
50
54
54
54
57
62
66
70
71
72
72
79
2.
82
2.1.
Methods
Pupillometry
2.1.1. Background
82
82
viii
2.1.1.1. Techniques
2.1.1.2. Data Acquisition
2.1.2. Pupil-Size Reporting Variables
2.1.2.1. Pupil Dilation Ratio
2.1.2.2. Measurement Issues
2.1.3. Pupil-Size Analysis
2.2.
Event-Related Potentials (ERPs)
2.2.1. ERP Data Acquisition
2.2.2. ERP Data Analysis
2.3.
Stimuli and Participants
2.3.1. Word Selection
2.3.2. Selection of Participants
3.
Replicating and Extending the Pupil Old/New Effect
Experiment 1 – Implicit vs. Explicit Tests of Recognition
3.1.1. Method
3.1.1.1. Participants
3.1.1.2. Materials/Apparatus
3.1.1.3. Design and Procedure
3.1.1.4. Pupil Recording
3.1.2. Results
3.1.2.1. Behavioural Data: Old/New Responses
3.1.2.2. Behavioural Data: Confidence
3.1.2.3. Pupil-Size Data
3.1.2.4. Pupil-Size Data: Confidence Analysis
3.1.3. Discussion
3.2.
Experiment 1b – Reading Condition
3.2.1. Method
3.2.1.1. Participants
3.2.1.2. Materials/Apparatus
3.2.1.3. Design and Procedure
3.2.1.4. Pupil Recording
3.2.2. Results
3.2.2.1. Behavioural Data
3.2.2.2. Pupil-Size Data
3.2.3. Discussion
3.3.
Experiment 2 – Short vs. Long Presentation Duration
3.3.1. Method
3.3.1.1. Participants
3.3.1.2. Materials/Apparatus
3.3.1.3. Design and Procedure
3.3.1.4. Pupil Recording
3.3.2. Results
3.3.2.1. Behavioural Data
3.3.2.2. Pupil-Size Data
3.3.3. Discussion
3.4.
General Discussion
3.1.
4.
The Role of Conscious Awareness in the PONE
Experiment 3 – Implicit Grammar
4.1.1. Method
4.1.1.1. Participants
4.1.1.2. Materials/Apparatus
4.1.1.3. Design and Procedure
4.1.1.4. Pupil Recording
4.1.2. Results
4.1.2.1. Behavioural Data
4.1.2.2. Pupil-Size Data
4.1.3. Discussion
4.2.
Experiment 4 – Implicit Grammar Replication
4.2.1. Method
4.2.1.1. Participants
4.1.
82
85
86
89
91
94
97
97
101
103
103
104
106
111
111
111
111
112
113
114
114
116
117
119
120
121
121
121
121
121
122
122
122
122
123
126
128
128
128
129
131
131
131
133
136
137
142
148
148
148
148
150
152
152
152
152
153
154
154
154
ix
4.3.
4.2.1.2. Materials/Apparatus/Design/Procedure
Results
4.2.2.1. Behavioural Data
4.2.2.2. Pupil-Size Data
General Discussion
154
155
155
156
158
5.
Malingering and the Old/New Response
161
4.2.2.
5.1.
Malingering and Deception
5.1.1. Background
5.1.1.1. Malingered Memory-Impairment
5.1.1.2. Psychophysiological Detection of Deception
5.1.1.3. Pupillometry Studies
5.2.
Experiment 5 – Malingering and the PONE
5.2.1. Method
5.2.1.1. Participants
5.2.1.2. Materials/Apparatus
5.2.1.3. Design and Procedure
5.2.1.4. Pupil Recording
5.2.2. Results
5.2.2.1. Behavioural Data: Old/New Responses
5.2.2.2. Behavioural Data: Confidence
5.2.2.3. Pupil-Size Data
5.2.2.4. Pupil-Size Data: Confidence Analysis
5.2.3. Discussion
5.3.
Experiment 6 – Methods of Malingering
5.3.1. Method
5.3.1.1. Participants
5.3.1.2. Materials/Apparatus
5.3.1.3. Design and Procedure
5.3.1.4. Pupil Recording
5.3.2. Results
5.3.2.1. Behavioural Data
5.3.2.2. Pupil-Size Data
5.3.3. Discussion
5.4.
Experiment 7 – Emulating Memory-impairment
5.4.1. Method
5.4.1.1. Participants
5.4.1.2. Materials/Apparatus
5.4.1.3. Design and Procedure
5.4.1.4. Pupil Recording
5.4.2. Results
5.4.2.1. Behavioural Data
5.4.2.2. Pupil-Size Data
5.4.3. Discussion
5.4.4. General Discussion
6.
6.1.
ERP and Pupil Old/New Effects
Introduction
6.1.1. Background to ERPs and Recognition Memory
6.1.2. Pupil Responses and ERPs
6.2.
Experiment 8 – Strength of Memory Effect
6.2.1. Method
6.2.1.1. Participants
6.2.1.2. Materials/Apparatus
6.2.1.3. Design and Procedure
6.2.1.4. Pupil Recording
6.2.1.5. Electrophysiological Recording and Analysis
6.2.2. Results
6.2.2.1. Behavioural Data
6.2.2.2. Pupil-Size Data
6.2.2.3. Event-Related Potentials
6.2.2.4. Old/New Effects for Weak and Strong Items
161
161
163
167
174
180
180
180
181
181
183
183
183
184
185
187
188
192
193
193
194
194
196
196
196
197
198
202
203
203
204
205
206
206
206
208
209
209
214
214
214
221
226
227
227
227
228
230
230
231
231
232
233
236
x
6.2.2.5. Weak Old vs. New Items
6.2.2.6. Strong Old vs. New Items
6.2.2.7. Effect of Presentation Frequency on Old/New Effect
6.2.2.8. Early Effects (80-150ms)
6.2.3. Discussion
6.3.
Experiment 9 – Pupil and Behavioural Data Only
6.3.1. Method
6.3.1.1. Participants
6.3.1.2. Materials/Apparatus/Design/Procedure
6.3.1.3. Pupil Recording
6.3.2. Results
6.3.2.1. Behavioural Data
6.3.2.2. Pupil-Size Data
6.3.3. Discussion
6.3.4. General Discussion
7.
General Discussion
7.1.1.
7.1.2.
8.
9.
Summary
Future Directions
Bibliography
Appendices
237
239
241
242
243
245
245
245
246
246
246
246
247
248
249
258
258
268
272
341
xi
List of Figures
Figure 1-1: Excerpt from Bartholomew’s De proprietatibus rerum (1240/1483). ......................... 1
Figure 1-2: Excerpt from Harvey’s Exercitatio duae anatomica de circulatione sanguinis (1649) 4
Figure 1-3: Two muscle groups regulate the size of the pupil. ................................................... 6
Figure 1-4: Pupillary constriction and dilation pathways. ........................................................... 8
Figure 1-5: The accommodation reflex – the pupil constricts as the lens thickens.................... 15
Figure 1-6: Concurrent pupil-size and LC neuron recording in the monkey .............................. 18
Figure 1-7: Percentage pupil-size change in relation to baseline pupil diameter. ..................... 23
Figure 1-8: Mean pupil-size in relation to 10 second lifting and resting periods ........................ 49
Figure 1-9: Human memory systems. ..................................................................................... 55
Figure 1-10: Decision space for the remember-know task ....................................................... 66
Figure 2-1: (a) Head-mounted and (b) tower-mounted EyeLink eye-trackers ........................... 84
Figure 2-2: Pupil as located by eye-tracking software ............................................................. 85
Figure 2-3: Experiment reported as (a) absolute pupil-size values (b) relative values. ............. 89
Figure 2-4: Sources of variation in measurements of pupil diameter........................................ 92
Figure 2-5: Traditional 10-20 and 10-10 electrode configurations. ........................................... 97
Figure 2-6: Geodesic sensor net 64 and 128 channel electrode maps. ................................... 98
Figure 2-7: 128 channel Geodesic sensor net worn by models. .............................................. 98
Figure 2-8: Modified combinational nomenclature for the 10-10 system .................................. 99
Figure 2-9: The 10-10 system overlaid on the Geodesic 128 sensor net. .............................. 100
Figure 3-1: Proportion of correct responses in each condition ............................................... 114
Figure 3-2: Proportion of correct responses to surface and semantic items in each condition 115
Figure 3-3: Average confidence rating for old and new responses to old and new items in the
explicit condition. .................................................................................................................. 116
Figure 3-4: Pupil dilation ratio for old and new items in each condition .................................. 117
Figure 3-5: Pupil dilation ratio for new, surface and semantic items in each condition............ 119
Figure 3-6: Proportion of correct responses in each condition ............................................... 122
Figure 3-7: Pupil dilation ratio for old and new items in each condition. ................................. 123
Figure 3-8: Proportion of correct responses for old and new items in each condition.. ........... 132
Figure 3-9: Pupil dilation ratio for old and new items in all four conditions ............................. 134
Figure 3-10: Pupil dilation ratio for correctly identified old and new items in all four conditions136
Figure 4-1: The two finite state grammars used to generate the strings ................................. 144
Figure 4-2: Pupil dilation ratio for old and new stimuli in both conditions................................ 153
Figure 4-3: Proportion of correct responses. ......................................................................... 156
Figure 4-4: Pupil dilation ratio to old and new items in each condition ................................... 157
Figure 4-5: Pupil dilation ratio to correctly identified old and new items in each condition. ..... 158
xii
Figure 5-1: Proportion of old responses to old and new items for standard and malingering
conditions. ............................................................................................................................ 184
Figure 5-2: Average confidence rating for correct old and new items in each condition.......... 185
Figure 5-3: Pupil dilation ratio for old and new items in standard, malingering and single response
conditions. ............................................................................................................................ 186
Figure 5-4: Proportion of correct responses to old and new items for Standard, Incomplete Effort
and Random conditions. ....................................................................................................... 197
Figure 5-5: Pupil dilation ratio for old and new items in Standard, Incomplete Effort, Random and
Quiet conditions ................................................................................................................... 198
Figure 5-6: Proportion of correct responses to old and new items for all conditions. .............. 207
Figure 5-7: Proportion of correct responses to all items for the four combinations of single and
learning task conditions ........................................................................................................ 207
Figure 5-8: Pupil dilation ratio for old and new items in all conditions .................................... 208
Figure 6-1: Experimental procedure...................................................................................... 229
Figure 6-2: Proportion of correct responses to new, weak and strong items .......................... 232
Figure 6-3: Pupil dilation ratio for correctly identified new, weak and strong items. ................ 233
Figure 6-4: Grand average ERPs for correctly identified new, weak (presented once) and strong
(presented three times) items at midline frontal, central and parietal electrodes. ................... 234
Figure 6-5: Topographical distribution of old/new differences in mean amplitude (µV) for weak
items (first row) and strong items (second row) between 300-700ms. ................................... 235
Figure 6-6: Mean amplitude for new and weak items at frontal, central and parietal electrodes at
300-500ms. .......................................................................................................................... 237
Figure 6-7: Mean amplitude for new and weak items at frontal, central and parietal electrodes at
500-700ms. .......................................................................................................................... 238
Figure 6-8: Mean amplitude for new and strong items at frontal, central and parietal electrodes at
300-500ms ........................................................................................................................... 239
Figure 6-9: Mean amplitude for new and strong items at frontal, central and parietal electrodes at
500-700ms. .......................................................................................................................... 240
Figure 6-10: Proportion of correct responses to new, weak and strong items ........................ 247
Figure 6-11: Pupil dilation ratio for correctly identified new, weak and strong items. .............. 248
1
1. Introduction – Pupil-Size and Cognitive Function
1
…the eye is ot o ly a passi e orga a d o e of the gate ays of
knowledge, but is also a portal through which the working of the brain
e o es a ifest. “a uel Wilks,
, p. -6.
The relationship between the eye and the inner workings of the mind has long been
the subject of philosophy and art, conjuring up powerful imagery and inspiring poems,
songs, and novels. Centuries before the causes of pupil-size changes were
contemplated in scientific and other literature, Franciscan monk Bartholomew The
Englishman wrote about the origins of the word “pupil” in his encyclopaedia On the
Properties of Things, thought to have been written in the 1240s (Keen, 2007; see
Figure 1-1):
Figure 1-1: Excerpt from Bartholomew’s De proprietatibus rerum (1240/1483).
Janisse cites a translation by Travisa (1495), which says, “the blacke of theye ... is
callyd Pupilla in latyn for small, ymages ben seen therin” (p. 1). Whilst Janisse (1977)
described this as one of the earliest references to the pupil, in the original text
Bartholomew refers to the much earlier etymology of seventh century Archbishop and
historian Isidore of Seville (560-636). Isidore believed the eyes were the sensory
2
organ closest to the soul: “… every indication of the mental state is in the eyes,
whence both distress and happiness show in the eyes” (Isidore, 636/2006, XI.i.36).
The section of Isidore’s etymology that Bartholomew refers to reads:
"The pupil (pupilla) is the middle point of the eye in which the power of
vision resides; because small images appear to us there, they are called
pupils, since small children are called pupils. There are many who use the
form pupula, but it is called pupilla because it is pure and unpolluted, just
like 'young girls'. Isidore,
/
, XI.i. .
Derived from the Latin word “pupilla”, meaning “young girls”, the pupil is named for
the tiny reflections of people that can be seen in someone’s eyes. In fact Isidore
based his works on the writings of earlier scholars, such as the Roman naturalist and
historian Gaius Plinius Secundus (Pliny the Elder, 23-79 AD) who wrote in his Natural
History encyclopaedia: “…the small pupil can reflect the entire image of a human
being” (Pliny, 79/1938, XI.LV). Pliny also commented on the mind-eye connection,
although he too did not explicitly link it to the pupil:
No ody has eyes of o ly o e olour … No other part of the ody supplies
greater indications of the mind – this is so with all animals alike, but
specially with man – that is, indications of self-restraint, mercy, pity,
hatred, love, sorrow, joy. Pli y, /
, XI.LIV .
From the earliest writings, these sentiments have been echoed by poets,
philosophers and authors including Guillaume de Salluste (1544-1590), Joshua
Sylvester (1563-1618), Shakespeare (1564-1616), Descartes (1596-1650), Johann
Kaspar Lavater (1741-1801), Byron (1788-1824) and Tennyson (1809-1892), all of
whom have written about the possibility of our eyes revealing the state of our mind
(Andreassi, 2000; Clark, 1885; Loewenfeld, 1958; Wilks, 1885). From the sixteenth
century onward there is a marked increase in the amount of writing about the pupil
3
(Loewenfeld, 1958). Berrien and Huntington (1943) argue that the pupil had been
seen as an emotional barometer since at least the 1800s, and that it had long been
associated with “arousal”. Honoré de Balzac (1799-1850) and Charlotte Brontë
(1816-1855) appear to be the first to specifically mention changes in pupil-size in their
texts (Clark, 1885; Wilks, 1883; 1885). In Jane Eyre Brontë (1847/1946) wrote, "Pain,
shame, ire, impatience, disgust, detestation, seemed momentarily to hold a quivering
conflict in the large pupil dilating under his ebon eyebrow" (p. 163). De Balzac
(1841/2000) described pupil constriction in response to positive emotions, “The blue
of the iris expanded like a flower, diminishing the dark circle of the pupil, and seeming
to float in a liquid and languishing light that was full of love” (p. 31). He also links
dilation of the pupil with negative emotions:
When Monsieur de Grandville… whom she declined to take as a husband,
kissed her hand with an earnest expression of regret, the new bishop
noticed the strange manner in which the black pupil of Veronique's eyes
suddenly spread over the blue of the iris, reducing it to a narrow circle. The
eye betrayed unmistakably some violent inward emotion. (de Balzac,
1841/2000, pp. 79-80).
Additionally, he asks whether dilation might reflect passion:
The pupils of her eyes, gifted with the power of great expansion, widened
until they covered the whole surface of the blue iris except for a tiny circle...
Was it the storm of restrained passions; was it some power coming from
the depths of the soul, which enlarged the pupils in full daylight as they
sometimes in other eyes enlarge by night, darkening the azure of those
elestial or s? de Balza ,
/
, p.
.
In a thought piece to Brain in 1883, and again in a letter to Nature in 1885, eminent
physician to Queen Victoria, Sir Samuel Wilks, drew attention to converging sources
of evidence from doctors and physiognomists, and his own observations of the mood
4
of his pet parrot, that in addition to responding to emotions, the pupil may manipulate
them in others (Wilks, 1883; 1885). For example, Foster wrote in his Textbook of
Physiology (1891) that the pupil dilates “as an effect of emotions” (p. 1172).
Whilst much of the early pupil literature comprises either medical accounts of the
physiological changes, or literary descriptions of the pupil-size changes
accompanying their characters’ emotions, the earliest scientific report of pupil
changes in response to internal states referred to by Wilks (1885) was by another
physician, William Harvey (1578-1657), who wrote the first complete and detailed
account of the circulatory system in the seventeenth century. Harvey writes, “In
anger the eyes are fiery, and the pupils contracted” (Harvey, 1649, p. 152; see Figure
1-2). Fontana (1765, cited in Loewenfeld, 1958) is thought to have provided the first
detailed account of the psychological stimuli that dilate the pupil, referred to in older
texts as dilatation.
Figure 1-2: Excerpt from Harvey’s Exercitatio duae anatomica de circulatione sanguinis (1649).
Parallel advances in the understanding of basic neurophysiology meant that the
relationship between pupil-size and the autonomic nervous system was established
by the 1850s (e.g., Bernard, 1852; Budge, & Waller, 1851; Kuntz, 1929). Although
Wilks (1885) felt that dilated pupils were associated with relaxed contemplation, and
constricted pupils with concentration, Fontana (1765, cited in Loewenfeld, 1958),
Gratiolet (Gratiolet, & Grandeau, 1865) and Hack Tuke (1884) proposed that fear
caused dilation, while Clark (1885) posited that any “strong mental emotion” would
produce dilation of the pupil (p. 433). The discussion even drew the attention of
5
Charles Darwin in Expression of the Emotions in Man and Animals (1872), although
he remained unconvinced and asked that further research be conducted.
Thus, even from the earliest writings, there appears to have been some agreement
that there are numerous non-luminance-based influences on pupil-size, but
considerable debate as to exactly what those influences are, and what their precise
effects on pupil-size are – a situation that some may argue remains today. This
thesis examines a relatively novel psychophysiological index of recognition memory,
the Pupil Old/New Effect (PONE). This chapter will provide the theoretical
background for the experiments presented in Chapters 3-6, starting with a brief
description of the anatomy and physiology of the pupil, followed by a summary of the
history of pupillometry, and research into the effects of cognitive processes on pupilsize up to the present day. The next section introduces some current models of
recognition memory and describes some of the key paradigms researchers have
used to test them. This is followed by a consideration of other psychophysiological
correlates of recognition memory, such as the Event-Related Potential (ERP) old/new
effect, which sets the scene for a comprehensive exploration of the literature on the
PONE to date.
1.1. The Pupil
1.1.1.
Anatomy & Physiology
The pupil is the circular aperture at the centre of the iris which allows light from our
environment to pass freely to the light-sensitive sensory cells of the retina. The size
of the pupil is determined by the iris muscles, the radial dilator pupillae and the
concentric smooth muscle circles of the sphincter pupillae, which work in opposition
6
to dilate (mydriasis) and constrict (miosis) the pupil (see Figure 1-3; Löwenstein, &
Loewenfeld, 1962; Beatty, & Lucero-Wagoner, 2000; Reeves, & Swenson, 2004).
circular sphincter pupillae
pupil
iris
radial dilator pupillae
Figure 1-3: Two muscle groups regulate the size of the pupil: circular sphincter pupillae contract to
make the pupil smaller, radial dilator pupillae contract to make the pupil larger.
The sphincter and dilator muscles are innervated by the Autonomic Nervous System
(ANS) parasympathetic and sympathetic pupillomotor fibres of the third cranial nerve
respectively (Foster, 1891; Reeves, & Swenson, 2004). The sympathetic and
parasympathetic systems work in opposition, and tonic activation in both systems
balance each other to produce an average waking pupil-size in ambient illumination of
2-6mm, with an average of 5mm and a range of 1-9mm (Beatty, & Lucero-Wagoner,
2000; Reeves, & Swenson, 2004). An increase in efferent activity to either muscle
group leads to increasing central inhibition of the motor-nucleus of the other group
(Miller, & Newman, 2005).
It has been shown that average ‘resting’ pupil-size decreases curvilinearly from about
age twenty, becoming asymptotic at around sixty years (Birren, Casperson, &
Botwinick, 1950), and that pupil-size becomes more variable with age (Kumnick,
1954; 1956). It has been suggested that this may be due to increased rigidity of the
7
sphincter muscle (Reeves, & Swenson, 2004; Spector, 1990) or decreasing
sympathetic activity (Bitsios, Prettyman, & Szabadi, 1996a). The pupils are ‘yoked’,
so that a change in one pupil shows a consensual response in the other because
each retina inputs equally to the pretectal region and Edinger-Westphal nuclei, and
the sphincter muscle of each pupil receives equal efferent output from these brain
regions (Reeves, & Swenson, 2004). Around one person in four has naturallyoccurring benign asymmetry of the pupils of up to 0.5mm, known as anisocoria
(Reeves, & Swenson, 2004).
1.1.1.1.
Pharmacology
The effects of certain substances on the pupil have been known since Ancient Roman
times, when the juices of the poisonous plant Atropa belladonna (deadly nightshade)
were used to dilate the pupil for cataract surgery (Loewenfeld, 1958). Apparently lost,
this knowledge was rediscovered in the seventeenth century when ladies purportedly
applied the plant to their eyes so that their dilated pupils would make them appear
more attractive to their unsuspecting admirers (belladonna meaning ‘beautiful woman’
in Italian) (Forbes, 1977; Wilks, 1883; Wootton, 1910). Belladonna contains atropine,
which blocks the parasympathetic input to the sphincter muscle.
Parasympathetic nerves transmit messages relating to “rest and digest” functions,
using Acetylcholine (ACh) neurotransmitter both centrally and peripherally, and the
efferent fibres innervating the sphincter muscle of the iris originate in the midbrain
Edinger-Westphal nucleus (Figure 1-4). ACh acts on the sphincter via muscarinic
receptors, leading to constriction of the pupil (Fountoulakis, 1999). Topically applied
pharmacological ACh agonists, such as pilocarpine, and ACh-breakdown inhibitors,
such as physostigmine, cause constriction of the sphincter, whereas ACh
antagonists, such as atropine, relax the sphincter, enhancing the efforts of the dilator
8
muscle (Miller, & Newman, 2005). Miosis can also be induced centrally by the effects
of chloroform, sedation, and through direct stimulation of the Edinger-Westphal
nucleus by morphine and other opioids (Foster, 1891; Knaggs, Crighton, Cobby,
Fletcher, & Hobbs, 2004; Miller, & Newman, 2005).
Sympathetic fibres carry signals relating to “fight or flight” responses, using ACh
centrally and Norepinephrine (NE, also known as noradrenaline) peripherally, and the
efferent fibres innervating the dilator muscle originate in the hypothalamic motor area
of the diencephalon (Figure 1-4). NE acts on the dilator via α-adrenergic receptors,
leading to dilation of the pupil (Fountoulakis, 1999). NE agonists, such as ephedrine,
and NE-reuptake inhibitors such as cocaine, cause dilation of the pupil, whereas NE
antagonists, such as thymoxamine, relax the dilator muscle and reduce its opposition
of the sphincter. Mydriasis can be induced centrally via Selective SerotoninReuptake Inhibitor (SSRI) antidepressants and Lysergic Acid Diethylamide (LSD), as
well as N-Methyl-D-Aspartate (NMDA) glutamate antagonists, such as ketamine, in
addition to opioid withdrawal and alcohol poisoning (Doughty, 2001; Foster, 1891).
Figure 1-4: Pupillary constriction and dilation pathways (from Beatty, & Lucero-Wagoner, 2000).
9
As there are a number of stages in the efferent pathways, and cholinergic
transmission is common to both sympathetic and parasympathetic pathways within
the central nervous system, some substances – such as nicotine (Loewenfeld, 1999),
antidepressants (Doughty, 2001) and alcohol (Skoglund, 1943) – can cause either
constriction or dilation depending on concentration and activity in other parts of the
pathway.
1.1.1.2.
Pathology
Certain health states and pathologies can affect the size of the pupil. Locally pupil
dilation can result from the increased intraocular pressure in glaucoma, and
constriction can occur in response to treatment when the pressure decreases (Miller,
& Newman, 2005). Neurological conditions affecting the central and peripheral
nervous system, such as epilepsy, migraine, multiple sclerosis and lesions in the
autonomic nervous system, can all have transient or long-lasting effects on both
pupil-size and pupil responses (Grunberger, Linzmayer, Majda, Reitner, & Walter,
1996; Harle, Wolffsohn, & Evans, 2005). Sympathetic lesions, such as those that
occur in Horner’s Syndrome, can result in pupil constriction, while diabetes mellitus
can lead to both sympathetic and parasympathetic autonomic neuropathy, resulting in
loss of the light reflex (Argyll-Robertson pupil), and smaller resting pupil-size (Reeves,
& Swenson, 2004). Argyll-Robertson pupil can also be caused by tertiary-stage
neurosyphilis (Reeves, & Swenson, 2004). Compared to healthy populations, pupil
abnormalities, such as increased or decreased diameter, decreased reactivity and
pupil asymmetry, are also found in patients with organic and functional mental health
problems such as Alzheimer’s, depression and schizophrenia (Granholm et al., 2003;
Sokolski, Nguyen, & DeMet, 2000; Steinhauer, Hakerem, & Spring, 1979),
neurodiverse conditions like autism and ADHD (Martineau et al., 2011; Zahn, Little, &
10
Wender, 1978), and substance issues, including alcohol and heroin misuse (Ghodse,
Greaves, & Lynch, 1999; Grunberger et al., 1998).
1.1.2.
Pupil-Size Change
Löwenstein and Loewenfeld (1962) argue convincingly that all physical stimuli arriving
at the senses, all somatic and visceral afferents, all mental processes including
intentional efforts and motor responses, emotions, and all centrally mediated arousal
responses, trigger a pupillary reflex dilation, also called the psychosensory reflex
(Foster, 1891; Hess, 1965; 1975; Loewenfeld, 1999; Löwenstein, & Loewenfeld,
1962). The exceptions to this are certain visual reflexes, such as changing focus
from far to near objects, and increased light falling on the retina, which cause
constriction (see section 1.1.2.1). Some relatively early experiments in the late
nineteenth century demonstrated pupillary dilations, occurring without changes in
blood pressure, in response to peripheral tactile and pain stimuli, in animals that were
conscious, under anaesthesia, and partially or completely paralysed with the
acetylcholine antagonist curare (Schiff, & Foa, 1874; Schiff, 1875). Löwenstein and
Friedman (1942) report that Schiff’s earlier work (~1867) referred to the pupil as the
body’s “finest esthesiometer” (device measuring the skin’s tactile sensitivity) (p. 969).
Interestingly, internal psychological events also cause pupil dilation and the first
cognitive pupillometry study in humans was probably conducted by Heinrich (1896)
who measured pupillary dilations evoked by mental multiplication. Such findings lead
Oswald Bumke to observe in 1911 that: “…every active intellectual process, every
psychical effort, every exertion of attention, every active mental image, regardless of
content, particularly every affect just as truly produces pupil enlargement…” (cited by
Hess, 1975, pp. 23-4). Löwenstein and Loewenfeld (1962) suggest that unlike the
pupil-size changes that occur in response to increased peripheral activity in the
11
autonomic nervous system, the psychosensory reflex is under higher cortical control
and reflects the level of cortico-thalamo-hypothalamic activity, which itself is
influenced by sensory stimulation, emotions and spontaneous thought (Kahneman,
1973). There are extensive cortical and limbic inputs to both the Edinger-Westphal
nucleus, increasing inhibition of the sphincter muscle, and to the hypothalamus,
causing the dilator muscle to contract (Silk et al., 2009). In addition, the anterior
cingulate cortex, thought to be involved in emotion regulation (Szabadi, & Bradshaw,
1996), inputs directly to the midbrain reticular formation which, when stimulated,
increases pupil-size (Beatty, 1986). Pupil dilation also results from direct stimulation
of limbic structures such as the amygdala (Koikegami, & Yoshida, 1953). However,
despite higher order influences on pupil-size, even the earliest writings suggested that
the pupil itself is not under voluntary control: “…it is not in our power to bring the will
to act directly on the iris by itself. This fact alone indicates that the nervous
mechanism of the pupil is of a special character…” (Foster, 1891, p. 1172).
Pupil-size is under the antagonistic control of both parasympathetic and sympathetic
inputs, and each pathway is subject to various types of inhibition and excitation at
each synapse. Therefore numerous internal and external factors, including stimulus
characteristics like illumination, colour, contrast, and duration, are known to have
individual and interacting effects on pupil-size (Yamaji, Hirata, & Usui, 2000). The
following section briefly describes the major sources of pupil-size change, starting
with changes in illumination via the light reflex, and accommodation via the near
reflex. It is important to note that even these relatively rapid and automatic reflex
processes are also modified by individual factors such as age, fatigue and emotional
state, and these interactions will also be discussed.
12
1.1.2.1.
The Light Reflex
The pupil is usually between 2-6mm in ambient illumination, dilating in dim light, and
constricting in bright light (Beatty, & Lucero-Wagoner, 2000; Reeves, & Swenson,
2004; Young, & Biersdorf, 1954). This constriction is known as the Pupillary Light
Reflex (PLR), and serves to regulate the total intensity of light entering the eye,
optimising image quality (Kardon, 1995). In response to large increases in luminance
the pupil can decrease in diameter by more than 50% in just 200ms (Miller, &
Newman, 2005), peaking between 500-1000ms (Beatty, & Lucero-Wagoner, 2000).
PLR amplitude and constriction velocity decrease with age, due to sympathetic deficit
and smaller initial pupil-size (Birren et al., 1950; Bitsios et al., 1996a).
Photosensitive retinal ganglion cells respond to increased light falling on the retina.
The PLR acts afferently via the optic nerve to the midbrain pretectal region, and
efferently to the Edinger-Westphal nucleus, then via parasympathetic fibres in the left
and right oculomotor nerves to the ciliary ganglions causing contraction of the pupil
sphincter muscle. Simultaneous inhibition of the sympathetic fibres innervating the
dilator muscles causes the antagonistic muscles to relax, and the pupil constricts to
reduce the amount of light falling on the retina (Loewenfeld, 1999). Both pupils
generally change equally even if light only enters one eye, unlike some animals, such
as frogs and birds, where pupil light reflexes are independent (Foster, 1891). In
humans this is due to the equal bilateral input from each retina to the pretectal region,
and bilateral output from the pretectal region to the Edinger-Westphal nuclei (Reeves,
& Swenson, 2004; Thompson, 1947).
Although the PLR is automatic, its amplitude can be reduced by evoking a
simultaneous psychosensory dilation with emotional or painful stimuli (Bender, 1933;
Gang, 1945; Miller, & Newman, 2005). For example, Bitsios, Szabadi and Bradshaw
13
(1996b; 2002) found that when a light stimulus follows the threat of an aversive
stimulus, such as electric shock, the reflexive constriction is smaller (fear-inhibited
light reflex), and is negatively correlated with self-reported anxiety. This amplitude
reduction can itself be attenuated by anxiolytic drugs such as diazepam (Bitsios,
Philpott, Langley, Bradshaw, & Szabadi, 1999). Sustained cognitive processing
during a task (such as counting backwards in intervals of 7) has also been shown to
diminish the PLR, an effect that may reflect cortical inhibition of the Edinger-Westphal
nuclei (Steinhauer, Condray, & Kasparek, 2000; Steinhauer, Siegle, Condray, &
Pless, 2004). In both the fear- and cognitive task-inhibited PLR overall pupil-size is
larger to begin with, due to increased emotional arousal and cognitive load (see
section 1.2). Consequently, careful consideration needs to be given to how pupil-size
change is measured and whether absolute changes in PLR amplitude should be of
the same magnitude when initial diameters vary (see Chapter 2, section 2.1.2.2, for
further discussion of this issue in pupillometry research).
1.1.2.2.
The Darkness Reflex
The darkness reflex involves dilation of the pupil due to cessation of the sympathetic
inhibition caused by a constant light source (Löwenstein, & Loewenfeld, 1964).
Pupils take longer (300ms) to begin dilating in response to darkness than to constrict
to light, and this latency does not change with age. However the maximum velocity of
constriction and maximum dilation, of between 3-9mm, both decrease from about age
fifteen (Birren et al., 1950; Bitsios et al., 1996a; Miller, & Newman, 2005). The
amplitude of the darkness reflex is related to the length of time in the dark (Stark,
1962, cited by Loewenfeld, 1999), and rather than just an absence of the PLR, it is
thought that signals from the retina may inhibit the oculomotor nerves. Retinal
disease, which leads to the loss of these signals, causes “paradoxical” constriction in
response to darkness (Miller, & Newman, 2005).
14
1.1.2.3.
The Lid-Closure Reflex
The lid-closure reflex manifests as a systematic miosis (of around 8% surface area),
occurring after blinks, which is followed by redilation (DeLaunay, 1949; Hupé, Lamirel,
& Lorenceau, 2009; Loewenfeld, 1999), although some researchers have observed
only a dilation after eye-blinks (Fukuda, Stern, Brown, & Russo, 2005).
Photosensitive retinal cells increase in sensitivity during brief (<500ms) interruptions
of light, such as during a blink, however the lid-closure reflex does not occur after
blinks in darkness (Hupé et al., 2009).
1.1.2.4.
The Accommodation Response
The pupil constricts when we change our focus from looking at a far away object to a
near object, and when the eyes converge, such as when looking at the tip of the nose
(Foster, 1891; Löwenstein, & Loewenfeld, 1964; Reeves, & Swenson, 2004). The
purpose of the accommodation response is to maintain a focussed image on the
surface of the retina. Unlike a camera, where the lens moves forward to focus on a
nearer object, the lens of the eye is elastic; it changes in thickness, becoming more
convex, and therefore increases in refractive power (Beatty, & Lucero-Wagoner,
2000; Foster, 1891; see Figure 1-5). Young healthy eyes are able to accommodate
within 350ms (Erichsen, Hodos, & Evinger, 2000), adjusting focus from distant to near
objects via three muscle groups – lens curvature is increased by contraction of the
ciliary muscle/release of the zonule fibres, the eyes converge through contraction of
the medial rectus muscle, and the pupillary sphincter muscle constricts the pupil
(Beatty, & Lucero-Wagoner, 2000).
15
parallel rays of light
from distant object
sphincter muscle constricts
rays of light from
nearby object
lens thickens
ciliary muscles contract
Figure 1-5: The accommodation reflex – the pupil constricts as the lens thickens (adapted from Nave,
2010).
The pupil constriction in the near reflex shares the same efferent parasympathetic
pathway with the PLR from the Edinger-Westphal nucleus onwards, and so is also
susceptible to central inhibition from concurrent psychosensory dilations (Miller, &
Newman, 2005). Its magnitude is also influenced by illumination, such that in dim
light the constriction associated with focussing on a near object is smaller than in
bright conditions (Miller, & Newman, 2005). Accommodation, convergence and PLR
constrictions are neurologically distinct and are dissociated in conditions like ArgyllRobertson pupil (where there is no PLR but normal responses to accommodation and
convergence), diphtheritic neuritis (where there is a preserved PLR but no change
with accommodation), and pretectal lesions (where normal constriction occurs for
accommodation, but not for convergence; Reeves, & Swenson, 2004; Spector, 1990).
1.1.2.5.
Pupillary Hippus
The iris is a vascular structure and rhythmic changes in pupil-size of around 1% occur
with heart beat and breathing due to fluctuations in blood pressure (Foster, 1891).
However, there are other rhythmic but irregular, oscillating, consensual (therefore of
central origin) contractions and dilations of reasonably large amplitude ~1mm (10-
16
20%) with a period of between 5-25s (0.04-0.2Hz; Bouma, & Baghuis, 1971;
McLaren, Erie, & Brubaker, 1992; Woodmansee, 1966). They occur under constant
illumination and fixation, and vary with intensity of illumination rather than pulse and
respiration (Loewenfeld, 1958). This pupillary unrest is also known as ‘hippus’
(possibly from hippos, Greek for ‘horse’, suggestive of a galloping rhythm; Beatty, &
Lucero-Wagoner, 2000), which was originally used to describe pathological changes
in pupil-size that can occur in phase with EEG recordings in seizure disorders (MüllerJensen, & Hagenah, 1978) or respiration in Cheyne-Stokes (Sullivan, Manfredi, &
Behnke, 1968). However, spontaneous consensual hippus usually has no clinical
significance, it is induced by changes in lighting level, becoming more obvious in
brighter light and when the pupil is small (Bouma, & Baghuis, 1971; Miller, &
Newman, 2005).
In addition to sensory and endogenous influences, hippus varies according to
“arousal” and “cognitive effort” (see sections 1.2.2 and 1.2.2.1). Hippus is amplified
by fatigue and passivity, but is suppressed by alertness and carrying out mental
activities such as mental arithmetic (Bouma, & Baghuis, 1971; Kahneman, 1971;
Miller, & Newman, 2005). Due to this suppression, a problem arises in taking
baseline measures of pupil-size before pupillometry experiments in that hippus may
be occurring during the baseline period, whereas it will be attenuated during the
experimental task, resulting in a hippus artefact in pupil-size data (Janisse, 1977; see
Chapter 2 section 2.1.2.2 for further discussion).
1.1.2.6.
Iris Colour
Iris colour has been considered a possible confound in pupillometry research. Dark
irises are associated with decreased sympathetic reactivity, muscle motility, and
contraction amplitude compared to paler irises (Beck, 1967; Dain, Cassimaty, &
17
Psarakis, 2004; Gambill, Ogle, & Kearns, 1967; Hess, 1975; Spector, 1990).
However, not all research into iris colour has found an effect on pupil-size (Birren et
al., 1950; Bradley et al., 2010; Goodrich, 1974; Kumnick, 1954; Wenger, & Videbeck,
1969) and Janisse (1977) suggests that previous findings may have been an artefact
of distinguishing pupil-size from a dark iris. The mixed findings may reflect the
numerous morphological and chemical factors which influence perceived iris colour
(with >240 degrees of freedom; Daugman, 2003) and, for these reasons, this variable
is not given further consideration.
1.1.2.7.
Role of the Locus Coeruleus in Stimulus-Evoked Dilations
The Locus Coeruleus (LC) is a neuromodulatory nucleus in the dorsal pons of the
brainstem which responds to salient stimuli (e.g., targets) with a transient increase in
firing rate. The LC projects throughout the forebrain, providing all of the forebrain and
most of the brain’s Norepinephrine (NE). This influence makes the LC partly
responsible for regulating all cognitive, emotional and motivational states (see
Berridge, & Waterhouse, 2003, for a review; Samuels, & Szabadi, 2008a; Sara,
2009). The LC receives afferents from the Anterior Cingulate Cortex (ACC) and
Orbitofrontal Cortex (OFC), which Aston-Jones and Cohen (2005) have argued are
involved in monitoring “task-related utility” (the cost/benefit of continuing with the
current task vs. looking for a new opportunity), and supplying information about
conflict and reward in the cognitive system (Aston-Jones et al., 2002; Botvinick,
Braver, Barch, Carter, & Cohen, 2001; Rajkowski, Lu, Zhu, Cohen, & Aston-Jones,
2000; Zhu, Iba, Rajkowski, & Aston-Jones, 2004). The LC-NE system also plays a
role in wakefulness (Aston-Jones, Foote, & Bloom, 1984; Jouvet,1969), simple
decision-making and the regulation of task engagement through distribution of
attentional resources (e.g., Aston-Jones, Rajkowski, & Kubiak, 1997; Clayton,
Rajkowski, Cohen, & Aston-Jones, 2004), and recent research has argued that it also
18
plays a major role in the modulation of the psychosensory pupillary response
(Gilzenrat, 2006).
NE is the neurotransmitter released by the sympathetic fibres that innervate the pupil
dilator muscle. Single-cell intracranial recordings and in-vivo stimulation of the LC in
conscious monkeys and rodents performing memory and sensory tasks show two
modes of activity – regular, continuous tonic firing (1-3Hz), interspersed with short
bursts of phasic firing (8-10Hz; Aston-Jones, Chiang, & Alexinsky, 1991; Aston-Jones
et al., 1997; Aston-Jones, Rajkowski, Kubiak, & Alexinsky, 1994). Both modes have
been shown to vary with vigilance and task performance measures, such as stimulus
processing efficiency (Aston-Jones, & Cohen, 2005; Rajkowski, Majczynski, Clayton,
& Aston-Jones, 2004; Samuels, & Szabadi, 2008b; Usher, Cohen, Servan-Schreiber,
Rajkowski, & Aston-Jones, 1999), and a high correlation (0.6) between spike
frequency and pupil diameter has been found, whereby large pupil diameter equates
to high LC activity (Rajkowski, Kubiak, & Aston-Jones, 1993; 1994; see Figure 1-6).
Figure 1-6: Concurrent pupil-size and LC neuron recording in the monkey (from Rajkowski, Kubiak, &
Aston-Jones, 1993).
19
A differentiation between tonic and phasic pupil activity has been observed in humans
(Dureman, & Scholander, 1962), and whilst no current technique allows direct
recording of LC neurons in humans, the relationship between pupil-size and task
performance found in other animals has been confirmed in human participants
(Gilzenrat, Cohen, Rajkowski, & Aston-Jones, 2003). Additionally the stimulusevoked pupil dilations seen in human vigilance experiments (e.g., Beatty, 1982a) are
consistent with the phasic pupil dilations which arise from phasic LC activity reported
in the animal literature (Aston-Jones, & Cohen, 2005; Murphy, Robertson, Balsters, &
O’Connell, 2011). Even though the relationship between pupil-size and LC activity
has not been fully characterised, and is subject to debate (see Nieuwenhuis, De
Geus, & Aston-Jones, 2011a), pharmacological up- or down-regulation of central NE
release in humans provides strong supporting evidence by mimicking increased and
decreased LC activity respectively. For example, sympathomimetic drugs such as
modafinil, yohimbine, and reboxetine increase subjective alertness by increasing
central NE; they also increase baseline pupil-size, reduce pupillary variability and
fatigue waves, and reduce the amplitude and velocity of the darkness reflex (Hou,
Freeman, Langley, Szabadi, & Bradshaw, 2005; Phillips, Bitsios, Szabadi, &
Bradshaw, 2000a; Phillips, Szabadi, & Bradshaw, 2000b). In contrast, sympatholytic
drugs such as clonidine and prazosin decrease subjective alertness by decreasing
central NE; they also decrease baseline pupil-size, increase pupillary variability and
fatigue waves, and increase the velocity of the darkness reflex (Hou et al., 2005;
Phillips et al., 2000a; 2000b). As a result of this converging evidence, researchers
increasingly use pupil-size as an indirect indicator of LC activity to investigate aspects
of human attention, such as orienting of attention to external cues (Gabay, Pertzov, &
Henik, 2011), switching attentional focus (Einhauser, Stout, Koch, & Carter, 2008),
and changes in cognitive control state (Gilzenrat, Nieuwenhuis, Jepma, & Cohen,
2010; Jepma, & Nieuwenhuis, 2011).
20
Connection between the LC and the pupil can also be seen in patient populations.
The LC receives input from the vagus nerve, known to play a role in memory
formation (Clark, Krahl, Smith, & Jensen, 1995; Clark, Naritoku, Smith, Browning, &
Jensen, 1999), via projections from the solitary tract. Individuals who benefit from
vagus nerve stimulation (which modulates NE release), such as those with treatmentresistant epilepsy, depression, anxiety disorders, Alzheimer's Disease (AD), migraine,
fibromyalgia, and tinnitus (for a review see Groves, & Brown, 2005; Engineer et al.,
2011; Ghanem, & Early, 2006; Lange et al., 2011), have been shown to demonstrate
atypical pupil responses. For example, migraine sufferers show various pupil
abnormalities such as anisocoria (Evans, & Jacobson, 2003), reduced velocity and
amplitude of the PLR (Mylius, Braune, & Schepelmann, 2003), and hyper- or hyporesponsivity to pharmacological agents that affect the autonomic system
(Fanciullacci, Galli, Pietrini, & Sicuteri, 1977). Harle et al. (2005) showed that
anisocoria and inter-ocular differences in PLR latency persist during the nonheadache phase, independent of time since last migraine, severity or frequency,
suggesting sustained autonomic imbalance in migraineurs. In AD patients, topical
application of tropicamide, a cholinergic antagonist, produces significantly larger pupil
dilations than for vascular dementia patients or young non-AD patients (Iijima et al.,
2003), and smaller peak PLR constriction amplitude (Granholm et al., 2003).
1.2. Pupillometry Research Literature
Pupillometry has been used to look at a wide variety of psychosensory and
physiological functions in a variety of animals (cats, chickens, dogs, fish, frogs,
guinea pigs, monkeys, pigeons, rabbits, and rats), producing a literature that has
grown exponentially. In a fascinating dissertation, Loewenfeld (1958) reviewed over
1300 pupil references dating back to the 1st century AD. She made a distinction
between 114 references to the pupil in historical pre-1830 literature, and 1204 pieces
21
from 1830-1957, because the existence of muscles within the iris, and their role in the
dilation of the pupil, was only established around this time. By the time of her later
publication, Loewenfeld (1999) listed over 15,000 references to pupil research leading
up to ~1985 (Steinhauer, 2002).
Psychosensory fluctuations in pupil-size are usually no more than 0.5-1.0mm and are
therefore difficult to see with the naked eye (Beatty, 1982b; Beatty, & LuceroWagoner, 2000; methods of measuring pupil-size in cognitive pupillometry are
described in Chapter 2). These tiny yet consistent pupillary changes have no
apparent functional purpose or evolutionary cost, and appear to reflect dynamic
changes in cognitive processing (Beatty, & Lucero-Wagoner, 2000). Although the
pupil is under peripheral autonomic control, without an obvious link to central
processing, evidence shows that the variations reliably and precisely track changes in
cognition (Beatty, 1982b; 1986; Goldwater, 1972). Described as “a permanently
implanted electrode” and “the only visible part of the brain” (Janisse, 1977, p. 1), the
pupil is of particular interest to cognitive psychophysiology researchers because it
potentially provides a unique physiological reporter variable to measure psychological
processes independently of subjective report (Beatty, & Lucero-Wagoner, 2000).
Task-evoked pupil-size changes begin 400ms post-stimulus (Partala, & Surakka,
2003), peaking after around 1-2s, and constricting once the task is complete, either
slowly (Kahneman, & Beatty, 1966; Hess, 1972) or rapidly (Bernhardt, Dabbs, & Riad,
1996) dependant on post-processing. Like other psychophysiological measures,
such as blood flow in functional Magnetic Resonance Imaging (fMRI), it has been
argued that pupillary changes provide an indirect measure of processing intensity
without causal links or face validity (Just, & Carpenter, 1993). As such they are also
subject to the problem of ‘psychophysiological inference’ (the assumption that a
22
physiological response has a consistent one-to-one and context-independent
relationship with the psychological variable of interest) originally raised by William
James (1890) (Cacioppo, & Tassinary, 1990; Cacioppo, Tassinary, & Berntson,
2000). However, Beatty and Lucero-Wagoner (2000) liken the use of Task-Evoked
Pupillary Responses (TEPRs) to the use of reporter genes in molecular biology,
which have advanced understanding of the genome, and suggest that TEPRs might
do the same in psychophysiology for understanding cognition.
Given the vast number of factors influencing pupil-size, and the numerous afferent
pathways involved, the question is whether it is possible to isolate and study the
systematic effects of individual influences within the pupillary signal. The next section
of this chapter will briefly review the main areas of human cognitive pupillometry from
the 1960s onwards.
1.2.1.
Eckhard H. Hess
Pupil changes in relation to cognitive activity were first demonstrated around the turn
of the twentieth century (e.g., Heinrich, 1896; Mentz, 1895; Roubinovitch, 1900), but
the findings remained largely within the European literature. It was not until the
1960s, and the experiments (and controversy) of Eckhard H. Hess, that a resurgence
of interest in pupil-size in North America lead to more systematic and thorough
investigation, using increasingly sophisticated pupillometry techniques and equipment
(see Beatty, 1982b; Hess, 1975; Kahneman, 1973; Steinhauer, 2002). Although by
no means the first, Hess is widely acknowledged as a key figure in the history of
cognitive pupillometry (called pupillography prior to the use of computerised
measures), establishing a clear relationship between psychological processes and
changes in pupil-size during three decades of published research (Janisse, 1977).
23
In their first study, Hess and Polt (1960) showed four male and two female
participants, pictures of a baby, mother and child, partially nude male, partially nude
female, and a landscape. Pupil-size was manually measured from a 16mm filmstrip,
of the participants’ eyes during the task, projected onto a screen. Male participants’
pupils dilated most to the picture of the nude female, whereas female participants’
pupils dilated most to the picture of mother and child. Hess and Polt (1960) replicated
their results, interpreting the findings as showing the interest value of the pictures,
whereby more interesting stimuli elicited larger pupil dilations. In 1964 Hess and Polt
conducted the first rigorous investigation of pupil dilation in relation to mental
arithmetic, confirming the findings of Heinrich (1896). Again, measuring pupil-size
from projected 16mm filmstrip, they asked five participants to carry out four mental
multiplications, giving a total of twenty data points. They found that pupil-size
increased whilst the answer was being calculated, and maximum dilation increased
from 10.8% to 21.6% approximately monotonically in proportion to calculation
difficulty (see Figure 1-7), suggesting that changes in pupil-size could be used to
directly measure cognitive activity as it occurs (Hess, & Polt, 1964).
Figure 1-7: Percentage pupil-size change in relation to baseline pupil diameter (in mm, not reported),
shows an almost perfect monotonic increase of pupil-size with task difficulty (adapted from Hess, &
Polt, 1964).
24
The finding that pupil dilation increases with increasing multiplication difficulty has
since been replicated by other researchers (Bradshaw, 1968a; Klingner, 2010;
Marshall, 2002; Payne, Perry, & Harasymiw, 1968). Ahern and Beatty (1979, 1981)
showed that the TEPR was smaller for more intelligent college students than for less
intelligent counterparts when carrying out the same arithmetic problems. The same
was true for digit span and sentence comprehension, suggesting more efficient
information processing in the participants of higher psychometric intelligence, who
required fewer cognitive resources (Ahern, & Beatty, 1979; 1981). However, Beatty
and Lucero-Wagoner (2000) observe that this research did not address the possible
role of practice and over-learning in the higher intelligence group.
Much of Hess’ later research was around the concept of “emotional valence” and his
controversial aversion-constriction hypothesis: the pupil dilates to positive-affect
stimuli and constricts to negative-affect stimuli (Hess, 1965; Hess, & Polt, 1960).
Hess (1972) showed participants affectively loaded photographs of crippled children
or mutilation, and reported an initial dilation followed by constriction caused by the
“shock value” of the stimuli. Other researchers have used the idea of emotional
valence to infer people’s attitudes and preferences from their pupil-size. For
example, it was shown that participants’ pupils would dilate to images of preferred
political leaders and candidates, and constrict to undesirable images (Barlow, 1969;
1970; Clark, & Ertas 1975; Hess, 1965).
1.2.1.1.
Criticisms of Hess and Early Work
Hess has been criticised by independent researchers and reviewers (e.g., Dooley, &
Lehr, 1967; Goldwater, 1972; Hakerem, 1973; Janisse, 1973; Mueller, 1970; Peavler,
& McLaughlin, 1967; Woodmansee, 1966; Zuckerman, 1971) for repeatedly using
very small sample sizes, not reporting all potentially relevant results, not using
25
evidence to back up his interpretations or conclusions, citing “unpublished” pilot data ,
pooling his data across experiments, rarely presenting appropriate statistical analyses
(or none at all before 1966), imprecise methods, providing insufficient detail for
replication, and claiming to have discovered things that other people had already
published first and with better methodology (Janisse, 1977; Löwenstein, &
Loewenfeld, 1962). The fact that Hess’ (1975) literature review ignores any research
disagreeing with his own findings has made it difficult to draw any conclusions from
Hess’ bulky and ambiguous literature (Janisse, 1977). Describing Hess’ work as
“inane twaddle”, eminent and world-renowned pupil researcher Irene Loewenfeld
(1999) dismissed most of the emotion-based pupil research, not in objection to the
existence of the well-established psychosensory reflex, but because she considered
the research methodology and analyses to be flawed, requiring replication with more
appropriate and rigorous techniques. Loewenfeld also passionately denounced Hess’
aversion-constriction theory due to her intimate knowledge of pupil physiology.
Hess answered his critics by stating that aversion-constriction was present only for
“certain” people and stimuli, rather than all people and all aversive stimuli. The
aversion-constriction hypothesis has not been replicated by other researchers using
carefully controlled studies (e.g., White, & Maltzman, 1978; Paivio, & Simpson, 1966;
Schaefer, Ferguson, Klein, & Rawson, 1968). It has instead been criticised on
methodological grounds in several literature reviews, which suggest instead that
constriction is the result of habituation or decreased interest in the experiment
(arousal decrement) (e.g., Woodmansee, 1966), and that the brief stimulus-evoked
phasic dilations are more likely to be the result of cognitive activation, whereas
emotional arousal effects are longer lasting and more likely to influence tonic or
baseline pupil-size (e.g., Beatty, 1982; Beatty, & Lucero-Wagoner, 2000; Goldwater,
1972; Janisse, 1977). There is plenty of evidence supporting psychosensory dilation,
26
but no good evidence for aversion-constriction (Janisse, 1977). Pupillometry has
particular difficulties, some of which are unique among psychophysiological
techniques (Janisse, 1977). Critics suggest that Hess did not control the luminance
of his stimuli or his laboratory, and that “aversive” stimuli were brighter, causing
constriction of the pupil through the PLR rather than an emotional response (Janisse,
1977). Beatty (1972) suggests precautions researchers that can take to equate
stimulus brightness and contrast; however this is much harder when conducting
pupillometry studies in naturalistic situations (Wang, 2010).
Due to the PLR, it is imperative that researchers control both background illumination,
for example constant artificial lighting, and the global and local luminance of the
stimuli, especially when presented on a computer screen (Janisse, 1977). This
second point has been the source of considerable controversy (e.g., Hess, Beaver, &
Shrout, 1975; Janisse, 1973; Loewenfeld, 1966; Woodmansee, 1966). Pictorial and
photographic stimuli have been strongly criticised because they vary greatly in
luminance both globally between stimuli and locally within different regions of a single
stimulus, which can create artifactual pupil-size changes (e.g., Goldwater, 1972;
Janisse, 1977; Woodmansee, 1970; Zuckerman, 1971). The light and dark properties
of images can generate pupil-size changes of around the same magnitude as
psychosensory changes (Janisse, 1977) and should be taken into consideration when
employing visual scanning (Pomplun, & Sunkara, 2003; Van Orden, Limbert, Makeig,
& Jung, 2001), visual search (Backs, & Walrath, 1992; Porter, Trościanko, & Gilchrist,
2007) or photographs (Dabbs, & Milun, 1999; Libby, Lacey, & Lacey, 1973). It is not
only Hess who has been criticised for failing to control luminance – many of the early
studies using pictorial stimuli did not take this into consideration, inaccurately
reporting pupil constriction in response to the affective quality of the stimuli (e.g.,
Tanck, & Robbins, 1970). Peavler and McLaughlin (1967) showed four female, and
27
four male participants, pictures including three of clothed females and one of a nude
female. They found that participants’ pupils only dilated to the nude picture, and
constricted to the images of clothed females; however the clothed pictures were
brighter than the others.
Tryon (1975) surveyed twenty possible sources of variation and confounds in
pupillometry, many of which have been mentioned above. When participants are
presented with multiple stimuli in an experiment, adaptation or habituation can occur,
where the overall diameter of the pupil decreases, the magnitude of the TEPRs
decrease, and the speed of contraction increases (Löwenstein, & Loewenfeld, 1952;
Lehr, & Bergum, 1966; Tryon, 1975). Goodrich (1974; 1975) highlights the fact that
actual pupil-size is distorted by the cornea, which has a lens power equivalent to 3848 dioptres (Janisse, 1977). Conducting experiments using participants under the
age of 30 (Woodmansee, 1966), having stimuli 3-4m from participants (Hakerem, &
Sutton, 1964), and using relatively short trials, all help to reduce pupil-size variation
due to the near-vision reflex, which occurs when participants lose or change focus
due to age, fatigue or boredom (Janisse, 1977). Even when luminance is held
constant, other visual features such as spatial frequency, patterns and movement can
also influence pupil-size (Barbur, Wolf, & Lennie, 1998; Nakayama, Yasuike, &
Shimizu, 1990; Sahraie, & Barbur, 1997; Slooter, & van Norren, 1980; Ukai, 1985).
Due to a phenomenon known as the pupillomotor Purkinje effect, the pupil dilates
more in response to coloured stimuli (chromatic) than grey-scale stimuli (achromatic),
and constricts more to shorter wavelengths as luminance increases (Bouma, 1962;
Kohn, & Clynes, 1969). In addition, colours can have emotional meaning (Bouma,
1962; Kohn, & Clynes, 1969; Miller, 1967). For these reasons, it is highly likely that a
visual stimulus will produce a change in pupil-size in a cognitive pupillometry study.
28
Important and more rigorous programs investigating psychological influences on
pupil-size have been carried out by researchers such as Kahneman, and Paivio and
Simpson (Janisse, 1977).
1.2.2.
Arousal
The term “arousal” is vague, somewhat contentious, and is used by different
researchers to refer to a variety of constructs, such as emotions, sexual attraction, or
attention, each of which has a large and overlapping literature (see Staal, 2004;
Neiss, 1988). The most popularized pupillometry arousal research has involved sex,
racism or fear; for example, using pictures of nudes (Lawless, & Wake, 1968), people
of different races (Woodmansee, 1967), or fear of electric shock (Polt, 1970). Stimuli
were designed to elevate participants’ mental and physical arousal, stimulating the
sympathetic nervous system and the release of adrenaline into the blood stream,
leading to pupil dilation.
Many experiments (e.g., Aboyoun, & Dabbs, 1998; Bull, & Shead, 1979; Hess, 1965;
Hicks, Reaney, & Hill, 1967; White, & Maltzman, 1978) have shown that pupil-size is
linearly related to the level of sexual arousal (Janisse, 1977). Janisse (1977) asks
whether pupillary dilation accompanying “Don Juan[’s] … statements of undying
devotion” (p. 11) is due to sexual arousal (Zuckerman, 1971) or to the fact that he is
lying (Bradley, & Janisse, 1975, cited in Janisse, 1975). Using erotic and suspense
films, Bernick, Kling and Borowitz (1971) were able to show that pupil-size may
discriminate sexual arousal from more generalised arousal. However, as Janisse
(1977) has indicated, most of the pupillometric research around sexual arousal used
pictorial or video stimuli, varying in luminance (e.g., Hess, Seltzer, & Shlien, 1965;
Nunnally, Knott, Duchnowski, & Parker, 1967; Peavler, & McLaughlin, 1967; Scott,
Wells, Wood, & Morgan, 1967; Lawless, & Wake, 1969). An additional potential
29
confound is that at that time in North America nude pictures were more novel than
clothed pictures, and the pupil also responds to novelty (Andreassi, 2000).
Since Hess’ original studies, and contrary to the aversion-constriction hypothesis,
other researchers have since shown that the pupil dilates to emotional stimuli of both
positive and negative valence (Janisse, & Peavler, 1974; Stelmack, & Mandelzys,
1975). For example, Guinan (1967) showed that average pupil-size was larger for
high emotionality words than low emotionality words, particularly in the first 2.5s, and
suggested that the emotional content of the stimuli was causing autonomic arousal.
The same has been found for acoustic stimuli by Partala and Surakka (2003) who
played participants ten positive, ten negative and ten neutral sounds, finding that their
pupils dilated more to the positive and negative sounds compared to the neutral
sounds (0.2mm vs. 0.14mm). However other researchers have found a U-shaped
function of pupil-size where the pupil is larger for neutral stimuli than for slightly
positive and negative stimuli (e.g., Gunther, & Lussier, 1975, cited in Janisse, 1977).
Urry et al. (2006) presented negative and neutral affective images and asked
participants to intentionally increase their emotional response to the stimuli (imagining
the situation happening to them), decrease their response (viewing the situation as
fake), or simply attend to the stimuli in a control condition. They showed that actively
enhancing emotional responses increased initial and sustained pupil dilation
compared to decreasing emotional responses, and that in both emotional regulation
conditions pupil-size was larger than in the unregulated ‘control’ condition. However
there were multiple influences on pupil-size including cognitive effort, imagery and
emotional arousal, whereas the ‘attend’ condition did not involve a
cognitive/imagination task and so was insufficient to act as control. More concerning
30
is the fact that the stimuli were colour photographs, therefore potentially introducing
visual confounds.
Further evidence for the link between emotion and pupil-size comes from studies
showing that depressed participants (medicated and unmedicated) show different
pupil responses to emotional stimuli compared to non-depressed participants. For
example, Siegle, Steinhauer, Carter, Ramel and Thase (2003a) found increased and
sustained pupil dilation in depressives compared to controls for up to 30s when
identifying the valence of emotional words. The extent of pupil-size increase also
correlated with self-reported rumination, suggesting that it reflected sustained
elaboration of emotional information processing in depressive participants (Siegle et
al., 2003a; Siegle, Granholm, Ingram, & Matt, 2001; Siegle, Steinhauer, Carter, &
Thase, submitted; Siegle, Steinhauer, & Thase, 2004).
A substantial subset of the arousal literature has focussed on the relationship
between pupil-size and state and trait anxiety, which lead to larger pupil-sizes
consistent with sympathetic arousal; a larger resting pupil diameter in perpetually
anxious participants was noted by Bumke as early as 1903 (cited in Janisse, 1977).
The influence of anxiety on pupil-size is so consistent and dramatic that Paivio and
Simpson (1966), and Kahneman (1973) discussed the potential confounding factor of
anxiety in pupillometry experiments designed to measure cognitive effort or task
difficulty. Carver (1971) and Johnson (1971) proposed that progressive pupil dilation
with increasing task difficulty occurred as a result of increased anxiety and emotional
arousal due to anticipation, because in many studies (e.g., Kahneman, & Beatty,
1966) participants were informed in advance about the increasing difficulty. However
Peavler (1974) found difficulty-related increases in the absence of prior knowledge of
the task, while Simpson and Molloy (1971) showed that task-related and difficulty-
31
related increases in pupil-size persist even in anxious participants, who show larger
baseline pupil-sizes and response-related dilations than non-anxious participants
when expected to make an overt response. This is due to factors such as
performance anxiety, anticipation of making a response, explicitly making a decision,
anxiety about making a mistake, being judged, or receiving feedback based on a
response (Paivio, & Simpson, 1966; Simpson, 1969; Simpson, & Molloy, 1971;
Simpson, & Paivio, 1968). However anxious participants had qualitatively the same
cognitive load pupil-response curve as in other experiments manipulating task
difficulty (Kahneman, 1973; Paivio, 1973). Even in non-anxious participants, the
offering of incentives and penalties (and therefore the introduction of an element of
risk) increases TEPRs (Kahneman, Peavler, & Onuska, 1968b; Kahneman, &
Peavler, 1969).
Painful or ‘startling’ stimuli, such as heat or loud noise, cause dilation with a latency of
300-500ms (Janisse, 1977; Loewenfeld, 1958; Nunnally et al., 1967). Chapman,
Oka, Bradshaw, Jacobson and Donaldson (1999) provided “noxious” electrical
fingertip stimulation at four intensities, increasing from faint to almost unbearable and
found that peak pupil-size rose with intensity. Polt (1970) showed that even the threat
of an electric shock caused pupil dilation, but this could also have been the result of
increased cognitive effort to answer correctly under the threat of shock. However,
most of the arousal research conducted in the 1960s and 1970s was subject to the
same criticisms as Hess’ emotion research. There was no established common way
of reporting or analysing results, luminance confounds were common, and little
communication or shared learning of methodological issues occurred between
researchers (Janisse, 1977).
32
1.2.2.1.
Fatigue & Sleepiness
Pupil-size also reflects decreases in arousal, and pupil changes are associated with
the sleep-wake cycle (Löwenstein, & Loewenfeld, 1964). Diameter is largest when
individuals are well-rested, decreasing with fatigue and reaching its smallest diameter
prior to sleep (Foster, 1891; Loewenfeld, 1999). Fatigue also increases amplitude
and frequency of pupillary hippus, particularly in darkness (Löwenstein, &
Loewenfeld, 1952; 1964; Yoss, Moyer, & Hollenhorst, 1970). Pathological sleep
deprivation, such as in Excessive Daytime Sleepiness and Narcolepsy, causes
distinctive patterns of pupillary movements for narcoleptics compared to healthy
participants, and for treated versus untreated narcoleptics, both in light and darkness.
For example, narcoleptics have smaller baseline pupil-sizes and show less random
pupillary noise (as opposed to spontaneous changes caused by hippus) compared to
healthy participants (O’Neill, Oroujeh, Keegan, & Merritt, 1996; O’Neill, Oroujeh, &
Merritt, 1998; Pressman et al., 1984; Yoss, 1970; Yoss, Moyer, & Ogle, 1969).
Narcoleptic participants also demonstrate differences on cognitive pupil measures, for
example working memory overload occurs after storing fewer digits; however earlier
and faster pupil dilation compensates for smaller baseline pupil-size to arrive at a
peak dilation equivalent to controls (O’Neill, & Trick, 2001). Parkinson’s Disease is
often accompanied by arousal symptoms such as sleepiness, which are positively
correlated with pupillary unrest (Jain et al., 2011).
Using techniques developed by Yoss (1969; 1970), variation in fatigue-related pupil
changes have the potential to be useful in monitoring alertness in professions where
vigilance is critical, such as drivers (Recarte, & Nunes, 2003; Recarte et al., 2008;
Walzl, Hagen, & Prummer, 2007; Yoss, 1969), pilots (Dehais, Causse, & Pastor,
2008; Yoss, Moyer, Carter, & Evans, 1970), naval vessel operators (de Greef,
Lafeber, van Oostendorp, & Lindenberg, 2009), industrial and construction workers
33
(Geacintov, & Peavler, 1974; Wilhelm et al., 2010) doctors (Wilhelm, Widmann, Durst,
Heine, & Otto, 2009) and telephone operators (Geacintov, & Peavler, 1974). It also
points to a promising avenue of pupil research in monitoring attentiveness and/or
affect in real world scenarios through Human-Computer Interaction (HCI) technology
(e.g., Bailey, & Iqbal, 2008; Iqbal et al., 2004; Lin, Imamiya, & Mao, 2008; Oliveira,
Aula, & Russell, 2009; Rowe, Sibert, & Irwin, 1998). Pupil-size correlates with task
difficulty during individual stages of a task, and decreases during transitions between
tasks (Iqbal et al., 2004; Bailey, & Iqbal, 2008; Oliveira et al., 2009). Researchers
have used changes in pupil-size to identify transitions between subsections of tasks
when a user could be interrupted with the least amount of disruption (Bailey, Busbey,
& Iqbal, 2007; Iqbal, Adamczyk, Zheng, & Bailey, 2005). Another aim of HCI is to
enable computers to establish a user’s affective state (e.g., Barreto, Zhai, Rishe, &
Gao, 2007; Lanatà, Armato, Valenza, & Scilingo, 2011). Gao, Barreto and Adjouadi
(2010) developed an algorithm which was able to identify stressed states in
participants with 77.8% accuracy using pupil-size, Galvanic Skin Response (GSR)
and pulse measurements.
1.2.3.
“Cognitive Effort” – Kahneman
The majority of pupil research from the 1960s onwards concerns the relationship
between pupil-size and “cognitive effort”, a complex concept which some have
described as the proportion of available attentional resources assigned to a task (see
Cain, 2007; Moray, 1979). Other related concepts include “mental workload” and
“processing load” due to the fact that cognitive resources are limited (Miller, 1956),
and performing two tasks simultaneously usually leads to decreased performance on
one or both (Beatty, 1982b; Kahneman, 1973). A key player in cognitive effort
research was Nobel Prize winner Daniel Kahneman, whom Janisse (1977) credits
34
with the methodical exploration of the concepts of cognitive “loading” and “unloading”,
“processing”, “mental effort” and “rehearsal” within pupillometric research.
Kahneman (1973) argued that general “arousal” can be seen as a response to task
demands, and therefore “mental effort” is part of arousal. However, he emphasised
the necessity for researchers to distinguish between dilations caused by processing
load and dilations caused by other elements of arousal, such as muscle activity or
anxiety, in rigorously designed experiments with careful consideration of potential
confounds (Kahneman, & Wright, 1971). Pupil dilations are largest during task
performance, compared to before and after, and those corresponding to correct
responses are usually larger than those for failures (Kahneman, 1973). In addition,
studies have shown that although behavioural responses have a small effect on pupilsize, due to performance anxiety and muscular exertion (see section 1.2.5), dilation
occurs due to task performance even in the absence of an overt response, so these
other factors cannot account for the majority of the pupil-size change (Kahneman et
al., 1968b).
Kahneman (1973) describes a useful physiological index of mental effort as one
which is responsive to variation within-task, faithfully tracking changes in participant
effort whilst they carry out the task, between-tasks, identifying which tasks are more
difficult and therefore require more effort, and between-participants, showing that
different people invest different amounts of effort in a task. Beatty (1982b) says that
pupil responses meet all three of these criteria (Beatty, 1982b). An interesting
example of between-participant differences is that of “intelligence”; however
pupillometric studies show mixed findings, with some researchers finding no
differences in pupil-size for participants of high and low intelligence (e.g., Bernick,
Altman, & Mintz, 1972; Daly, 1966; Simpson, & Molloy, 1971), whilst others have
35
found larger pupil-sizes for participants with lower IQ (e.g., Ahern, & Beatty, 1979;
1981; Verney, Granholm, Marshall, Malcarne, & Saccuzzo, 2005), or, in contrast,
larger pupil-sizes for participants with higher IQ (e.g., Boersma, Wilton, Barham, &
Muir, 1970). Verney et al. (2005), for example, controlled stimulus brightness and
found that more intelligent participants performed better at a visual backwards
masking task, which evoked smaller pupil responses, compared to less intelligent
participants who performed worse (detected fewer targets, allocated more attentional
resources to non-target stimuli) and had larger pupil-sizes. If more intelligent
participants perform better on tasks because they find tasks easier than less
intelligent participants, it might be expected that their pupils would be smaller;
however if they perform better because they exert more effort, it might be expected
that their pupils would be larger than less intelligent participants (Janisse, 1977). The
question becomes whether pupil-size measures difficulty or effort, and whether effort
and perception of difficulty vary with intelligence (Janisse, 1977). As dilation varies
with effort, which varies with intelligence and cognitive resources, then individuals will
vary on the amount of effort required to carry out the same task and the amount of
effort-related dilation, but the two may be indistinguishable.
Kuc and Janisse (1967; 1976, cited by Janisse, 1977) compared successful and
unsuccessful digit span recall at a 50% difficulty threshold (50% of trials were correct
at that level of difficulty), and measured intelligence using a specific measure (Digit
Span Forward subscale of WAIS, 1955) following completion of the task. Using this
approach they found that pupil-size was larger for correct trials and suggested this
was due to increased mental effort leading to success, since task difficulty was held
constant, and subjective difficulty was associated with incorrect trials which had
smaller pupils. They also found no significant main effect of intelligence overall,
although there was a trend towards larger pupil-size during loading for the high
36
intelligence group, who also gave more correct answers (Janisse, 1977). Potential
confounds included the larger number of verbal responses associated with correct
trials than incorrect trials, and confidence, although Janisse (1977) concludes that
pupil-size changes are far more likely to reflect effort than difficulty and that using
“intelligence” as a factor in pupillometry studies is not straightforward. However,
using measures that are task-specific, rather than general, and making comparisons
of correct and incorrect trials may simplify interpretation of results and lead to
comparisons of more appropriate groups (Janisse, 1977). Daly (1966, cited by
Janisse, 1977) suggested that fluctuations in pupil-size, which decrease when
participants concentrate, might be a better measure of problem-solving efficiency than
maximum pupil-size (Kahneman, 1971; Janisse, 1977).
Other aspects of task performance such as accuracy, motivation and memory load
have all been linked to pupil-size changes (Janisse, 1977). Pupil-size has repeatedly
been shown to increase with increasing mental effort in a variety of tasks (see Table
1; for reviews see Beatty, 1982b; Beatty, & Lucero-Wagoner, 2000; Goldwater, 1972;
Janisse, 1977). Kahneman (1973) asks whether pupil-size changes that occur in
response to different types of task can be legitimately compared in terms of the
amount of effort expended. For example, measurement at a single time point may
not represent total effort for tasks involving sustained effort (Kahneman, 1973).
Tasks involving rapidly decaying short term memory, such as digit-span or pitchdiscrimination, or tasks requiring participants to respond quickly to stimuli, generate
both time-pressure and large pupil-sizes (Kahneman, 1973). ‘Difficulty’ manipulations
do not increase task difficulty equally between different types of requirement, for
example there is a larger gap between easy and difficult arithmetical problems, due to
storage and rehearsal, than between easy and difficult sentence comprehension
problems (Elshtain, & Schaefer, 1968; Kahneman, 1973).
37
Task
Key Pupil-Size Findings
Selected Authors
Signal detection
Pupil response only present when
participants reported flash, not when
identical flash went unreported; large
response to low probability or omitted
stimuli, similar to P300
Max increased monotonically as
comparison & reference tones became
closer in pitch, and therefore harder to
discriminate
Complex sentences induced larger
pupillary responses than simpler sentences
(0.25mm vs. 0.21mm) and increased
latency to peak by 116ms
Increases incrementally during loading,
returning to baseline during recall; peak
increased with string length; overload
occurs sooner in schizophrenic participants
Light: Hakerem, & Sutton,
1966; tones: Beatty, 1975
Pitch
discrimination
Sentence
comprehension
Digit recall
Dual task
performance
Incentive
Random motor
responses
Visual search
Spatial ability
Processing
speed
Deception
Larger for dual-task than single visual
search, but similar for single digit
transformation; during dual-task, errors
increase & decrease with pupil-size,
suggesting maximal processing capacity
reached
Larger for high than low incentives
larger when participant chose to move
than when instructed
Larger for more difficult searches;
increases as task progresses (memory
load?)
Increased with angular disparity when
judging irregular hexagons; greater
dilations for low than high spatial ability
participants
Increased at 75% & 100% max processing
speed capacity, constriction at 125%
Larger for guilty/lying than
innocent/truthful participants
Kahneman, & Beatty, 1967;
Schlemmer, Kulke, Kuchinke,
& Van Der Meer, 2005
Just, & Carpenter, 1993;
Wright, & Kahneman, 1971
Beatty, 1966; Beatty, &
Kahneman, 1966; Granholm
et al., 1997; Kahneman, &
Beatty, 1966; Simpson, &
Hale, 1969
Kahneman, Beatty, &
Pollack, 1967
Kahneman, & Peavler, 1969;
Kahneman, Peavler, &
Onuska, 1968
Simpson, & Hale, 1969
Pomplun, & Sunkara, 2003;
Porter, Trościanko, &
Gilchrist, 2007
Just, & Carpenter, 1995
Poock, 1973
Berrien, & Huntington, 1943;
Dionisio, Granholm, Hillix, &
Perrine, 2001
Table 1: Selection of tasks used to investigate the relationship between pupil-size and “mental effort”.
Some researchers have found larger pupil-sizes during tasks that intuitively feel ‘easy’
than for tasks we might consider more difficult. For example, dilations to pairedassociate recall are 4-6 times larger than during learning (Kahneman, & Peavler,
38
1969); and the pupil dilates more whilst retaining five digits for immediate recall,
which most participants perform with ‘ease’, than whilst attempting to listen to and
comprehend a long complex message (Carver, 1971). Large dilations also
accompany the prompted recall of over-learned personal information such as age or
phone number, which should be ‘easy’ to retrieve (Beatty, & Kahneman, 1966;
Kahneman, 1973; Schaefer et al., 1968). It may be misleading to conclude that more
effort is required for recall than for learning, or that recall is more ‘difficult’, when the
task demands are different (Kahneman, 1973).
Steinhauer et al. (2004) used a novel approach to look at the sympathetic and
parasympathetic contributions to pupil-size during sustained cognitive effort.
Participants performed an easy task (add 1) and difficult task (subtract 7) in normal
light and complete darkness. Their pupils dilated equally to both tasks in darkness,
but dilated more to the hard task in the light. Steinhauer et al. (2004) suggested that
in addition to sympathetic dilation, there was also parasympathetic inhibition of
constriction for the difficult task in the light, allowing larger dilations to occur. They
repeated the experiment having used eye-drops to selectively block the
parasympathetic sphincter muscle (tropicamide), or the sympathetic dilator muscles
(dapiprazole), or neither (placebo). Dilation was seen in each condition, however
effects of task demands and light condition on pupil-size, and the previously seen
interaction between the two, were only present when dapiprazole was used to block
the sympathetic dilator muscles. It was absent when tropicamide was used to block
the parasympathetic constrictions, suggesting that it is increased parasympathetic
inhibition of pupil constriction that leads to larger pupil dilation during more difficult
tasks, whereas sympathetic activity is less differentially affected (Steinhauer et al.,
2004). Direct cortical input, indirect cortico-thalamic-hypothalamic input and arousalrelated reticular pathway activity have all been linked to inhibition of the Edinger-
39
Westphal region through which the pupillary sphincter muscle is controlled (Bonvallet,
& Zbrozyna, 1963; Löwenstein, 1955).
1.2.3.1.
Signal Detection & Discrimination
One of the most straightforward cognitive tasks is asking a participant to indicate
when they detect a near-threshold sensory stimulus, for example, a weak flash of light
in a darkened room. Signal detection as applied to pupillometry was first reported by
Hakerem and Sutton (1966), who recorded tiny pupil responses (<0.1mm) to brief
flashes of light near the visual threshold followed by a tone. Participants were dark
adapted for 20 minutes prior to the experiment and asked to press a button after the
tone if they had seen the light (Hakerem, & Sutton, 1966). Button presses were
counterbalanced across experiments to control for response-related dilations, and
participants experienced pupil dilations to light stimuli that were too weak to evoke a
light-reflex. Interestingly the stimulus-evoked response was only present in trials
where participants reported seeing the flash, not in trials where an identical flash went
unreported. The relationship between pupil-size and conscious awareness is
returned to in the general discussion of Chapter 3, section 3.4.
In a standard auditory signal detection experiment Beatty and Wagoner (1975; 1976)
asked participants to detect 100ms 1kHz tones in a white noise background and
press one of four buttons (yes-certain, yes-uncertain, no-uncertain, and no-certain).
In a 2 (signal: present, absent) x 2 (decision: yes, no) x 2 (confidence: certain,
uncertain) design, they found that pupil-size was the same for the conditions where
the signal was absent. For the signal present conditions, pupil changes were related
to stimulus-response category, whereby yes-certain decisions evoked the largest
dilations, followed by yes-uncertain decisions, no-uncertain decisions, and no-certain
decisions. Beatty and Wagoner concluded that pupil-size changes were only related
40
to outcome in the presence of a signal. The findings support a cognitive or effortbased, rather than anxiety or emotion-based, interpretation of pupillary dilation since
the two ‘uncertain’ types of trial evoked an intermediate sized response (Beatty, &
Lucero-Wagoner, 2000; Janisse, 1977).
Discriminating between two signals, presented either simultaneously or in succession,
is more complex and resource intensive than signal detection, leading to larger
overall pupil dilation in signal discrimination than in signal detection (Beatty, & LuceroWagoner, 2000). For example, maximum pupil-size increases monotonically as a
comparison and reference tone become closer in pitch, making them harder to
discriminate (Kahneman, & Beatty, 1967).
1.2.3.2.
Working Memory
In a series of paced auditory serial recall/digit span tasks Kahneman and Beatty
(1966) presented participants with digit strings of 3-7 items at a rate of one per
second (loading phase), asking them to recall the items in the same order at the
same rate (unloading phase). During loading pupil-size increased incrementally with
each successive item, due to increasing rehearsal in working memory (Baddeley, &
Hitch, 1974; Kahneman, & Beatty, 1966), and peak size increased with increasing
string length (0.1mm vs. 0.55mm for 3 and 7 digits). This was followed in the
unloading phase by a brief dilation (approximately one second before unloading
began), then decrements as items were recalled, due to cessation of rehearsal
(Johnson, 1971), until pupil-size returned to baseline at the end of the trial.
Kahneman and Beatty (1966) replicated the loading and unloading effect by asking
participants to add 1 to each digit (e.g., hear 3-9-1-6, report 4-0-2-7), and in a
different task, with 4-item strings of words (high-frequency monosyllabic nouns), and
found that the increased task difficulty lead to larger pupil dilations, particularly for the
41
digit transformation. Kahneman and Beatty’s (1966) classic experiments have been
replicated using modern remote eye-tracking techniques and the same results have
been found (Klingner, 2010).
Peavler (1974) investigated the effects of saturating central processing resources by
using strings of up to 13 digits in a paced recall task and found that “overloading” did
not lead to further increases in pupil-size. After increases in response to the
participants’ maximal information processing capacity (Miller, 1956; Poock, 1973) of
7-8 digits, the pupil-size remained just below the maximum until recall (Peaver, 1974).
Other investigators have found that pupil-size starts to decline once demands outstrip
available resources, particularly in populations with cognitive impairments such as in
schizophrenia (Granholm, Asarnow, Sarkin, & Dykes, 1996; Granholm, Morris,
Sarkin, Asarnow, & Jeste, 1997), suggesting that differences arose from variation in
instructions – Peavler (1974) asked participants to try their best, so they may have
actively maintained their maximum digit strings through rehearsal (Granholm et al.,
1996; 1997). These studies also contribute compelling evidence that it is cognitive,
rather than emotional or anxiety-related changes, which cause the pupil-size
increases as failure-related emotional responses should cause dilation to peak
following overload (Beatty, & Lucero-Wagoner, 2000).
1.2.3.3.
Visual Search
Increasing pupil-size has been demonstrated in visual search, which involves working
memory (e.g., Porter et al., 2007). By comparing counting and searching within the
same arrays, Porter et al. (2007) found that counting involved sustained pupil dilation
throughout the task, whereas search involved little dilation to begin with and
increasing dilation throughout the task as spatial memory demands increased.
Manipulating search difficulty by increasing the number and variety of distracters, they
42
found that more difficult searches, and larger or more complex spatial arrays,
produced larger pupil-sizes than easier searchers and smaller or less complex arrays.
1.2.3.4.
Effort and the Red Pupillary Reflex
The fundus or “red” reflex occurs when light shone into the eye is reflected by the
ocular fundus, and is seen most commonly in everyday life as “red eye” in
photographs. Kruger (1975; 1977a; 1977b) found a 10% increase in red reflex
luminance accompanying increases in arithmetic task difficulty, which he attributed to
an increase in accommodation (Kruger, 1977a). He confirmed this by using
cycloplegic drugs (which paralyse the ciliary muscles and prevent accommodation),
and found that luminance was no longer related to task difficulty (Kruger, 1977b).
However, cognitive effort-related pupil-size increase is a potential confound as this
would cause a brighter red reflex by allowing more light into the eye. Cycloplegic
drugs cause mydriasis as well as accommodative paralysis, which would also explain
the loss of relationship between red reflex and task difficulty; other researchers have
found that accommodative response remains constant or even decreases as
cognitive processing increases (Jainta, Hoormann, & Jaschinski, 2008; Davies,
Wolffsohn, & Gilmartin, 2005).
1.2.3.5.
Language & Comprehension
There is a robust literature linking pupil-size to variations in the complexity of
language processing, such as speech perception (Zekveld, Kramer, & Festen, 2011),
language translation (Hyönä, Tommola, & Alaja, 1995), letter perception (Beatty, &
Wagoner, 1978), syntax (Schluroff, 1982) and semantic processing (Hyönä et al.,
1995). For example, increased grammatical complexity and ambiguity increases
processing load, short-term memory demands and decision time, producing larger
peak dilations and longer peak latencies than simpler sentences (Just, & Carpenter,
43
1993; Schluroff, 1982; Schluroff et al., 1986; Stanners, Headley, & Clark, 1972;
Wright, & Kahneman, 1971). Beatty and Wagoner (1978) asked participants to
decide whether letter pairs were physically the same (AA, aa), phonetically the same
(Aa), or from the same category (vowels: Ae). Larger pupil-sizes were found to
phonetic pairs than physical pairs, and larger pupil-sizes to vowel pairs than other
pairs, concluding that pupil-size reflected the amount of processing needed to decide
whether letters were the same (Beatty, & Lucero-Wagoner, 2000). However,
researchers have found that changes in cognitive load associated with reading texts
of varying difficulty is not reflected in pupil-size (Schultheis, & Jameson, 2004; Iqbal,
Zheng, & Bailey, 2004), perhaps due to minor differences between easy and difficulty
reading tasks.
1.2.3.6.
Imagery
Paivio and Simpson (1966) asked participants to create images in their mind of
concrete and abstract nouns. Measuring manually from projected 16mm film strip,
they found that when an overt response was required, abstract words evoked
significantly larger pupil-sizes with longer latency and duration compared to concrete
words, and that both abstract and concrete words evoked larger pupil-sizes than
control slides of equivalent brightness. This finding has been replicated several times
(Colman, & Paivio, 1969; Paivio, & Simpson, 1968; Simpson, & Climan, 1971;
Simpson, Molloy, Hale, & Climan, 1968; Simpson, & Paivio, 1966; 1968) and
interpreted as abstract words being more difficult to imagine, and therefore taking
longer and requiring more cognitive effort (Kahneman, 1973).
1.2.3.7.
Lie Detection
Despite being under autonomic control, relatively little research has been conducted
into the activity of the pupil during lying (Janisse, 1977). The majority of studies have
44
concluded that deception-related changes in pupil-size are related to additional
cognitive effort involved in fabricating answers (e.g., Dionisio, Granholm, Hillix, &
Perrine, 2001) and/or increased anxiety related to concealing information and the fear
of being caught (e.g., Berrien, & Huntington, 1943; Bradley, & Janisse, 1975, cited in
Janisse, 1977). Several studies have involved participants committing a mock crime
and hiding it from the experimenter (e.g., Lubow, & Fein, 1996), and only a couple
have investigated pupil-size whilst asking participants to lie in a memory paradigm
(Dionisio et al., 2001; Heaver, & Hutton, 2011). Detailed discussion of pupillometry
studies involving lie-detection can be found in Chapter 5.
1.2.3.8.
Auditory Stimuli
Importantly, cognitive influences on pupil-size are not dependant on the visual
modality, a number of studies have shown that these effects also occur for auditory
stimuli. Klingner (2010) replicated three classic cognitive pupillometry studies (digit
span, mental arithmetic, and vigilance) using the same stimuli presented auditorally
and visually under conditions of equivalent brightness and contrast. In all three, pupilsize showed qualitatively the same pattern of dilations with increasing task difficulty in
both modalities, however pupil-size was on average almost twice as large for auditory
stimuli as for visual stimuli (Klingner, 2010). The qualitative similarity in results
suggests that pupil responses reflect post-perception task demands, and that there is
increased processing load for auditory tasks, whereas visual presentation facilitates
comprehension and processing (Klingner, 2010).
1.2.4.
Attention
“Attention” remains one of the most enigmatic concepts in cognitive psychophysiology
(Beatty, & Lucero-Wagoner, 2000) and definitions made over 100 years ago remain
45
relevant today: “the taking possession by the mind, in clear and vivid form, of one out
of what seem several simultaneously possible objects or trains of thought.
Focalisation, concentration, of consciousness are of its essence” (James, 1890, pp.
403-4). The influence of attention on pupil-size is “widely accepted” (Janisse, 1977,
p. 2), and changes in pupil-size appear to closely chart central attentional resources
(Beatty, & Lucero-Wagoner, 2000). Researchers have related pupillary dilation
observations with attention-related concepts such as “interest” and “novelty” as part of
the autonomic orienting response to salient stimuli (Lynn, 1966; Nieuwenhuis et al.,
2011a; Pavlov, 1927; Sokolov, 1963). Libby et al. (1973) noted that interesting or
attention-getting pictures evoke the largest pupil-sizes, and Pratt (1970) reported that
more complex and interesting (less predictable) shapes produce larger pupil dilations.
Over repeated exposure, the effect of novel, unfamiliar or unpredictable stimuli
lessens (Pratt, 1970).
However, as discussed in section 1.2.1.1, pictorial stimuli confound pupil-size studies.
An alternative method of assessing attention effects without visual confounds is to
use auditory stimuli. Beatty (1988) recorded pupil-size whilst participants were
instructed to detect target tones presented to one ear and to ignore all tones
presented to the other ear. Similar to findings in the ERP literature (e.g., Hink, Van
Voorhis, & Hillyard, 1977), he found no detectable increase in pupil-size to tones in
the unattended channel, tiny (0.015mm) dilations to non-targets in the attended
channel and large stimulus-evoked dilations to targets (Beatty, 1988). In an auditory
vigilance task Beatty (1982a) asked participants to sustain attention to a task and
detect infrequent target tones amongst non-target tones for 48 minutes. As the task
progressed, behavioural and pupil-size data demonstrated a vigilance decrement, as
participants became less sensitive to targets, more conservative in their judgement
criteria, and experienced smaller stimulus-evoked dilations.
46
When focused attention increases, sympathetic activity increases (sympathetic
dominance) and parasympathetic activity decreases, however, there is often only a
small correlation between markers of sympathetic dominance, such as pupil diameter,
GSR, blood pressure and heart rate (Kahneman, 1973). Autonomic markers often
respond differently to stimuli, for example the largest pupil-sizes and slowest heart
rates may occur to the most interesting pictures in a selection, referred to by Lacey
(1967; Libby et al., 1973) as “directional fractionation”.
1.2.4.1.
Locus Coeruleus
As discussed in section 1.1.2.7, the activity of the Locus Coeruleus (LC) has been
demonstrated to change with pupil-size, level of arousal and attention to a task. The
LC is thought to be involved in the flexible allocation of cognitive control as a function
of internal and external requirements (Cohen, Botvinick, & Carter, 2000; Cohen,
Aston-Jones, & Gilzenrat, 2004). The Adaptive Gain Theory (AGT) of LC-NE function
(Aston-Jones, & Cohen, 2005) states that phasic LC activity optimises task
engagement for exploitation of a known source of reward, and that tonic activity
promotes disengagement to allow exploration for new sources. This exploit-explore
balance is important in the cognitive control and adaptive regulation of behaviour, and
the model has enabled the successful prediction of pupil response in tasks involving
target detection, conflict and reward (Gilzenrat, 2006). Very recently baseline pupilsize has been shown to correspond to performance dynamics and task engagement
in humans, whereby increases in prestimulus pupil-size are associated with
decreasing task engagement, and decreases in prestimulus pupil-size correspond
linearly with increases in task-evoked dilations (phasic responses) and task
engagement (Gilzenrat et al., 2010; Jepma, & Nieuwenhuis, 2011).
47
Gilzenrat et al. (2010) used an auditory target detection task to show that on a trialby-trial basis baseline pupil-size accurately predicted engagement and
disengagement from task performance according to AGT. Manipulation of conflict
and reward within tasks caused participants to either persist with the task or choose
to give up, and this behaviour was preceded by increases and decreases in baseline
pupil-size respectively. Smaller baseline pupils were inversely followed by larger
stimulus-evoked dilations, as described by the phasic mode of the LC. Gilzenrat et al.
(2010) provided the first evidence of this sort, and felt that pupil-size could be used to
index control state, even though in this experiment LC activity was inferred from pupil
dynamics rather than direct recording of neuronal responses.
Nieuwenhuis, Gilzenrat, Holmes and Cohen (2005b) extended the LC-NE model to
describe the “attentional blink” (Raymond, Shapiro, & Arnell, 1992), a temporary
inability to perceive the second target when two targets are presented temporally very
close together. The authors proposed that this impairment results from the neural
refractory period of the LC-NE system as observed in monkeys after a phasic
response to a target (e.g., Usher et al., 1999). The P300 ERP component is also
suppressed for the second target under the same experimental conditions,
particularly when the P300 to the first target is large, leading Nieuwenhuis et al.
(2005b) to suggest that it also reflects LC-NE phasic activity,
1.2.4.2.
Blinks
As discussed in section 1.1.2.3, after a blink the pupil briefly contracts before
redilating (DeLaunay, 1949). A number of studies have now shown a link between
cognition and spontaneous blinks, which has lead to some researchers excluding
post-blink periods from their analyses (e.g., Hupé et al., 2009). Studies have shown
longer blink suppression and higher blink rates accompanying high cognitive load
48
than low cognitive load, and shorter blink latencies for positive compared to negative
stimuli (e.g., Ohira, 1996; Ohira, Winton, & Oyama, 1998; Recarte, Perez, Conchillo,
& Nunes, 2008). Siegle, Ichikawa and Steinhauer (2008) suggest that because they
occur under cortical control, blinks distinguish separate stages of information
processing, relating to changes in cognitive states or resource allocation, and
regulating the time in which information is acquired (Fogarty, & Stern, 1989; Stern,
Walrath, & Goldstein, 1984). Siegle et al. (2008) proposed that blinks complement
pupil-size changes in the analysis of cognitive tasks.
1.2.4.3.
Decision Making & Uncertainty
The act of making a simple decision in a laboratory setting leads to a significant
increase in pupil-size compared to being instructed (Simpson, & Hale, 1969). Most
pupillometric studies require participants to respond, therefore unless the variable
under investigation is ‘choice’, task instructions need to be very clear so as to remove
decisions and uncertainty from requirements (Janisse, 1977). As stimulus uncertainty
increases, baseline pupil-size increases and phasic pupil amplitude decreases
(Bradshaw, 1968b; 1969; Levine, 1969, cited in Hakerem, 1974; Pratt, 1970). With
the exception of Simpson (1969), the literature shows a sympathetic-like dilation in
anticipation of parts of each trial, such as onsets, offsets, changing stimuli, or
response prompts (Bradshaw, 1969; Janisse, 1977; Nunnally et al., 1967).
1.2.5.
Physical Effort
The pupil dilates in response to physical effort (Foster, 1891; Hakerem, & Sutton,
1966; Hupé et al., 2009; Nunnally et al., 1967). Nunnally et al. (1967) asked
participants to lift weights of increasing then decreasing mass for ten seconds each,
with a ten second rest in between, whilst measuring pupil-size. They found that mean
49
pupil-size was larger during the lifting period than the following rest period, and that
as the mass of the weight increased and decreased so did pupil-size. Nunnally et al.
(1967) reported pilot results showing the same relationship between pupil-size and
fist clenching exercises, and concluded that that pupil-size is related to muscle
tension (see Figure 1-8).
Figure 1-8: Mean pupil-size (magnified x17.5) in relation to 10 second lifting periods and intervening
10 second resting periods (adapted from Nunnally, Knott, Duchnowski, & Parker, 1967).
Participants do not need to lift large weights for motor activity to influence their pupilsize, and, in addition to task demands and performance anxiety, an overt response
may induce pupil-size changes through muscle tension from moving the mouth or a
finger (Paivio, & Simpson, 1966; Richer, & Beatty, 1985; Simpson, 1969; Simpson, &
Molloy, 1971; Simpson, & Paivio, 1968). Hupé et al. (2009) showed participants
ambiguous moving stimuli and asked them to press a button when their bistable
percept changed. They found that 70% of the 5% average change in pupil-size area
was due to making the motor response.
50
Despite the fact that overt responses cannot account for all pupil-size changes, some
researchers have shown that pupil responses only occur when participants are
required to make an overt response to the task (Hakerem, & Sutton, 1966; Paivio, &
Simpson, 1966; Simpson, & Paivio, 1968). For example, in Hakerem and Sutton’s
(1966) signal detection task, when participants were not required to indicate when
they had detected a flash, pupil-size did not increase to any of the flashes, whereas
when a decision was required, pupil-size increased when flashes were detected.
Counterbalancing the conditions in which a response was required ensured that the
difference was not due purely to motor effort, and Hakerem and Sutton suggested the
results reflected a higher level of vigilance in the reporting conditions. It is not difficult
to imagine that participants may be less engaged with the task if their performance is
not being externally assessed. However, Simpson and Paivio (1966) used an
imagery task to show that larger pupil dilations occurred to abstract words than
concrete words even without a motor response. As this is in contrast with findings by
the same authors that the difference only occurred when a response was required
(Paivio, & Simpson, 1966; Simpson, & Paivio, 1968), it highlights the importance and
likely contribution of task instructions and continued participant motivation.
1.2.6.
Pupil-Size and Concurrent Psychophysiological
Measures
Pupil measures may be especially suited to experimental studies because like other
psychophysiological markers, pupil-size can covertly and continuously monitor the
time course of cognitive processes. Its millisecond time resolution enables
researchers to observe the dynamics of processing demands with minimal latency,
and data collection is not dependent on participant response (Hyönä et al., 1995;
Kramer, 1991). To give a fuller picture, changes in pupil-size can be combined with
51
other psychophysiological markers of cognitive effort such as heart rate variance (Lin
et al., 2008; Rowe et al., 1998), functional Magnetic Resonance Imaging (fMRI; Just,
Carpenter, & Miyake, 2003), Electroencephalography (EEG; Schacter, 1977), EventRelated Potentials (ERPs; Just et al., 2003; Kok, 1997), plasma 17hydroxycorticosteroid levels (Bernick et al., 1971), Positron Emission Tomography
(PET; Just, Carpenter, Keller, Eddy, & Thulborn, 1996), GSR (Colman, & Paivio,
1969; Kahneman et al., 1969), and other eye-tracking measures such as blink rate
(Recarte et al., 2008; Siegle et al., 2008), or fixation time (de Greef et al., 2009).
As it is under autonomic control, the pupil demonstrates characteristics such as
arousal decrement (Woodmansee, 1966), parasympathetic rebound (Rubin, 1964,
cited by Janisse, 1964), and habituation (Löwenstein, & Loewenfeld, 1962). Wilder’s
(1958) Law of Initial Values (LIV) states that for physiological responses, as the initial
value gets higher so the response becomes smaller for enhancing stimuli and larger
for reducing stimuli (Jin, 1992). However, despite sympathetic and parasympathetic
input, at times the pupil does not respond in line with other peripheral autonomic
efferents, such as heart rate. Examples of this “directional fractionation” (Lacey,
1959; 1967) include the effects of modafinil, which increases noradrenergic input to
the pupil without influencing heart rate and salivation (Hou et al., 2005). Libby et al.
(1973) found that changes in heart-rate and pupil-size share only 19% of their
variance, and that there is great individual variation in correlation between the two
variables (between -0.52 and 0.39, average 0.35). Pupil-size has a greater
correlation with GSR of 0.30-0.50 (Colman, & Paivio, 1969; Scott et al., 1967), but
Colman and Paivio (1969) suggest that pupil-size “may be a more sensitive peripheral
response than GSR during cognitive tasks” (p. 296), for example, differentiating
concrete from abstract imagery, and easy from difficult paired-associate learning
(Colman, & Paivio, 1969; McElvain, 1970). It is difficult to establish from these
52
observations how much of this variation is due to pupil-size and GSR indexing related
but different processes that sometimes co-occur, leading to a higher correlation under
certain conditions, or whether they index the same processes but that pupil-size is
just more sensitive.
In a paced serial recall/digit span task Kahneman et al. (1969) found that increases in
GSR and pupil-size tracked increases in task difficulty, whereas heart rate decreased
as difficulty increased. Of the three measures, pupil-size, measured with a “highspeed” (1Hz) infrared film camera, was the most consistent (Janisse, 1977). A
negative correlation between pupil-size and heart-rate was also found by Kuc and
Janisse (1967, cited by Janisse, 1977) in participants performing digit span under
stress (r = -.55), replicating previous findings (Clark, 1975; Libby et al., 1973),
whereas consistent with the views of Löwenstein and Loewenfeld (1962) there was a
minimal positive relationship under low-stress conditions (0.14). This was because
although pupil-size increased more during loading, overall pupil-size under both
conditions was the same, yet heart-rate was faster throughout the high-stress
condition than the low-stress condition. Additionally, the authors concluded that pupilsize reflected participant intelligence, correctness of answer and cognitive aspects of
the task, whereas heart rate reflected emotional aspects (Janisse, 1977). The
pupillary system is therefore a sensitive, specific and comparatively low-noise
measure of psychophysiological changes (Beatty, & Lucero-Wagoner, 2000).
A growing number of studies have combined pupillometry with neuroimaging
techniques such as ERP or fMRI (e.g., Brown, Kinderman, Siegle, Granholm, Wong,
& Buxton, 1999; Conway, Jones, DeBruine, Little, & Sahraie, 2008; Friedman,
Hakerem, Sutton, & Fleiss, 1973; Hawkes, & Stow, 1981; Just, & Carpenter, 1993;
Just et al., 1996; 2003; Kuipers, & Thierry, 2011; Ledoux et al., 2010; Muller-Jensen,
53
& Hagenah, 1979; Murphy et al., 2011; Siegle, Steinhauer, Stenger, Konecky, &
Carter, 2003b; Steinhauer, 1982; Van Droof et al., 2010). Techniques can be
compared if researchers use the same paradigms (e.g., Vilberg, & Rugg, 2008), and
pupil-size is easily acquired alongside other techniques, providing complementary
data about the processes under investigation (Just et al., 2003).
Siegle et al. (2003b) recorded concurrent pupil-size and fMRI data whilst participants
carried out a digit sorting task, and found activity in the middle frontal gyrus had a
similar time-course to pupil-size. By recording pupil data on its own outside the fMRI
scanner using the same participants, the researchers were able to show that the
parametric increase with task difficulty was the same in both contexts. They were
also able to use the individual variation in pupil-size to model the activity in the middle
frontal gyrus to improve sensitivity and specificity, showing that pupil-size accurately
reflects task-related cognitive activity as measured by fMRI (Siegle et al., 2003b).
However, because pupil-size is measured continuously and tracks changes in task
requirements with low latency (0.1-0.5s) it may be a more consistent measure of
general cognitive effort than measures such as GSR (Kramer, 1991).
Just and colleagues (e.g., 1993; 1996; 2003) have investigated psychophysiological
indices of working memory load for two decades, using a wide range of executive
processing, language processing, spatial and memory task whilst recording ERPs,
fMRI and pupillometry data. Just et al. (2003) reviewed the literature and showed that
similarities exist between all three measures, for example, response magnitude
during tasks. They concluded that when being used to measure the same paradigms
the techniques tap the same common process, “capacity utilisation” or cognitive load.
54
1.2.7.
Memory
Although working memory has received considerable attention (see section 1.2.3.1),
very few researchers have studied the influence of Long-Term Memory (LTM)
retrieval such as recognition on pupil-size. In the next section current models of
recognition memory will be outlined, before returning to a discussion of pupil-size and
recognition memory in section 1.3.2.3.
1.3. Recognition Memory
1.3.1.
Models of Recognition Memory
Almost everything we do relies, more or less, on our capacity to learn, store and
retrieve information from memory. However, the precise structure of human memory
and the underlying cognitive processes are so complex, that even after centuries of
research using increasingly sophisticated methods, the current literature still contains
ongoing debate, and reports divided opinion on how we should model this
fundamental function (e.g., Diana, Reder, Arndt, & Park, 2006; Dunn, 2004;
Macmillan, & Rotello, 2006; Malmberg, Holden, & Shiffrin, 2004; Murdock, 2006;
Park, Reder, & Dickison, 2005; Parks, & Yonelinas, 2007; Rotello, Macmillan, &
Reeder, 2004; Tulving, 1985b; Wixted, 2007b; Wixted, & Stretch, 2004; Yonelinas,
2002). The remainder of this chapter is primarily concerned with the retrieval of
information from long term memory (Tulving, 1983; see Figure 1-9) in simple item
recognition, rather than associative or plurality recognition, which are thought to rely
on different underlying processes (Westerman, 2001).
55
Central
executive
Visuospatial
Episodic
Phonological
sketchpad
buffer
loop
Visual
Long term
semantics
Language
memory
Figure 1-9: Human memory systems (from Baddeley, 2000).
Recognition is the awareness that something has been encountered before, and
models of recognition memory are broadly divided into single- (e.g., Wixted, &
Stretch, 2004) and dual-process models (e.g., Diana et al., 2006; Yonelinas, 2002).
Current dual-process models of recognition memory (e.g., Atkinson, & Juola, 1973;
1974; Mandler, 1980; Tulving, 1985a; Yonelinas, 1994; 1997; 1999; Yonelinas, &
Jacoby, 1996) assume that the recognition of previously encountered faces, objects
or words occurs due to two independent mnemonic processes – recollection, a slow
and effortful conscious process where specific contextual information concerning the
original learning experience is retrieved from episodic memory, and familiarity, a
relatively rapid and automatic process which provides a context-free sense that an
item is known but without detailing why (see Yonelinas, 2002, for a review).
In contrast, single-process models (e.g., Donaldson, 1996; Gillund, & Shiffrin, 1984)
propose that the qualitatively different experiences of recollection and familiarity
originate from a single common neurocognitive process (Squire, Wixted, & Clark,
56
2007; Wixted, & Stretch, 2004). Therefore much of the debate in the literature
centres upon whether recognition is based on one (familiarity) or two (familiarity and
recollection) variables, whether the variable(s) draw on multiple sources of
information (global, specific), and if a weighted summed source should be considered
a single or multiple sources (e.g., Rotello et al., 2004, STREAK model) (Malmberg,
2008).
When recognition memory is explicitly tested under experiment conditions,
participants are typically presented with a set of learning items (which may be written
or auditorally presented words, images, faces, tones), followed immediately, or after a
delay, by a second set of items containing the original learning items (old) and items
that were not presented at learning (new), in the same modality or a different
modality. There are different recognition tasks that may be employed, for example: 1)
old/new task where participants just have to state whether a presented stimulus has
been seen before in the experimental context; 2) rating task where participants also
say how confident they are that they have (not) seen the stimulus before, according to
a scale; 3) a two-alternative forced-choice task where participants have to decide
which of two stimuli they saw before (see Malmberg, 2010).
Familiarity is proposed to be a continuous signal and stimuli such as words, pictures
and everyday objects will already be associated with a certain amount of familiarity.
However, a larger degree of familiarity is gained from exposure in the study phase of
an experiment, allowing them to still be useful stimuli in standard old-new recognition
tasks. When previously studied items (old) are intermixed with items not presented
during learning (new), participants are typically able to correctly identify at least 70%
of old stimuli (hits) and 80% of new stimuli (correct rejections) (Achilles, 1920;
Yonelinas, 1994). More new items are correctly identified than old items, and there
57
are usually more misses (old items identified as new) than false alarms (new items
identified as old) (Achilles, 1920). In contrast to familiarity, recollection is suggested
to be an all-or-nothing threshold retrieval process, involving the recovery of episodic
detail from the original learning instance (Yonelinas, 1994). Supposedly, in a
recognition task, new items may evoke familiarity but they won’t generate recollection,
whereas old items elicit a stronger sense of familiarity, plus recollection, and hence
result in a positive recognition decision.
Frequently, participants may be asked to decide whether they “remember” seeing an
old stimulus at learning (conscious recollection leading to an R response), or whether
they just “know” that it is old (feeling of familiarity leading to a K response), an
introspective report known as a remember-know or R-K judgement (Gardiner, 1988;
Tulving, 1985a). Source memory has been used as a more objective measure of
whether the recognition decision is based on recollection or not, for example
Yonelinas (1994; 2001b) asked participants which of two learned word lists each old
item appeared on. If an item was identified as being on the correct list, and was given
a high confidence rating, it was assumed to have been recollected. Researchers
have used various terms to label the strength variables underlying recognition, such
as global and specific (Rotello et al., 2004), item and associative (Murdock, 2006),
and semantic and episodic (Reder et al., 2000), but this thesis will generally use
familiarity and recollective strength (Wixted, & Stretch, 2004).
1.3.1.1.
Single-Process Models
Widespread interest in recognition memory did not occur until the “cognitive
revolution” in the 1960s because recognition was considered simpler and more
straightforward than recall, partly due to its higher accuracy and perceived ease
(Malmberg, 2010). Global memory models endeavour to account for task
58
performance in all conditions under a unitary theoretical framework (Malmberg, 2010).
However, early models of recognition memory were signal detection measurement
models (e.g., Banks, 1970; Bernbach, 1967; Kintsch, 1967; Lockhart, & Murdock,
1970), where recognition was based on the comparison of a single continuous
variable, supposed to be memory strength or familiarity, to a criterion value.
Judgement criteria are subjective, with conservative criteria generating fewer false
alarms, but also fewer correct hits, and an individual’s criterion may be modified by
accumulating positive and negative feedback (cf., random walk theory; Ratcliff, 1978;
Ratcliff, & Murdock, 1976; Murdock, 1985). Recognition occurs because old items
are more familiar than new items, however, measurement models do not explain how
stimulus familiarity is generated (Malmberg, 2008; 2010).
In the 1980s global matching process models were developed to explain the
generation of familiarity in signal detection models of recognition (Clark, & Gronlund,
1996; Gillund, & Shiffrin, 1984; Hintzman, 1988; Humphreys, Bain, & Pike, 1989;
Murdock, 1982; single-process recollection models are relatively rare, except
Yonelinas, 1999; Diller, Nobel, & Shiffrin, 2001). Global matching models state that
familiarity during a recognition test results from an assessment of the similarity
between a stimulus and all information held in memory relating to the learning phase
(Gillund, & Shiffrin, 1984; Hintzman, 1988; Humphreys et al., 1989; Murdock, 1982;
Norman, & O’Reilly, 2003; Shiffrin, & Steyvers, 1997). The recognition decision is still
based on a continuous variable, familiarity, which is generated by matching a retrieval
cue (a transient representation of the stimulus) against the large number of traces in
memory (Malmberg, 2008). The more similar a retrieval cue is to traces in memory,
the more familiar it will ‘feel’, therefore as old items have been seen before in that
context, and will closely resemble at least one trace, they will seem more familiar than
new items (Malmberg, 2008).
59
Despite the benefits of being global models, which make few assumptions, global
matching models in their present state were unable to account for phenomena such
as list-length and -strength interference effects (Ratcliff, Clark, & Shiffrin, 1990),
Receiver Operating Characteristic (ROC) curves (Ratcliff, Sheu, & Gronlund, 1992)
and mirror effects (Glanzer, & Adams, 1985; Malmberg, 2008). In response to these
challenges, new global matching models were developed with a Bayesian approach
which assumes that memory systems have evolved to be optimal and adaptive, and
aims to achieve maximal accuracy on the basis of the available information. For
example, the Retrieving Effectively from Memory model (REM; Shiffrin, & Steyvers,
1997; 1998), the Theory Of Distributed Associative Memory model (TODAM;
Murdock, 1997; 2006), the Subjective Likelihood Model (SliM; McClelland, &
Chappell, 1998) and the Bind-Cue-And-Decide Memory model (BCDMEM, Dennis, &
Humphreys, 2001). Familiarity computations strengthen the likeness between
retrieval cues and their trace, whilst lessening the likeness with other memory traces,
known as differentiation (Criss, 2006).
Threshold models of recognition memory differ from signal-detection models (Krantz,
1969; Macmillan, & Creelman, 1990). The high-threshold model describes two item
states, detected in memory and not detected in memory. A high threshold must be
met for an item to achieve detection status, so only old items surpass the threshold,
however false alarms can result from participants guessing when items are not
detected (Malmberg, 2008). The double high-threshold model describes three item
states and two high thresholds: a threshold only old items achieve to reach the
detect-old state, a threshold only new items achieve to reach the detect-new state,
and an indeterminate state for items that fail to reach either threshold, which may
result in false alarms or misses (Malmberg, 2008). Threshold models are not widely
accepted, partly because they predict the same manipulation effects on both single-
60
item recognition memory and recall, whereas many factors affect the two types of
memory performance in different ways, for example: word-frequency (e.g., Balota, &
Neely, 1980; Gregg, 1976; MacLeod, & Kampe, 1996), emotion (Hertel, & Parks,
2002), age (Craik, & McDowd, 1987), alcohol (Soderlund, Parker, Schwartz, &
Tulving, 2005), primacy and recency (Achilles, 1920; Mulhall, 1915), and types of
neurological impairment (Malmberg, 2010). These interactions have not been
explained by threshold models; however, thresholds are often a component of dualprocess models (Malmberg, 2008).
Whilst familiarity and recollection are behaviourally dissociable, it is unclear whether
they are neurally distinct, i.e. whether separate anatomical structures or neuronal
populations subserve familiarity and recollection (Rutishauser, Schuman, & Mamelak,
2008). Whilst several researchers propose that the hippocampus is concerned only
with recollection (e.g., Eldridge, Knowlton, Furmanski, Bookheimer, & Engel, 2000;
Holdstock et al., 2002; Yonelinas, 2001a), patients with hippocampal lesions often
have a general loss of memory capacity rather than a specific recollection impairment
(Manns, Hopkins, Reed, Kitchener, & Squire, 2003; Stark, Bayley, & Squire, 2002;
Stark, & Squire, 2003; Wais, Wixted, Hopkins, & Squire, 2006). An fMRI study by
Hannula and Ranganath (2009) showed that conscious recollection does not
automatically accompany hippocampal activation during associative recognition with
objects and scenes. Concomitant eye-tracking showed that participants fixated
stimuli for longer on correct trials than incorrect trials, which also resulted in higher
levels of activity within the hippocampus, but that this only lead to recollection when
accompanied by prefrontal cortex activity (Hannula, & Ranganath, 2009).
Powerful evidence for single-process models of recognition memory comes from
single-cell recordings. Rutishauser et al. (2008) made recordings from individual
61
neurons in the human hippocampus and amygdala, both part of the Medial Temporal
Lobe (MTL), whilst epileptic participants performed an item recognition task. By
asking participants to retrieve spatial locations of stimuli as well as their old/new
status, they were able to determine whether they had been able to recollect episodic
detail of the learning context. They found that neuronal activity increased in response
to the second presentation (familiarity) of an old stimulus compared to the response
to initial presentation at learning, regardless of successful recollection, but that the
amount of change determined whether or not the stimulus location was recollected
(Rutishauser et al., 2008). Consequently they concluded that human MTL neuron
firing rates signal information pertaining to the phenomenological experiences of both
recollection and familiarity, and proposed that their findings support a ‘continuous
strength of memory’ model whereby stronger neuronal activity represents stronger
memories (Rutishauser 2008; Rutishauser et al., 2008).
A different method of analysing EEG data involves looking at activity within different
frequency oscillation bands (e.g., Klimesch, 1995). Gruber, Tsivilis, Giabbiconi and
Muller (2008) analysed oscillatory EEG activity during a source discrimination
recognition study of pictures of objects. They found that Induced Gamma Band
Responses (iGBRs: 35-80Hz; 210-330ms) were not sensitive to source memory,
whereas Induced Theta Band Responses were (iTBRs: 4.0-7.5Hz; 600-1200ms).
iGBRs were higher for correctly identified “old” stimuli compared to “new” stimuli,
suggesting that increased familiarity results from increased neuronal spike activity.
Gruber et al. (2008) proposed that recollection was reflected in the theta band,
whereas familiarity was reflected in the gamma band.
62
1.3.1.2.
Dual-Process Models
Familiarity-only single-process models were criticised for being overly simplistic, and
starting in the 1970s dual-process theories of recognition memory were developed,
allowing recognition to be based on either item familiarity (signal detection), or
episodic recollection of the learning context (threshold; best analogised by Mandler,
1980; Atkinson, & Juola, 1974; Kelley, & Wixted, 2001; Malmberg et al., 2004; Reder
et al., 2000; Rotello et al., 2004).
The aim of dual-process theory is to quantify the recollective contribution to
recognition (Malmberg, 2008). Many studies show recollection and familiarity can be
differentiated behaviourally and these findings are used to support the argument that
they have different underlying neural mechanisms (Yonelinas, 2002). The two types
of response have been shown to respond differently to various manipulations (see
Gardiner, & Java, 1993; Gardiner, & Richardson-Klavehn, 2000; Rajaram, &
Roediger, 1997, for reviews). For example, “remember” judgements are impaired by
the performance of a secondary task, such as auditory vigilance, whilst learning,
whereas “know” responses are not (Gardiner, & Parkin, 1990). Repetition priming
manipulations, where the stimulus is presented very briefly right before testing,
enhance “know” responses but do not influence “remember” responses (Huber, Clark,
Curran, & Winkielman, 2008). Changing the modality of stimuli between learning and
test from pictures to words enhances “remember” responses but decreases “know”
responses (Rajaram, 1993). Interestingly, manipulations which are known to affect
remember judgements, also affect explicit tests of memory, and those known to affect
know judgements also affect implicit tests of memory (Paller, Voss, & Boehm, 2007;
Rajaram, & Roediger, 1997; Voss, & Paller, 2008; Yonelinas, 2002).
63
Dual-process models suggest that familiarity should act faster than recollection.
Delayed recollection occurs in real-life as well as the lab, for example recognising
someone’s face but not knowing who they are until after they’ve walked past
(Mandler, 2008; Mandler, & Boeck, 1974; Rabinowitz, & Graesser, 1976). Atkinson
and Juola (1973) suggest that familiarity is activated initially as a fast search, with the
slower, more thorough recollection process occurring only if familiarity is
unsuccessful. Mandler’s (1980) more recent “horse race” model proposes that both
process occur in parallel, and that the relatively automatic familiarity process finishes
before the more deliberate, intensive recollection search. It was Hintzman and
Curran (1994) who first looked at this using a response-deadline procedure (Dosher,
1984; Gronlund, & Ratcliff, 1989; Hintzman, & Curran, 1997; Reed, 1973). At
recognition, the time between stimulus presentation and participants’ response was
varied randomly. When participants were forced to respond quickly there was an
increase in false alarm to similar lures, whereas when participants had longer to
respond they made fewer false alarms. On average, participants were able to
discriminate old and new words after 420ms, but took 520ms to discriminate old items
from similar lures. The authors suggested that when distinguishing between old and
new items familiarity was rapid and accurate, but increased the number of false
alarms to lures, which were only correctly rejected once the slower recollection
process failed to produce specific contextual detail (Hintzman, & Curran, 1994).
Hintzman and Curran have used a global-matching approach to behaviourally
dissociate familiarity and recollection by manipulating the similarity between items at
learning and recognition in a plurality recognition paradigm (Hintzman, & Curran,
1994; 1995; Hintzman, Curran, & Oppy, 1992). Participants were asked to remember
plural and singular words (e.g., frog, books), including their grammatical number, and
then were tested with old words (e.g., books), new words and plurality reversed lures
64
(e.g., frogs). Due to the higher familiarity of the lures, they produced many more false
alarms than new words. By increasing the number of times items were presented
during the learning phase up to 20, and asking participants to judge at recognition
how many times the item had been presented (frequency judgements of new and
similar words should be zero), they were able to manipulate familiarity (Hintzman, &
Curran, 1995; Hintzman et al., 1992). Increasing presentation frequency increased
frequency judgements (indexing familiarity of old and similar items), but not false
alarms (indexing recollection of specific information about plurality; Hintzman, &
Curran, 1995; Hintzman et al., 1992).
ERP studies provide evidence in support of dual-process models of recognition
memory, due to the dissociable neural signatures evoked by recollection (late
parietal) and familiarity (early mid-frontal; Curran, Tepe, & Piatt, 2006). Duarte,
Ranganath, Winward, Hayward and Knight (2004) showed thirteen undergraduate
participants 350 grey-scale pictures of common objects (e.g., duck, baseball) and
asked them to judge either whether or not the object was alive, or whether or not it
was hand-operated (source discrimination). At recognition 300 old stimuli were
shown on screen for 180ms intermixed with 150 new stimuli, and participants were
asked to make a recognition judgement, and if they judged it to be old, also make a
R-K decision and an encoding category decision (“animate” or “manipulable”; Duarte
et al., 2004). They found that items given a K response evoked an earlier positivity at
frontal sites (150-450ms), and items given an R response evoked a positive-going
ERP at frontal (300-600ms) and parietal (450-800ms) sites (recollection > familiarity >
misses). Interestingly, ERPs recorded at encoding also differed between recollected
and familiar items. Items later recognised on the basis of familiarity evoked a leftlateralised positivity at anterior sites (300-450ms), whereas items later recognised
due to recollection evoked a right-lateralised positivity at anterior sites (300-450ms)
65
and bilaterally (450-600ms). Duarte et al. (2004) concluded that recollection and
familiarity are manifestations of functionally, temporally and topographically
dissociated patterns of neural activity during both encoding and retrieval. Numerous
ERP studies have replicated these findings, showing very similar time windows, with
familiarity occurring ~300-500ms and recollection ~500-800ms (see Düzel, Yonelinas,
Mangun, Heinze, & Tulving, 1997; Wilding, & Rugg, 1997a; Rugg et al., 1998a;
Curran, Schacter, Johnson, & Spinks, 2001; Curran, 2000; Wolk et al., 2006; Ally et
al., 2008). Recollection and familiarity ERPs have also been doubly dissociated using
manipulations such as picture superiority (Cohn, Moscovitch, & Davidson, 2010;
Curran, & Doyle, 2011).
In contrast to neuropsychological evidence showing only general memory-impairment
(Manns et al., 2003; Stark et al., 2002; Wais et al., 2006), or single cell recordings
showing neurons which respond in situations of recollection and familiarity
(Rutishauser et al., 2008), a double dissociation of recollection and familiarity exists
whereby some amnesic patients with selective hippocampal damage have
recollection impairments with intact familiarity (Aggleton, & Brown, 1999). Others
have impaired familiarity with intact recollection, such as patient N.B. who had
entorhinal and perirhinal cortex damage within the MTL (Bowles et al., 2007; 2010).
Düzel, Vargha-Khadem, Heinze and Mishkin (2001) reported data from a patient with
hippocampal damage who showed an absence of the Late Positive Component (LPC)
in the 500-700ms window, normally associated with recollection, but preserved
FN400 old/new effect in the 300-500ms window, thought to represent familiarity (see
Chapter 6, section 6.1.1 for further description of these components). Such evidence
may support dual-process models over unitary memory strength models (Squire et
al., 2007), although recollection and familiarity may share a notable neuroanatomical
overlap (Medina, 2008).
66
1.3.1.3.
Evaluating Models
Responses can be viewed as occupying one of three regions of a two-dimensional
‘decision space’ (see Figure 1-10), which differ on whether a “‘remember” response
depends on just recollection, the sum of recollection and familiarity, or the difference
between recollection and familiarity (Rotello, & Macmillan, 2006). The process-pure
(e.g., dual-process; Yonelinas, 2001b), model states that if there is sufficient
recollective information, an R response will be made, and that a K response only
occurs if there is sufficient familiarity but insufficient recollective information (Rotello,
& Macmillan, 2006). If there is neither enough recollective or familiarity information,
then a “new” response is made (see Figure 1-10A; Rotello, & Macmillan, 2006). In
the one-dimensional (e.g., single-process; Wixted, & Stretch, 2004), model
recollective and familiarity information is added together, so R responses are given to
items for which the sum exceeds a higher threshold than for K responses, which are
given to items for which the sum exceeds a higher threshold than for “new”
responses, but lower than for R responses (see Figure 1-10B; Rotello, & Macmillan,
2006).
Figure 1-10: Decision space for the remember-know task (without ratings). (A) process-pure (dualprocess) model, (B) one –dimensional (single-process) model, (C) sum-difference model (STREAK), xaxis shows familiarity/global memory strength, y-axis shows recollection/specific memory strength
(from Rotello, & Macmillan, 2006).
67
In the STREAK model (Rotello, & Macmillan, 2006), either recollective or familiarity
information means that an “old” response is given, but the strength components are
oppositional, balancing out to establish the response (Rotello, & Macmillan, 2006).
An R response occurs when there is relatively more specific contribution than global,
and a K response occurs when there is relatively more global contribution than
specific (see Figure 1-10C; Rotello, & Macmillan, 2006).
Rotello and Macmillan’s (2006) STREAK model is a single-process account. They
asked participants to make binary and trinary R-K decisions with confidence ratings
and found that 48/70 participants produced data that fit a one-dimensional strength
model, and only 4/70 produced data that fit a dual-process model. They concluded
that R and K responses depend on a single strength variable. By including
confidence ratings to differentiate between models, Rotello and Macmillan (2006)
may have changed the nature of the task into something that is more quantitative
than the qualitative difference between recollection and familiarity. Also, recollective
answers were excluded from the ratings scale (1-6, sure new to sure knew) for binary
tasks, whereas new items were excluded from ratings (1-3 for details and feeling of
knowing) for trinary tasks. In addition, they omitted the old-new paradigm, where the
old-new decision is made first, then for old items an R-K judgement is made (see
Rugg, & Yonelinas, 2003), which might fit a dual-process model.
However, although some data fit well, dual-process models have been criticised
because it is difficult to separately and empirically estimate the contributions of
recollection and familiarity; in addition recollection may simply be a stronger
representation of familiarity, evoking additional detail. Single-process models are
special instances of the more complex dual-process models, which revert to singleprocess when recollection does not occur (Malmberg, 2010). From the point of view
68
of philosophy of science, greater parsimony comes from single-process theories,
which regard recognition as a strength continuum rather than separate categories
(Curran, DeBuse, Woroch, & Hirshman, 2006; Medina, 2008), as it is not desirable to
over complicate models when a single-process model is sufficient to explain observed
data (see Diana et al., 2006; Dunn, 2004; Wixted, & Stretch, 2004; Yonelinas, 2002).
A model should summarise the data with fewer parameters than data points, and not
just repeat the data with a saturated model (Rotello, & Macmillan, 2006). Due to the
lack of parsimony, some researchers have argued against the need for dual-process
models, which also often don’t explain how the memory signal is generated (e.g.,
Gillund, & Shiffrin, 1984). However, unlike single-process models, which are
interested in memory strength, dual-process models are able to explain the dynamics
and organisation of recognition memory, which single-process models are not (e.g.,
Atkinson, & Juola, 1974; Mandler, 1980).
Different models are better able to account for particular recognition memory
phenomena, but this isolationist approach has failed to reach a wider consensus.
Malmberg’s (2008) framework explains more of the empirical data than other current
models, including accuracy and retrieval dynamics of single-item recognition,
associative recognition, and plurality discrimination. He proposes that an individual
selects, from among several possible recognition strategies, the one most likely to
generate an accurate answer most efficiently, thus accounting for the fit of different
related models under different experimental conditions (Malmberg, 2008). This is
supported by evidence showing that participants’ decision rules vary depending on
instructions. For example strategies are different when asked to make two
consecutive binary decisions (whether an item is “remembered”, before deciding
whether non-remembered items are “known” or new), compared to when asked to
69
make a single trinary decision (“Is the item remembered, known or new?”) (Rotello, &
Macmillan, 2006; Brown, & Bodner, 2011).
However, some researchers (e.g., Wixted, 2007a; Wixted, & Stretch, 2004; Rotello et
al., 2004) propose that rather than being exclusive theories, dual-process and signaldetection can be integrated into a single model of recognition memory, such as the
signal detection unequal variance model (Wixted, 2007a), STREAK (Rotello et al.,
2004), and single-trace dual-process models (e.g., Greve, Donaldson, & van Rossum,
2010). These models assume that as well as familiarity, recollection also lies on a
continuum, and that rather than recognition decisions being based on either
recollection OR familiarity, both sources of memory information are combined into a
unitary combined memory strength that is then compared with a criterion value to
make a recognition decision (Wixted, & Stretch, 2004; Wixted, 2007a). This is
supported by the fact that recollection can be graded, for example some contextual
information recollected vs. all contextual information recollected (Ingram, Mickes, &
Wixted, 2011; Wixted, 2007a). This view even unites apparently contrasting
neuroanatomical evidence from lesion studies, and also that of the single-cell
recordings made by Rutishauser et al. (2008). The hippocampal neurons measured
as having increasing activity with increasing memory strength may be involved in
summating the signal from separate populations of neurons representing familiarity
and recollection.
It has been suggested that ERPs provide evidence in support of a combined memory
strength model of recognition memory. Finnigan, Humphreys, Dennis and Geffen
(2002) manipulated memory strength by repeating half of the old items three times
during learning (strong) and the other half only once (weak), giving three ‘strengths’ of
item at recognition – new, weak and strong. Finnigan et al. (2002) demonstrated that
70
the FN400 was sensitive to memory strength in a graded fashion, with new items
being most negative, followed by weak and then strong items being most positive at
parietal electrodes. The LPC was found to be sensitive to decisional factors such as
confidence and accuracy, and upon visual inspection of the grand-averages appears
to show the same pattern at parietal electrodes. They argue that their data provide
evidence for a memory strength model.
It has been proposed that rather than distinguishing between recollection and
familiarity, the traditional remember-know paradigm in fact distinguishes weak from
strong memories (Wixted, & Mickes, 2010). Experimental design also influences
participant strategy when deciding whether to respond “know” or “remember”
(Kapucu, Macmillan, & Rotello, 2010; Rotello, & Macmillan, 2006). The issues of
precisely how recollection and familiarity exist, what they comprise and how they act
may still not be fully understood, but the recollection-familiarity distinction continues to
be useful in studying recognition memory.
1.3.2.
Psychophysiological Correlates of Recognition Memory
Processes
The old/new paradigm is an informative experimental design that lends itself to
combination with a psychophysiological technique, such as ERPs, fMRI, or
pupillometry. Words or pictures learned during a study phase are mixed with new
items, and participants respond “old” when they recognise an item from the study list
and “new” to items that aren’t recognised. This paradigm can be used to measure
differences in physiological responses to old and new items. The “old/new effect”
was first established in the ERP literature, but has since been studied using ERPs,
fMRI and Positron Emission Tomography (PET). Some of this literature is reviewed
below.
71
1.3.2.1.
Event-Related Potentials
ERP studies reveal more positive-going deflections associated with items correctly
identified as old compared to items correctly identified as new, known as the ERP
old/new effect (e.g., Karis, Fabiani, & Donchin, 1984; Maratos, Allan, & Rugg, 2000;
Sanquist, Rohrbaugh, Syndulko, & Lindsley, 1980). There are two main ERP old/new
effects reported in the literature, which consistently distinguish between old and new
items during recognition memory tests, the parietal Late Positive Component (LPC)
and the frontal N400 (FN400) old/new effects (for a review see Johnson, 1995). For
example, Allan and Rugg (1997) demonstrated that in an old/new recognition test
their 18 participants averaged more positive and longer-lasting left posterior ERPs for
hits than for correct rejections. Whilst both ERP effects are present over both
hemispheres, and larger for old items than new items, the frontal old/new effect,
thought to reflect familiarity, is larger over the midline, and the parietal old/new effect,
thought to reflect recollection, is larger over the left hemisphere (Allan, & Rugg, 1997;
Curran, 2000; Curran, & Cleary, 2003; Curran et al., 2001; Curran et al., 2006;
Goldmann et al., 2003; Rugg, & Allan, 2000; Vilberg, & Rugg, 2008; see Chapter 6,
section 6.1.1 for further discussion of these components). The ERP old/new effect
may provide a reliable, quantifiable marker of recognition processes, which can then
be compared to participant report to link the experience of recognition to underlying
neurocognitive activity.
The ERP old/new effect is comparable between epileptic patients who have and who
have not had a medial temporal lobectomy (Rugg, Roberts, Potter, Pickles, & Nagy,
1991), and between participants with Alzheimer’s Disease and healthy controls
(Friedman, Hamberger, Stern, & Marder, 1992; Rugg et al., 1994). Verleger (1995)
interprets this as meaning that the ERP old/new effect does not originate in the
hippocampus. However, Smith and Halgren (1989) found that the ERP old/new effect
72
was reduced in participants who had had a temporal lobectomy. The variety of
results is likely to reflect homogeneity in the precise lesions within these patient
populations, and the effects of individual anatomy on the magnitude of the electric
field measurable at the scalp (Luck, 2005).
1.3.2.2.
PET and fMRI
Neuroimaging techniques like PET and fMRI are also able to detect distinct patterns
of activity in response to old and new items. For example, a large number of fMRI
studies have shown increased activity in the anterior left frontal cortex, and medial
and lateral parietal cortex in response to correctly identified old compared to new
items, suggesting involvement in item-related retrieval success (e.g., Browndyke et
al., 2008; Donaldson, Petersen, Ollinger, & Buckner, 2001; Henson et al., 2005;
Henson, Rugg, Shallice, Josephs, & Dolan, 1999; Konishi, Wheeler, Donaldson, &
Buckner, 2000; Vilberg, & Rugg, 2008). Habib and LePage (1999) conducted a metaanalysis of PET studies and found increased blood flow to regions of the inferior and
medial parietal lobe, and the left middle frontal gyrus, when participants viewed old
items compared to new items. They concluded that stimuli need to be learned and
tested in the same modality for this old/new effect to occur, suggesting a response to
context/modality rather than just semantic or conceptual information.
1.3.2.3.
Pupil-Size
A less studied marker responsive to memory processes is pupil dilation. As
discussed in section 1.2, primarily since the 1960s, researchers have investigated the
relationship between pupil-size and a variety of psychological processes (for reviews
see Andreassi, 2000; Beatty, & Lucero-Wagoner, 2000; Goldwater, 1972; Janisse,
1977). Early studies did not specifically look at recognition memory, instead the
majority of the research concentrated on “arousal” and “mental effort”.
73
Authors
Sample
Beatty, &
Kahneman,
1966
Kuc, &
Janisse, 1967
Craven, 1972
4 male
undergraduate
students
Gardner, Mo,
& Krinsky,
1974b
Stelmack, &
Leckett,
1974
Gardner,
Beltramo, &
Krinsky, 1975
4
Krinsky, &
Gardner,
1977
6 undergraduate
students
Gardner,
Philp, &
Radacy, 1978
7 children
7-9 yrs
Garrett,
Harrison, &
Kelly, 1989
24 male
undergraduate
students
20 undergraduate
students
Otero,
Weekes, &
Hutton, 2006
Measurement
Key Pupil-Size Findings
Digit recall:
unfamiliar 7
digit number,
and familiar
name
Digit span: digits
Absolute, mm
10% increase following
familiar name; larger
pupil-size recalling wellknown than unknown
telephone number
Larger on correct trials
compared to incorrect
trials
Dilation to presented
words, which could
represent recognition
Old > new
Words
Gardner, Mo,
& Borrego,
1974a
Maw, &
Pomplun,
2004
Task & Stimuli
3 male
graduate
students
6 college
students
36 under& postgraduate
students
Recognition:
nonsense words
of consonantvowelconsonant
Recognition: 3
consonant
trigrams
Recognition:
taboo & neutral
words
Delayed recall:
digits
Comparison
tone
recognition:
1000Hz tone
Delayed recall:
digits
Passive viewing
of pictures of
nudes
Recognition:
pictures of
famous/nonfamous faces
Old/new, R-K
DRM
recognition:
words &
pictures
Old > new
Standard
photographic
technique; mm
Craig video camera;
measured from TV,
1Hz; grand average z
scores of mean size
Grand average z
scores of mean
dilation
Polymetrics
pupillometer,
manually scored
from strip-chart;
grand average z
scores of mean size
Sony video camera,
zoom f75 measured
on high res monitor;
log10 mean dilation
SR Research EyeLink
II; PDR of trial pixel
area to baseline at
initial setup
SR Research EyeLink
II & Data Viewer;
PDR of max size
during trial to
participant average
across trials
Decreasing pupil-size
increases recognition
threshold
Largest during recall,
increased during
presentation, decreased
during delay
Largest during recall,
increased during
presentation, decreased
during delay
Largest during recall,
increased during
presentation and
decreased during delay
Slide x trial interaction –
increased dilation on
subsequent presentations
for some slides
Famous faces > nonfamous faces
Old > critical distracters >
new items; remember >
know
74
Authors
Sample
Task & Stimuli
Measurement
Key Pupil-Size Findings
Laeng et al.,
2007
3 amnesic
patients
Old/new
recognition &
confidence:
colour pictures
New > old
Võ et al.,
2008
19
students
Otero,
Weekes, &
Hutton, 2011
45
students
Speeded old/
new
recognition:
positive/neutral
/negative words
Old/new, R-K
recognition:
picture names
LOP, old/new,
R-K recognition:
spoken nouns
DRM
recognition:
object names
Old/new
recognition:
words, feigning
memory loss
SensoMotoric
Remote Eye-Tracker;
single baseline to
blank white screen
from single patient
SensoMotoric HiSpeed eye-tracker,
250Hz; peak dilation
minus baseline
34 undergraduate
37
Heaver, &
Hutton, 2011
26
Kafkas, &
Montaldi,
2011
41 psychology
undergraduate
Old/new, R-K
encoding:
object pictures
Papesh, &
Goldinger,
2011
30
Papesh,
Goldinger, &
Hout, 2011
29
students
Naber,
Rutishauser,
& Einhäuser,
unpublished
Van Rijn,
Dalenberg,
Borst, &
Sprenger,
submitted
32
Old/new;
auditory low/
high frequency
words
Old/new
recognition:
spoken words &
nonwords in
two voices
Old/new
recognition:
photographs;
confidence
Learned names
of brain areas
19 psychology
undergraduate
SR Research EyeLink
II & Data Viewer;
PDR of max size
during trial baseline
Old > new; old neutral >
old positive > old
negative > new negative >
new positive > new
neutral
Remember > know > new
Deep > shallow > new
Old > false alarms > new
SR Research EyeLink
II & Data Viewer;
500Hz, PDR of max
size during trial to
trial baseline
Eye-Trac 6000; 60Hz;
peak pupil diameter
in mm minus
average trial baseline
size
Tobii 1750; 50Hz,
average pupil
diameter and
baseline diameter
Tobii 1750, E-Prime;
50Hz, peak diameter
minus average
diameter during trial
baseline
SR Research EyeLink
2000 & Matlab,
500Hz; diameter, no
baseline
SR Research EyeLink
1000; 500Hz;
percentage change
relative to baseline
Old > new under standard
instructions and when
asked to malinger or
respond new to all
items
At encoding
misses>familiar
>recollected
Old>misses/new
particularly for low
frequency words
At encoding hits>misses;
at encoding & retrieval
nonwords>words; at
retrieval original>
familiar> new voices
More constriction to new
items than old items
Pupil response decreased
with repeated
presentations
Table 2: Studies of recognition memory and pupil-size; PDR = Pupil Dilation Ratio, see section 2.1.2.1.
75
However, some studies reported results that could also be interpreted as
demonstrating recognition effects of pupil-size (see Table 2). For example, Craven
(1972, cited by Janisse, 1977) observed dilation to presented word stimuli. Also in
Kuc and Janisse’s (1967, cited by Janisse, 1977) digit span study, the larger pupilsize on correct trials compared to incorrect trials could represent a stronger memory
signal leading to recall success.
In recent years studies have begun to look directly at recognition memory and
suggest a possible relationship with pupil-size. In probably the first LTM pupil study,
Beatty and Kahneman (1966) investigated pupil-size and memory load, finding the
same sort of pupil-size changes to processing load as occurs with digit recall in STM
tasks. They compared pupil responses when participants recalled an unfamiliar 7
digit number provided by the experimenter, compared to recalling their own telephone
number from long-term memory. Long-term memory retrieval of a well-known
telephone number evoked larger pupil-sizes (0.5mm) than the seemingly more
effortful recall of an unknown number (0.34mm). They suggested that the pupil
reflected the retrieval of information from long-term memory (Janisse, 1977).
Gardner, Mo and Borrego (1974a) presented four participants with previously seen
(“well-formed” memories created during the learning phase) and unseen (not
presented during learning) nonsense words comprising a Consonant, Vowel and
Consonant (CVC). Gardner et al. (1974a) reported pupil dilation to “old” CVCs, and
constriction to “new” CVCs for all four participants. Following on from this study,
Gardner, Mo and Krinsky (1974b) attempted to replicate the results using high
frequency words presented in the auditory rather than visual modality to guard
against pupil-size changes as a result of visual features (see section 1.2.1.1). This
time they found no significant differences between pupil dilation to old and new
words. However, their sample size was underpowered with again only four
76
participants, and Gardner et al. (1974b) concluded that stimuli were too high
frequency and were equally familiar regardless of whether they were on the learning
list or not. However when using strings of three randomly generated consonants
(trigrams) in their second experiment, to remove pre-existing stimulus familiarity, they
found that mean pupil-size increased more to old learned items than to novel items
(Gardner et al., 1974b). Gardner, Beltramo and Krinsky (1975) felt pupil-size
reflected the cognitive effort arising from the storage and retrieval from memory, and
found constriction during retention when participants reported rehearsing information.
Gardner et al. (1978) suggested that rather than indicating general mental effort,
“pupillary dilation is specific to mental encoding and retrieval of information” (p. 168).
More recently Maw and Pomplun (2004) showed 20 undergraduate participants 40
famous and 40 non-famous faces, with equal numbers of male and female faces.
Wearing an EyeLink II eye-tracker, participants were asked to press a button to
indicate whether or not they recognised each face. Although the main focus of the
study was the eye-tracking data, they found that maximum pupil-size increased
relative to baseline in response to famous faces but not to non-famous faces (Maw, &
Pomplun, 2004). The authors asserted that pupil-size represents memory processes
associated with recognising a face, however they did not test pupil responses to nonfamous familiar faces, or non-face stimuli (Maw, & Pomplun, 2004).
In the first robust study using modern eye-tracking methods to look explicitly at
recognition memory, Otero, Weekes and Hutton (2006) showed 36 participants words
and pictures of everyday objects, and found that maximum pupil-size was consistently
larger when participants viewed old items previously encountered during learning,
compared to new items, independent of encoding modality (pictures vs. words). In
addition, pupil-size in response to semantically-related lures was also larger than for
77
correctly identified new items. In two follow-up studies the authors replicated their
findings using concrete nouns presented visually, and extended this to show that the
Pupil Old/New Effect (PONE) also occurred with spoken word stimuli (Otero, Weekes,
& Hutton, 2011).
Other researchers have confirmed the PONE. Võ and colleagues (2008) showed 19
participants words varying in emotional content (positive, negative and neutral) to
investigate the influence of affect on pupil-size during word recognition. During a
rapid recognition test they found larger pupil-sizes to correctly classified old words
than correctly classified new words, and that the PONE was reduced for words with
positive or negative emotional valence. They claimed to have introduced the PONE
for the first time, however, as discussed above this is not strictly the case, although
the emotional attenuation of the PONE was a novel effect. An alternative explanation
for some findings may be that stored information, such as a face or telephone
number, has emotional associations that enlarge the pupil (see section 1.2.2; Janisse,
1977). Porter et al. (2007) state that cortical areas active during tasks evoking pupil
responses are closely interconnected with areas implicated in memory, such as the
limbic and reticular activating systems, which are also involved in emotional arousal
(Brown et al., 1999; Löwenstein, & Loewenfeld, 1962). Silk et al. (2009) found that
recall of emotional words evoked larger pupil-sizes than non-emotional words,
however Võ et al.’s (2008) recognition paradigm supports earlier findings of a Ushaped pupil-size function where larger pupils are found in response to neutral stimuli
than for slightly positive and negative stimuli (e.g., Levine, & Hakerem, 1969, cited in
Janisse, 1974; Gunther, & Lussier, 1975, cited in Janisse, 1977).
In explaining their findings, Võ et al. (2008) proposed that the PONE represents the
greater cognitive effort required to correctly identify old compared to new stimuli,
78
based on extensive previous research demonstrating the relationship between pupilsize and cognitive effort (see section 1.2.2.1). They argued that recollection requires
the retrieval of qualitative contextual information, including the experience of an old
item during the study phase, which is more cognitively demanding than the correct
rejection of a new item, which does not. Võ et al. (2008) suggested that the
attenuated pupil response to emotionally valent words reflects the relative ease, and
therefore reduced cognitive load, with which words are recognized due to their
associations.
Whilst building on a substantial body of research demonstrating links between pupilsize and cognitive load, there are problems with a cognitive load account of the
PONE. Firstly, although the correct rejection of new items may not involve precisely
the same recollective processes as occur during recognition of old items, it is not
clear why recognition of previously presented items should necessarily be more
cognitively demanding than the correct rejection of novel items, and it is certainly not
the case that no cognitive effort is involved. Correct rejection may involve an effortful
memory search, and studies have found that it typically takes longer than correct
recognition (e.g., Ratcliff, & Murdock, 1976), particularly for items involving “recall-toreject” (Leding, & Lampinen, 2009). For example, in a remember/know recognition
memory ERP paradigm, Wiese and Daum (2006) found that the average response
time for hits was 1144ms, whereas the average response time for correct rejection of
non-critical lures was 1355ms. This suggests that recognizing an old item is not
necessarily more cognitively demanding than correctly rejecting a new item.
An alternative interpretation is put forward by Otero et al. (2011) who advance that,
like Finnigan et al.’s (2002) graded memory strength ERPs, the PONE represents a
combined memory signal strength. They suggest that recognition of old stimuli is
79
more automatic than rejecting new stimuli, which may be somewhat familiar but
generate no further detail on which to base a decision (Otero et al., 2011). Their
proposal is supported by findings that pupil-size is larger for items which are
recollected compared to items which are known, and is larger for known than new
items. In addition, pupil-size for false alarms (new items incorrectly identified as old)
is intermediate between correctly identified old and new items. Otero et al. (2011)
argue that both familiarity and recollection vary on a strength continuum, and that old
stimuli elicit a stronger summed familiarity and recollection signal than new stimuli,
leading to larger pupil-sizes as a direct result of the greater combined memory
strength. This explanation is supported by Papesh, Goldinger and Hout (2011) who
found that “stronger” memories were associated with larger pupil-sizes than weaker
memories. Another interesting finding was that items later correctly recognised
evoked larger pupil dilations at learning than items that were subsequently forgotten,
an effect also demonstrated in the ERP literature (e.g., Karis et al., 1984; Uhl et al.,
1990; Fabiani, & Donchin, 1995), and suggesting greater effort went into encoding
(Papesh et al., 2011).
1.4. Summary
Beatty and Lucero-Wagoner (2000) sum up by saying, “Pupillometry has served
psychophysiology well in the study of the dynamics of human cognitive processing”
(p. 159). Pupillometry is one of the more affordable psychophysiological techniques,
is portable, non-invasive, and does not rely on behavioural responses. Changes in
pupil-size consistently and reliably report the time-course of within-task, between-task
and individual variations in cognitive processing, so the relatively compact literature is
surprising. Despite an entire chapter on pupillometry in the second edition of the
Handbook of Psychophysiology, by the third edition in 2007, pupillometry is not
mentioned. Pupillometry may still be trying to dispense with the bad reputation some
80
researchers gave it in the 1960s, but equipment and techniques have improved and
there has been a recent revival in its application to areas such as cognitive load,
emotion processing, deception, and recognition memory.
The following chapters and series of experiments explore the PONE under a variety
of conditions. Experiments 1 and 2 aim to replicate the PONE in a standard explicit
test of memory, and determine whether a similar effect can be observed in an
“implicit” test of memory. It is clear that the PONE is now well established for
“explicit” recognition, but as yet it is still not clear what exactly the effect represents,
whether it is associated with specific a mnemonic process, or whether an old/new
effect can also be observed when memory is tested “implicitly”.
Experiments 3 and 4 aim to further investigate the mnemonic processes associated
with pupil dilation by measuring pupil-size in an Artificial Grammar Learning (AGL)
condition, proposing that for the implicit condition, conscious recollection will not be
available, as both the “grammatical” and “nongrammatical” strings presented in the
recognition phase will be different to those presented in the learning phase. Previous
research (Reber, 1967; 1969; Scott, & Dienes, 2008; 2009) has indicated that implicit
learning of grammatical letter strings evokes a greater sense of familiarity than nongrammatical strings. This would suggest that familiarity alone is sufficient basis for a
recognition judgement. It was predicted that a PONE would occur in the implicit
condition, reflecting familiarity signal strength, but that this effect would be smaller
than that in a standard test of memory.
Experiments 5, 6 and 7 use changes in pupil-size to explore the role of conscious
awareness in the PONE by drawing on the psychophysiology of deception and
malingering literature. Experiment 5 explores whether the PONE is under voluntary
control by asking participants to perform at their best, to deliberately perform poorly,
81
or to respond “new” to all items, in a standard recognition memory test. If, like the
ERP old/new effect, the PONE is not under voluntary control, pupil-size should
increase for old items compared to new items, even when participants say that they
do not recognise stimuli. Next Experiment 6 provides participants with instructions for
three different types of malingering strategy that might be used – not paying attention
during learning, randomly preloading a response, and not responding during
recognition. This experiment aims to artificially reduce performance measures during
a standard test of recognition memory and observe any effects on the PONE. Then
Experiment 7 asks participants to perform a secondary task during learning and
recognition in a divided attention paradigm, and looks at how genuinely reduced
recognition performance (simulating memory-impairment) affects the PONE. It was
predicted that interfering with the encoding and/or retrieval of stimuli would reduce the
magnitude of the PONE compared to when participants performed a single task at
learning and recognition.
The last empirical chapter contains Experiments 8 and 9, which explore the effect of a
graded memory strength manipulation on the PONE, in line with Otero et al.’s (2011)
memory strength explanation, and explores the idea that the old/new effects seen in
ERPs and pupil-size may index the same mnemonic processes. Few studies have
measured ERPs and pupil-size simultaneously, with none having examined
recognition memory specifically, therefore Experiment 8 recorded concurrent ERP
and pupil-size data. It was predicted that the strength manipulation would also
produce a graded effect whereby pupil-size was larger for strong items (seen three
times at learning) than weak items (seen once at learning), and larger for weak than
new items due to the differences in memory strength.
82
2. Methods
2
In Chapter 1, the pupillometry literature and relevant recognition memory literature
was reviewed. The aim of Chapter 2 is to describe the general methods used to
collect data for this thesis.
2.1. Pupillometry
2.1.1.
Background
Like fMRI, ERPs and other psychophysiological measures, great care must be taken
when attempting to draw inferences about cognitive function from pupil-size data.
One significant problem pertaining to the interpretation of any continually changing
signal (such as changes in pupil-size) is identifying individual contributions. For
example, during any task there may be several (potentially overlapping) events that
cause changes in pupil-size. Whether or not the contribution of individual events can
be quantified depends on careful experimental design and use of control conditions
where the only variable thought to change between conditions is the one of interest
(e.g., Partala, & Surakka, 2003; Oliveira et al., 2009).
2.1.1.1.
Techniques
One of the earliest references to a pupillometer is by Archimedes (212-187 BCE,
cited by Schweitzer, 1956). However, it was not until the nineteenth century that
more objective photographic methods were developed (for a fascinating review of the
history of pupillometry see Hakerem, 1967). At the time of the resurgence of interest
in pupillometry during the 1960s, most studies made use of 16mm movie cameras
with mirrors and lenses to enlarge the eye. Images were recorded on infra-red film at
1-2Hz, and once the film was developed vertical or horizontal pupil diameter was
83
measured manually with a ruler or grid from an average of 20 individual frames per
stimulus, which were projected onto a screen or table (e.g., Hess, & Polt, 1960; 1964;
Kahneman, & Beatty, 1967; Kahneman et al., 1969; Paivio, & Simpson, 1966). For
known camera-to-eye distances and magnifications, actual pupil-size in mm could
then be calculated from measured pupil-size (Janisse, 1977). Popularised by Hess
(1965), infra-red photography had the advantage that infra-red light does not trigger
the light reflex, is reflected by the iris whilst being absorbed by the pupil, and is still
detectable in a variety of light conditions (Hakerem, 1967).
Hand measurement was time-consuming and imprecise; Janisse (1977) reported a
colleague manually measuring 100,000 frames for a single study. The first device
that measured changes in pupil-size “online” electronically with a signal processor
was the 60Hz Löwenstein Pupillograph developed by Löwenstein and Loewenfeld
(1958) which scanned the eyes with a low intensity infra-red beam (Hakerem, 1967).
This type of photoelectric device was developed into “television” pupillometers during
the 1970s and over the intervening years the technology, resolution (~0.001mm) and
sampling rates (up to 1000Hz) vastly improved. Modern pupillometry research uses
mobile or semi-mobile video-based eye-trackers, such as the EyeLink II (SR
Research, Ontario, Canada), which are often head-mounted with two small infra-red
cameras angled towards the eyes, and measure variables such as gaze position,
saccades, blinks, fixations and pupil-size (see Figure 2-1a) (Wang, 2010).
A key advantage of modern pupillometry is that it is non-invasive; in particular remote
infra-red eye-tracking equipment, such as the EyeLink 1000 (SR Research, Ontario,
Canada), can be used to measure pupil-size from a desktop position without the need
for head-mounted equipment or restraints such as chin and head rests (see Figure
2-1b). This is of particular benefit with populations or paradigms where a head-
84
mounted or restrained eye-tracker would interfere with the task or be impractical, for
example studies involving developmental populations (e.g., Chatham, Frank, &
Munakata, 2009) or concurrent ERP acquisition.
a
b
Figure 2-1: (a) Head-mounted and (b) tower-mounted EyeLink eye-trackers.
Yet, in remote set-ups the camera is further from the eye (50-100cm) giving less
precise measurements and meaning that the pupil-camera distance has to be
estimated for each frame, due to unrestrained head movement (Klingner, 2010). As
very few pupillometry studies have been conducted using remote eye-trackers (e.g.,
Klingner, 2010; Klingner, Kumar, & Hanrahan, 2008), these systems are less
validated and findings less replicated than those with head-mounted eye-trackers.
However, researchers are developing minimal-calibration and calibration-free eyetracking techniques (Hansen, & Pece, 2005; Ohno, & Mukawa, 2004), combined with
increasing affordability and availability, this situation will soon change.
All experiments reported in the present thesis measured pupil size (and gaze
position) using either a head-mounted EyeLink II eye-tracker (Experiments 1, 3, 4, 5,
6, 7 and 9) or a tower/desk-mounted EyeLink 1000 (Experiments 2 and 6), both
manufactured by SR Research, Ontario, Canada. As gaze-tracking requires precise
85
localisation of the centre of the pupil, the EyeLink eye-trackers routinely and precisely
measure pupil-size in camera pixels as a by-product. When not fixating centrally, the
pupil becomes distorted (an ellipse), and whilst pupil size can be approximated by the
eye-tracker using a foreshortening division, this links pupil size with gaze position and
introduces a potential confound. Distortion also means that the pupil area measure is
more stable than pupil diameter, which is calculated based on the assumption that the
pupil at a particular moment is circular, and which will not be the case if the
participant is looking away from the centre of the screen. Therefore, in the
experiments reported here, small stimuli were presented in the centre of the screen
and participants were asked to look straight ahead.
2.1.1.2.
Data Acquisition
Once participants were seated comfortably with the head-mounted eye-tracker, or
with their chin in the chin-rest of the tower-mounted eye-tracker, a nine-point
calibration and validation procedure was carried out to ensure test-retest accuracy of
<0.5º of visual angle. Whilst this procedure is less critical for pupillometry studies
than gaze-tracking studies, a good calibration can only be performed if the thresholds
for pupil colouring have been set properly. An automatic thresholding option was
used to set the pupil colouring threshold – the greyscale threshold at which the host
PC determines that a dark circular area is the pupil (see Figure 2-2).
Figure 2-2: Pupil (shown as blue on computer display) as located by eye-tracking software.
86
2.1.2.
Pupil-Size Reporting Variables
There are many different ways to report the effects of task on pupil-size, with a
number of interconnected issues: should measurements report a size metric
(minimum, average, maximum or difference) or latency (onset, offset or peak)? If
size, should it be in diameter or area, measured in absolute units (mm) or relative
units (percentage or ratio)? Should results be adjusted with a pre- or post-stimulus
baseline, or an overall average (Janisse, 1977)? For example, is a 1mm or a 10%
change to a small pupil equivalent to a 1mm or 10% change when the pupil is already
large? An apparently equivalent change in diameter is very different when
considering the change in area, with larger increases for pupils with larger baselines.
What if two participants end up with a final pupil-size of 7mm diameter, but started
from different baselines – is it fair to conclude that one participant made more effort
(see Chapter 1, section 1.2.3), that the other was more anxious/motivated/aroused
(see Chapter 1, section 1.2.2), or that this task may have a maximum processing load
generating a maximum pupil-size? To explain the approach used in this thesis, these
issues are first given further consideration.
Wilder’s (1958) Law of Initial Values (LIV) states that “the change of any function of
an organism due to a stimulus depends, to a large degree, on the prestimulus level of
that function” (p. 199), meaning that trials with a larger baseline pupil-size will show a
smaller increase in response to the stimulus than trials with a smaller baseline pupilsize, and vice versa. The smaller pupil has more “room” to change, whereas the
larger pupil may experience a “ceiling” effect.
According to Janisse (1977), by the mid 1970s 90% of pupillometry studies reported
relative changes in diameter as a percentage of a baseline, with only a small number
reporting percent change in pupil area. Later studies commonly reported pupil
87
diameter change in millimetres by subtracting the baseline from the trial peak, which
Beatty and Lucero-Wagoner (2000) suggest is a more complete and appropriate
measure. They also argue that stimulus-evoked pupil-size changes have been
demonstrated to be independent of baseline diameter across a wide range of initial
values, tasks and laboratories (e.g., Bradshaw, 1969; 1970; Beatty, 1982b), and feel
that smaller baseline values inflate percentage measures of pupil-size change
(Beatty, & Lucero-Wagoner, 2000).
Dureman and Scholander (1962) highlight the antagonistic nature of the
psychosensory dilation and light reflexes, whereby as the pupil dilates in response to
a stimulus, the additional light falling on the retina triggers constriction of the sphincter
muscle in opposition to the dilator muscle. These influences are not necessarily
linearly related, and Dureman and Scholander (1962) suggest that because
resistance from the sphincter increases “as a positive function of the… pre-stimulus
pupillary area” (p. 51), it generates more opposition when the pupil dilates from 5.5 to
6.5mm, than from 3.0 to 4.0mm. The absolute change in diameter is the same,
whereas the pupillary area changes by 12mm2 and 7mm2 respectively, reflecting the
larger amount of activity required to produce a 1.0mm change in an already larger
pupil. They therefore prefer area measures for both changes in pupil-size and
maximum dilation (Dureman, & Scholander, 1962).
Janisse (1977) suggests that the “best” pupil-size index may be context-specific, and
that no single measure is suitable for all experimental situations. For modern eyetrackers, calibration errors, and individual differences such as eye size, camera-pupil
distance, and the refractive power of the cornea and participant glass/contacts, lead
to difficulties in back-calculating absolute size from camera pixel-count. Recent
research has therefore been carried out using relative percentage and ratio measures
88
of area rather than absolute measures of diameter (e.g., Bailey, & Iqbal, 2008;
Heaver, & Hutton, 2010; 2011; Hupé et al., 2009; Maw, & Pomplun, 2003; Kang et al.,
2009) and relative measures more easily allow for comparisons between individuals
and groups, as they account for individual differences in baseline or peak through
normalisation (e.g., Conati, & Merten, 2007).
There is no statistical test to directly compare experimental effects within a group to
the same effect within a different group (manipulation by group interactions in a
between-group design) due to pre-existing differences in the variable of interest, or
level of noise, leading to non-equivalent groups (Luck, 2010; Nieuwenhuis,
Forstmann, & Wagenmakers, 2011b). One way in which researchers have attempted
to address this issue is to compare relative rather than absolute effects. For example,
if in an experiment a group of participants in condition A have an average pupil-size
measured by the eye-tracker as 2,000 camera pixels larger for old items than new
items, but another group in condition B only show a difference of 200 pixels larger for
old then new items, this would produce a significant main effect of group and a
significant item-type by group interaction based on absolute values. If group A’s
pupils were larger to start with, and in fact changed from 8,000 to 10,000 pixels,
whereas group B’s pupils were smaller and changed from 800-1000 pixels, this is a
relative change of 25% for both groups, and an analysis would reveal a significant
main effect of item-type, but no interaction with condition and no main effect of group
(see Figure 2-3). Some experiments require a between-group design, for example if
naïve participants are needed in both conditions, therefore a relative measure of
change helps to counter between-group differences.
5000
100
4500
4000
3500
3000
A
2500
B
2000
1500
1000
500
0
new
Relative pupil size (%age)
Absolute pupil size (pixels)
89
old
80
70
60
50
A
40
B
30
20
10
0
new
Item type
a
90
old
Item type
b
Figure 2-3: (a) Experiment reported as absolute pupil-size values, (b) Same experiment reported in
relative values.
Having considered the issues above, the output variables of our equipment and the
current standard practise with modern eye-trackers, we decided to use an area
measure rather than a diameter measure, and a relative measure (pixel ratio) rather
than an absolute measure. As this thesis is concerned with the magnitude of the
memory signal we chose to use a size change metric rather than latency. Maximum
pupil-size was used rather than average pupil-size because although maximum
measures may be sensitive to random noise at the peak, making the maximum
slightly larger than the true value, the average measure would mean excluding a large
proportion of trials where participants either blinked, looked around the screen or the
eye-tracker momentarily lost the pupil (situations which reduce measured pupil-size).
2.1.2.1.
Pupil Dilation Ratio
The EyeLink II (500Hz) and EyeLink 1000 (1000Hz) eye-trackers used to collect
pupil-size data for this thesis provide an arbitrary unit of measurement, reflecting the
number of camera pixels occluded by the pupil image as determined by the EyeLink
host software, together with other metrics including number and duration of fixations,
eye position and blinks, which can be analysed in Data Viewer (SR Research,
90
Ontario, Canada). The number of pixels occluded by the pupil typically falls between
800-2000 units (±1 unit) and 10% of variance is due to factors such as the distance
between the camera and the eye, angle of the camera (the EyeLink II camera is
positioned below the eye rather than in front of it where it would occlude vision; the
EyeLink 1000 when positioned on the desktop is below eye-line and the angle is
therefore affected by participant height), gaze position, and individual differences
such as pupil position, corneal distortion, resting pupil-size and overall eye size. The
measurements are difficult to convert to absolute units, and whilst diameter is
measured, area is recommended by the manufacturers (SR Research, Ontario,
Canada).
In order to gauge the degree of stimulus-evoked pupil response and generate a
comparable measure, Maw and Pomplun (2004) devised a Pupil Dilation Ratio (PDR)
by dividing maximum trial pupil-size by a single baseline measured immediately after
initial calibration of the EyeLink II, and found PDR was significantly larger to famous
faces than non-famous faces. However, because during long experiments the iris
muscle fatigues (Löwenstein, & Loewenfeld, 1964; Peavler, 1974), stimulus-evoked
responses diminish (Francis, & Kelly, 1969; Lehr, & Bergum, 1966; Löwenstein, &
Loewenfeld, 1952) and baseline pupil-size decreases due to autonomic arousal
decrement (Lehr, & Bergum, 1966; Sternbach, 1966; Woodmansee, 1966), the
experiments in this thesis took a baseline measure at the beginning of each trial
(Otero et al., 2011). PDRs reported here represent the maximum pupil-size during
the 1750 or 2000ms trial period as a proportion of the maximum pupil-size during the
250 or 200ms pre-stimulus baseline period (Otero et al., 2011).
To reduce fatigue, and the potential effects of loss of interest or boredom,
experiments were also limited to 30 minutes of measurement, stimuli were presented
91
in a random order, and rest breaks were offered between blocks of more than 50
trials (Klingner, 2010; Sternbach, 1966). In an experimental trial, pupil-size usually
reaches maximum >1000ms after stimulus presentation (Beatty, 1982b), whereas
after ~2000ms participants may lose focus on the stimuli, and occasionally look away
from the centre of the screen, which can lead to confounds in pupil-size data (Otero,
2010). Therefore recognition trials were 2000ms long in order to ensure maximum
dilation was captured. Data was recorded from one eye (typically the right eye)
because the pupils are yoked (see Chapter 1, section 1.1.1; Reeves, & Swenson,
2004). Mean pupil-size can offer a more robust measure of response in situations
where trials differ in length, however trials in this thesis are of equal length between
participants, therefore measurements of maximum pupil-size were recorded rather
than mean pupil-size, consistent with the literature.
2.1.2.2.
Measurement Issues
As argued earlier, not all changes in pupil-size are necessarily due to the
experimental effect under investigation, and Loewenfeld (1958) reports that externally
triggered changes in pupil-size are overlaid on a signal with a variable level of noise.
It is therefore highly probable that changes in pupil size caused by “internal” events
are also superimposed on this varying signal (see Figure 2-4). As discussed in
Chapter 1 section 1.1.2.5, one source of background noise is endogenous pupillary
unrest, or hippus (Woodmansee, 1966), which may change diameter by 1% every
second, and up to 10-20% every few seconds (Woodmansee, 1966). Hippus is
amplified by fatigue and passivity, and suppressed by alertness and mental activity
(Bouma, & Baghuis, 1971; Kahneman, 1971; Miller, & Newman, 2005). This means
that a pre-stimulus baseline measure of pupil-size may include more hippus than the
trial measurement. Researchers have taken a variety of approaches to dealing with
hippus, including averaging over repeated measures of at least 8 trials per participant
92
to increase the signal-to-noise ratio (Hakerem, & Sutton, 1964; Woodmansee, 1966),
selecting participants who are familiar with the testing environment and procedure,
and who are alert and well-rested (Janisse, 1977), and using range correction,
designed for assessing heart-rate and electrodermal responses, to reduce noise by
computing each individual’s possible range of pupil-sizes and expressing the actual
value as a proportion of the individualised range (Lykken, 1972).
Figure 2-4: Sources of variation in measurements of pupil diameter (from Klingner, 2010).
However, Kahneman (1973) was confident that task-related focus was sufficient to
reduce pupillary hippus, stating that changes in pupil-size are so reliable and
predictable, that he took no further steps to control for it. In order to minimise the
influence of artefacts, the baseline measure of pupil-size in this thesis is maximum
pupil-size. This is because the PONE is concerned with increases in the maximum
pupil-size in response to stimuli – by measuring the maximum size during the
baseline, the likelihood that any baseline to trial difference is simply the difference
between the pupil at minimum and maximum amplitude during hippus is reduced.
Gaze position affects the size of the pupil as perceived by eye-trackers such as the
EyeLink, which measure pupil-size in eye-tracker camera pixels (Pomplun, &
Sunkara, 2003; Pomplun, Sunkara, Fairley, & Xiao, 2009) (Tobii eye-trackers
93
measure the length of an ellipse fitted to the pupil which is less distorted by
perspective; Klingner, 2010). Due to effects of gaze position on the EyeLink, in this
thesis all events were presented in the centre of the monitor. Each trial did not begin
until the participant had fixated the centrally-positioned drift correction dot, which was
then followed by a fixation cross. To prevent luminance changes, which could trigger
the PLR (see section 1.2.1.1), an isoluminant mask consisting of “&&&&&&” or
“HHHHHH” (matched with stimuli for character length) preceded each stimulus. This
was followed by the stimulus, which either remained in position for the duration of the
trial, or was replaced by the isoluminant mask. Within an experimental condition
stimuli were the same number of characters, presented in a Monospaced font, and all
words subtended no more than 3o of visual angle to ensure they fell within the fovea,
reducing the likelihood that participants would need to make a second fixation to read
the stimulus and induce local luminance changes or distortions in pupil shape.
Stimuli were achromatic, stationary and of constant contrast in order to control for
pupil-size changes in response to visual stimulus features, and participants were
asked to remain still during the experiment to prevent accommodation-related
changes (Loewy, 1990).
Another source of noise are blinks and the lid-closure reflex (see Chapter 1 section
1.1.2.3), which causes both pupils to briefly contract and redilate. In a methodology
paper, Nakayama (2006) found that blinks had a significant effect on both pupil-size
and Pupil Unrest Index (PUI) when participants carried out a mental arithmetic task,
and that estimation of pupil-size during blinks provided a pupil grand average that
was more sensitive to the experimental manipulation in a small sample size (n=5). An
alternative method to blink estimation or correction (Klingner, 2010), and the one
used in this thesis, is blink suppression – asking participants to try to blink only
between trials, as trials were only 2000ms long. The experimenter could see the eye
94
image during the experiment and wait until after a blink occurred to trigger the next
trial. Blink reduction was especially important in Experiment 8, which used ERP
measures, as eye blinks and eye movements have a detrimental effect on EEG
recording (see section 2.2.1) due to large electrical signals produced by the eye
muscles, and whilst random blinks will average out, stimulus-linked blinks will average
into the grand average.
By pooling data from 20,000 binocular blinks Klingner (2010) asserted that stimuluslinked blinks were associated with a reduction in pupil-size in the subsequent 1000ms
by ~0.03mm, and an increase in pupil-size between 1000-2000ms by ~0.05mm
(Klingner, 2010). However, increased cognitive load is known to be associated with
both higher blink rates and increased pupil-size, so this is not surprising (see Chapter
1, section 1.2.4.2). As stimuli were presented on all trials, the procedure was the
same for old and new items, and blinks were minimised as far as possible, it is
unlikely that blinks accounts for the difference in pupil-size for old and new items. To
check this, the number of blinks made during each trial was automatically recorded
and output alongside the pupil-size data. Paired-sample t-tests on blink rate between
conditions were performed across all experiments and no significant differences in
blink rates for old and new items were found.
2.1.3.
Pupil-Size Analysis
During an experiment the EyeLink records raw data every 1-2ms (depending on
sampling rate) including a timestamp, the X and Y position of the eye(s) being tracked
in screen pixel co-ordinates, pupil-size and event-related messages signalling when
the display software has reached particular points in the experiment, for example
mask and stimulus onset and offset. Raw data is imported into Data Viewer (SR
Research, Ontario, Canada), which allows the specification of time windows for the
95
extraction of calculated variables, such as maximum pupil-size, into summary trial
reports which then were analysed in Excel (Microsoft) and SPSS 18 (IBM).
As seen in Chapter 1, task-evoked increases in pupil-size are usually less than 0.51.0mm (Klingner, 2010) or 10-20% of baseline (Beatty, 1982), which is equivalent in
magnitude to the constant background variation caused by other influences such as
hippus. This makes it virtually impossible to identify the task-related signal from noise
on any individual trial. One method of enhancing the signal-to-noise ratio is to
average multiple repeated trials of the same task (Beatty, & Lucero-Wagoner, 2000;
Pomplun, & Sunkara, 2003), leading to consistent task-evoked responses averaging
in, whilst random variations such as hippus should average out. This is the approach
taken in this thesis; the mean number of trials per participant per condition was 33.9
old (range = 20.5-58.5, SD = 4.65) and 37.7 new (range = 20.6-73.1, SD = 4.29).
Within experiments, statistical comparisons of pupil-size for old and new items were
made for all trials and/or only correct trials for old and new items. The analyses
restricted to the correct responses allowed us to be sure that any differences in pupil
size between old and new items were not due to some “error” response that may
occur when participants realise they have made an incorrect response. In some
instances it was not appropriate to analyse only correct trials, for example in
Experiment 6 where participants were randomly preloading answers, or in Experiment
5 where they were instructed to say “new” to all items. Unfortunately it was not
possible in most cases to analyse changes in pupil-size associated with incorrect
responses, even though previous research has shown an interesting effect of an
intermediate pupil-size for false alarms (Otero et al., 2011) – as insufficient false
alarms and misses were made by participants to produce a meaningful average for
analysis.
96
Other techniques that researchers have used to analyse pupil data include waveform
analysis (e.g., Kuipers, & Thierry, 2011), wavelet transforms or decomposition to find
brief discontinuities that differentiate cognition from light reflexes (e.g., Leal, Neves, &
Vieira, 2011; Marshall, 2002; 2007), frequency-domain analysis (Kumar, n.d., cited in
Klingner, 2010; Moloney et al., 2006; Nakayama, & Shimizu, 2004), principle and
independent component analysis (Jainta, & Baccino, 2010), analysis of average pupilsize (e.g., Klingner, 2010), and analysis of area under the pupil-response curve
(Webb, Honts, Kircher, Bernhardt, & Cook, 2009). Oliveira et al. (2009) used
Principle Component Analysis (PCA) to isolate changes in pupil diameter due to the
local luminance changes from changes due to stimuli in a web search task. Jainta
and Baccino (2010) used PCA and Independent Component Analysis (ICA) to reveal
the main and hidden contributions to pupil responses from participants who were
reading or performing easy or difficult mental arithmetic. They identified three
components in the individual pupil responses, only one of which changed with task
difficulty and accounted for 50% of variance during the most difficult task. Jainta and
Baccino (2010) proposed that this component might be mental effort, but did not
speculate as to the nature of the other two components.
The focus of the present thesis was to characterise the recently identified PONE in
terms of the cognitive processes that may underlie it. As such, the relatively
straightforward PDR was used as the methods described above are more suited to
characterising the nature of the pupil response itself, possibly with a view to exploring
its neural underpinnings.
97
2.2. Event-Related Potentials (ERPs)
2.2.1.
ERP Data Acquisition
Event-Related Potentials (ERPs) are averaged waveforms identified as positive or
negative deflections of the Electroencephalograph (EEG) voltage (Luck, 2005; see
Chapter 6, section 6.1.1 for further discussion). Measured from the scalp, EEG
recordings are made using arrays of electrodes in predetermined positions, covering
the majority of the participants head, often as part of a net or cap. The traditional
international 10–20 and 10–10 electrode configurations have 22 and 42 electrodes
respectively (see Figure 2-5; Jasper, 1958; Michel et al., 2004; Pivik et al., 1993).
Figure 2-5: Traditional 10-20 and 10-10 electrode configurations (adapted from Reynolds, & Richards,
2009).
“Geodesic sensor nets”, such as the Electrical Geodesics Inc. (EGI) nets used in
Experiment 8 of this thesis (see Chapter 6, section 6.2), have a high-density (or
dense-array) electrode configuration of 64, 128, or 256 equidistant electrodes,
approximately 35-40 mm apart (for adults, depending on head size and net size)
98
covering most of the scalp surface (Electrical Geodesics Inc; Tucker, 1993; Tucker,
Liotti, Potts, Russell, & Posner, 1994, see Figure 2-6 and Figure 2-7).
Figure 2-6: Geodesic sensor net 64 and 128 channel electrode maps (adapted from Reynolds, &
Richards, 2009).
Figure 2-7: 128 channel Geodesic sensor net worn by models.
99
The traditional 10-20 positions have been updated to replace T3/T4 with T7/T8, and
T5/T6 with P7/P8 in line with guidelines issued by the American
Electroencephalographic Society (1991; 1994), and, to make room for P9/P10,
electrodes P7/P8 were moved to a more superior site (see Figure 2-8).
Figure 2-8: Modified combinational nomenclature for the 10-10 system (from the American Clinical
Neurophysiology Society, 2006).
The geodesic configuration differs from the electrode placement sites of the
International 10-10 and 10-20 systems, but an approximate correspondence between
the two has been established (Luu, & Ferree, 2000; Srinivasan, Tucker, & Murias,
1998). Luu and Ferree (2000) computed corresponding positions between the two
systems using maximum arc length distance of 0.20 of the radius (see Figure 2-9).
100
Figure 2-9: The 10-10 system overlaid on the Geodesic 128 sensor net (adapted from Luu, & Ferree,
2000).
The changes in voltage measured by EEG equipment are tiny (often less than 10µV),
and typically smaller than the background electrical fluctuations from electronic
equipment (including experimental apparatus), skin potentials, muscle and eyemovements (Luck, 2005). In order to accurately measure dipoles at the surface of the
scalp the signal must be amplified 10,000-50,000 times, along with the noise,
therefore it is important to reduce sources of noise as far as possible. Ways of doing
this include shielding electronic equipment whilst keeping it as far from the participant
101
as possible, reducing the impedance of the skin via abrasion or using a highimpedance system, asking participants to relax so that they are not clenching their
facial or neck muscles, and asking participants to fixate the centre of the screen and
only blink between trials (Luck, 2005). As occurred in Experiment 8, participants can
be seated in a Faraday cage, which shields the entire experimental setup from
outside electromagnetic radiation; however care must be taken with any equipment
(e.g., monitors, eye-trackers) used inside the cage as emissions will be trapped within
the cage. Endogenous noise can also occur in the form of alpha-waves in
participants who are tired, bored or sleepy, similar to hippus. This can be reduced by
using well rested participants and offering rest breaks and water between blocks.
2.2.2.
ERP Data Analysis
Analysis of EEG data was carried out by segmenting the continuous epoch into
sections that began 200ms prior to stimulus presentation (-200ms) and ended
1000ms after stimulus presentation, using event markers communicated by the
experimental software (E-Prime 2.0) to the EEG software (Net Station), and grouped
according to condition. Grand-average waveforms were generated for each
participant, baseline-corrected using the period -200 to 0ms, and re-referenced offline
to average mastoid electrodes (Nunez, 1981) after these channels were verified as
having made a good, relatively artefact-free recording (it is not possible to make this
check for the online reference during recording). Average mastoid reference was
selected as this is a commonly used reference, allowing comparison with other
studies, it is also a convenient site that does not cause discomfort or distraction,
provides good electrical conduction, and given that all references have their
limitations it is as good a reference as any (Luck, 2005; however see also Dien,
1998).
102
Net Station Waveform Tools (Electrical Geodesics, Inc) were used to extract the
mean amplitude for the windows of interest, and data were analysed using SPSS 18
(IBM). Data were analysed in their raw form, without normalisation conversion to
relative differences as only one electrode factor had more than two levels, and
although theoretically appealing, Urbach and Kutas (2002) state that normalisation
fails to achieve the desired effect of removing significant condition by electrode site
interactions (Luck, 2005). To reduce potential violations of sphericity, and retain
topographical detail, lateral electrode position was analysed using two factors with
two levels (hemisphere: left, right; site: superior, inferior), rather than one factor with
four levels (Luck, 2005). Analyses included parallel strings of electrodes (e.g.,
MacKenzie, & Donaldson, 2007) rather than groups of electrodes (e.g., Curran,
2000).
In line with other ERP memory research (e.g., Curran, 2000; Finnigan et al., 2002;
MacKenzie, & Donaldson, 2007), multiple univariate analyses were performed, rather
than a single multivariate analysis (where Mauchly’s test indicated that the
assumption of sphericity had been violated, degrees of freedom were corrected using
*
Greenhouse-Geisser estimates of sphericity (ε < 0.75) or +Huynh-Feldt estimates of
sphericity (ε > 0.75)). This was because whilst MANOVA would determine whether or
not there was an effect, it would not reveal where, and would still require multiple
follow up ANOVA. Kiebel and Friston (2004) have stated that multivariate and mass
univariate are not dissimilar, and Groppe, Urbach and Kutas (2011) review four
promising methods of mass univariate analysis that control for familywise error and
are particularly suited for exploring ERP data.
Luck (2010) questions whether it is legitimate to compare grand averages of
conditions containing different numbers of trials as is common in ERP experiments.
103
Issues may arise because waveforms formed from fewer trials will contain more noise
due to a lower signal-to-noise ratio than averages of larger numbers of trials (Luck,
2010). Variation in noise is more of a concern when measuring peak amplitude rather
than mean amplitude; it biases the measurement because a spurious peak has more
influence over the final value of the peak measurement due to fewer contributing trials
(Luck, 2010). The ERP experiment reported in this thesis is concerned with mean
amplitude, which is an unbiased measure even when trial numbers differ, and so
perfectly good trials do not need to be discarded simply to even the numbers (Luck,
2010).
2.3. Stimuli and Participants
2.3.1.
Word Selection
With the exception of the artificial grammar condition of Experiments 3 and 4 (see
Chapter 4), study and recognition lists for the experiments in this thesis were created
using nouns selected from the MRC Psycholinguistic Database (Coltheart, 1981).
Items within a list were matched for length (5, 6, or 7 letters long), and between lists
were matched for lexico-semantic features such as frequency, familiarity and
imageability, according to K-F norms (Kucera, & Francis, 1967), as these are known
to affect both memory performance (e.g., Balota, & Neely, 1980; Bauer, Olheiser,
Altarriba, & Landi, 2009; Deese, 1960; Gorman, 1961; Gregg, 1976; Schulman,
1967), and pupil-size (Colman, & Paivio, 1969; 1970; Kahneman, & Peavler, 1969;
McElvain, 1970; Paivio, & Simpson, 1966; 1968; Simpson, & Paivio, 1968; see
Chapter 1, section 1.2.3.6). For example, an item is more likely to be correctly
recognised as “old” if it is relatively uncommon (Shepard, 1967), or if it is concrete
rather than abstract (Gorman, 1961). Words with emotional or offensive content were
excluded due to their potentially biasing effects on both memory (Bauer et al., 2009)
104
and pupil-size (Johnson, 1971; Stelmack, & Mandelzys, 1975; see Chapter 1, section
1.2.2). Different word lists were used for each experiment to prevent confounds if a
participant took part in more than one experiment, which was particularly important for
experiments with implicit tests of memory as performance on these has been shown
to be more enduring than on explicit tests of memory (Allen and Reber, 1980).
2.3.2.
Selection of Participants
Participants were required to be native English speakers, and to have normal, or
corrected-to-normal, vision in at least one eye, with glasses or contact lenses to be
brought to the experiment if required. Whilst both contact lenses and glasses have
effects on the refraction of infra-red light (Dahlberg, 2010; Wang, 2010), experiments
in this thesis involve participants fixating the centre of the screen, minimising artifacts
normally associated with gaze-tracking. Participants were prevented from
participating in to both Experiments 3 and 4, or in both Experiments 8 and 9, as the
stimuli used were identical.
Although resting pupil-size decreases with age, pupil responses appear to remain
relatively unchanged during adulthood (see Chapter 1, section 1.1.1; Kim, Beversdorf,
& Heilman, 2000; Kumnick, 1956; Porter et al., 2010). However, the correlation
between pupil-size and age is slightly lower in psychiatric populations than healthy
controls, whereby resting pupil-size in participants with mental health problems does
not decrease with age as much as for healthy participants, possibly due to comorbid
anxiety (Liakos, & Crisp, 1971). Even treated and remitted schizophrenics have
abnormal pupil responses, such as decreased dilations to stimuli and faster working
memory overload compared to controls (Andreassi, 2000; Granholm, & Verney, 2004;
Minassian, Granholm, Verney, & Perry, 2004). Therefore it was relatively important to
keep age constant as baseline measurements are used to calculate PDR, and it was
105
important for the present experiments to recruit participants without significant mental
health difficulties. The average age of 377 participants (113 male) across all
experiments was 24.4 years (SD = 6.9 years).
As pupil-size is influenced by thoughts and feelings, including physical sensations
such as pain or discomfort, and background noises and distractions, care was taken
to seat participants comfortably in an adjustable chair, maintain the laboratory at an
adequate temperature, provide water and breaks if required and remove sources of
distraction.
106
3. Replicating and Extending the Pupil Old/New Effect
3
In Chapter 1, the pupillometry literature and relevant recognition memory literature
was reviewed, showing that since the 1960s a number of studies have, directly or
indirectly, measured the effect of encountering novel versus learned stimuli on pupilsize (e.g. Bradshaw, 1967; Bradshaw, 1968; Garrett, Harrison, & Kelly, 1989;
Gardner et al., 1974a; 1974b; Maw, & Pomplun, 2004; Otero et al., 2006; 2011; Võ et
al., 2008). It is clear that the Pupil Old/New Effect (PONE) is now well established for
“explicit” recognition, but what is not clear is exactly what the effect represents,
whether it is associated with specific mnemonic processes, or to what extent the
PONE is linked to conscious awareness. As yet no research has investigated
whether the PONE can also be observed when memory is tested “implicitly”. The
aims of the two experiments presented here were to replicate the PONE in a standard
explicit memory recognition test, and explore whether a similar effect can be
observed in an “implicit” recognition test.
A large body of literature suggests that a distinction can be made between explicit
and implicit memory. Implicit memory is defined experimentally as a change in
performance that results from previous exposure to items, but in the absence of a
conscious recollective experience of the exposure itself (Dienes, & Berry, 1997;
Stevens, Wig, & Schacter, 2008). One line of research that has been used to
support the distinction between implicit and explicit memory is experimental
dissociations in healthy participants – manipulations which affect performance on one
but not the other type of memory task (see Foster, & Jelicic, 1999). For example, a
Levels Of Processing (LOP) manipulation (Craik, & Lockhart, 1972) enhances
performance on explicit tests of memory for items processed more deeply during
study (semantic processing; e.g., “Generate a sentence using this word”) compared
107
to items processed shallowly during study (surface processing; e.g., “What colour is
the text?”), but does not differentially affect performance when memory is tested
implicitly, with techniques such as perceptual identification (Jacoby, & Dallas, 1981),
word stem completion (Graf, Mandler, & Haden, 1982), and word fragment
completion (Roediger, Weldon, Stadler, & Riegler, 1992). In contrast, “surface”
manipulations (such as keeping the font of items constant between study and test)
enhance performance on implicit tests of memory, but do not affect explicit tests of
memory (Stevens, Wig, & Schacter, 2008).
In addition, a growing literature demonstrates implicit-explicit dissociations in
neuropsychological patients, including relatively intact implicit memory in patients with
amnesia (e.g., Laeng et al., 2007; Verfaellie, Bauer, & Bowers, 1991; for a review see
Schacter, McAndrews, & Moscovitch, 1988). For example, Nissen and Bullemer
(1987; see also Nissen, Willingham, & Hartman, 1989) showed that when presented
with a ten-trial repeating light sequence, which the participants then had to recreate,
the performance of participants with Korsakoff’s amnesia improved as the sequence
was repeated, consistent with performance of control participants, even though unlike
controls the Korsakoff’s participants were not consciously aware of the pattern.
As discussed in Chapter 1, section 1.3.1.2, familiarity as measured by “know”
responses (Gardiner, 1988; Tulving, 1985a) has been shown to respond in a similar
manner to so called “implicit” memory in a variety of experimental manipulations
(Paller, Voss, & Boehm, 2007; Yonelinas, 2002). For example, priming manipulations
where the stimulus is briefly presented prior to testing lead to feelings of familiarity
(Jacoby, & Whitehouse, 1989), which enhanced “know” responses, without
influencing “remember” responses (Huber, Clark, Curran, & Winkielman, 2008).
Similarly, familiarity and recollection often dissociate in neuropsychological patients
108
with amnesia, Parkinson’s disease or Alzheimer’s (Cohn et al., 2010; O'Connor, &
Ally, 2010; Weiermann, Stephan, Kaelin-Lang, & Meier, 2010; Yonelinas et al., 1998).
To date, only one study appears to have assessed the relationship between pupil-size
and recognition memory performance in amnesic patients. Laeng et al. (2007) tested
two patients with amnesia by reading unfamiliar or fictional short facts (e.g., “penguins
lay blue eggs”), while an image related to a word in the sentence (e.g., “eggs”) was
presented in one of four boxes on a computer screen. The patients were asked
questions based on the facts (e.g., “what colour are penguins’ eggs?”), and despite
answering very few questions correctly, their eyes focused on the box in which the
relevant image had been presented. The authors argued that this finding suggests
the patients had an implicit memory for the location of the picture. In a second
experiment, Laeng et al. (2007) carried out a picture based old/new recognition task
with three amnesic patients. One patient answered “new” to every question, whilst
the others made 46.6% and 70.2% correct decisions. Interestingly, in contrast to
most recent research which has found significantly larger pupil-sizes for old words,
the amnesic patients’ pupil-sizes were greater for new than old words. The reasons
for this discrepancy are not clear, but given their amnesic status, it might be argued
that any correct recognition would be based on implicit or non-recollective memory
processes (suggesting that the standard PONE reflects primarily recollective
mnemonic processes).
There are a number of issues which hinder interpretation of the Laeng et al. (2007)
study. The sample was very small, the amnesic patients had different aetiologies
and, as is clear from their performance, a wide range of memory difficulties.
Importantly, baseline pupil measurements were only taken from a single participant
during the blank screen between pictures. Stimuli were a mixture of colour pictures
109
and photographs of objects and faces, and only one participant had not seen the
same set of stimuli in the two preceding experiments. Areas of the brain have been
shown to respond to stimulus novelty (Habib, & LePage, 1999; Tulving, & Kroll,
1995), showing decreased activation to repeated presentation (repetition
suppression; e.g., Schacter, & Buckner, 1998; Buckner et al., 1998; Grill-Spector, &
Malach, 2001; van Turennout, Ellmore, & Martin, 2000), and some studies have
demonstrated a “novelty” pupil response (Andreassi, 2000; Janisse, 1977), which
may be part of an orienting response to salient stimuli (Lynn, 1966; Pavlov, 1927;
Sokolov, 1963). It is therefore possible that in the absence of a PONE in amnesic
participants, the novelty response is instead the most visible influence on pupil-size
between new and old items.
Implicit tests of memory offer a way to study the influence of familiarity on recognition
decisions in the absence of recollective processes, and LOP manipulations (e.g.,
Craik, & Lockhart, 1972) have been shown to have different effects on explicit and
implicit tests of memory (e.g., Jacoby, & Dallas, 1981). The first experiment
combined a LOP manipulation at study with explicit and implicit tests of recognition
memory. Its aim was to replicate the PONE, and determine whether the PONE can
also be observed when participants are exposed to novel and learned items, but are
not asked to make a recognition decision based on conscious recollection. To this
end, a standard recognition task was used in one condition (called the “explicit”
condition), and a perceptual fluency recognition task (as used by Jacoby and Dallas,
1981) was used in the second condition (hereafter referred to as the “implicit”
condition). In perceptual fluency tasks, the recognition of very briefly presented items
is facilitated if they have previously been encountered during the learning phase,
without participants necessarily being able to consciously recollect the initial learning
experience. The LOP manipulation employed at learning was included in an attempt
110
to replicate Otero et al.’s (2011) finding that the PONE was larger for items which had
been encoded with deep orienting instructions compared to those encoded with
shallow orienting instructions, and larger for shallow items than new items.
On the basis of previous research, pupil-size was expected to increase for old
compared to new items in the explicit recognition test. However, Võ et al.’s (2008)
cognitive load account of the PONE (see Chapter 1, section 1.2.2.1) does not
consider whether the PONE should also be observed when participants recognise
previously encountered stimuli based on a familiarity judgement rather than
recollection. If a PONE was observed during a recognition judgment made on the
basis of familiarity alone, this would undermine the claim that pupil dilation reflects a)
recollective processes and b) cognitive load. Due to their emphasis on conscious
recollection, Võ et al.’s (2008) cognitive account would predict that the PONE in the
explicit condition should be smaller for deeply encoded items (less effort needed for
recollection) than shallowly encoded items, whereas Otero et al.’s (2011) “memory
signal” explanation predicts that deeply encoded items should elicit a larger pupil-size
(stronger memory signal) than shallowly encoded items, which should elicit a larger
pupil-size than new items.
It was predicted that the PONE would not be observed for either semantic or shallow
items in the implicit recognition task because conscious recollective processes would
not be involved in perceptual priming. If, however, the absence of PONE allowed the
novelty pupil effect to dominate pupil-size (as may have occurred in Laeng et al.,
2007), it was predicted that new items might elicit a larger pupil-size than surface and
semantic items in the implicit condition. An open prediction was made as to whether
there would be a difference between surface and semantic items in the implicit
condition. An LOP effect was also predicted between conditions in the behavioural
111
data whereby a higher number of semantic (deeply encoded) items would be correctly
recognised in the explicit condition due to deep encoding, and a higher number of
surface (shallowly encoded) items would be correctly recognised in the implicit
condition due to perceptual fluency.
3.1. Experiment 1 – Implicit vs. Explicit Tests of
Recognition
3.1.1.
Method
3.1.1.1.
Participants
Fifty participants (20 male; age range: 18.4-36.5, M = 23.4, SD = 4.01) with normal or
corrected-to-normal vision in at least one eye, were recruited from the psychology
course-credit and subject pools at the University of Sussex, and through personal
contact. Participants were briefed with a detailed consent form (specific to the
condition to which they were allocated) and verbal description of the methods and
procedure, and were invited to ask questions. Participation was on a voluntary basis,
and participants were thanked and debriefed at the end with a verbal description of
the aims of the study and the opportunity to ask further questions. The experiment
was approved by the relevant ethics committee.
3.1.1.2.
Materials/Apparatus
Two word lists were created using nouns selected from the MRC Psycholinguistic
Database. The learning list comprised 40 items, whilst the recognition list contained
those 40 nouns plus 40 new items. All items were 6 letters long, and lists were
matched for concreteness (learning items range = 305-634, M = 523, new items
range = 296-635, M = 525), familiarity (learning items range = 436-621, M = 547, new
112
items range = 428-632, M = 549), imageability (learning items range = 368-643, M =
551, new items range = 324-646, M = 552) and frequency (learning items range = 21348, M = 95, new items range = 18-492, M = 95), according to the K-F norms. Words
likely to elicit a strong emotional response were removed. The learning list and
recognition test were presented in black 20pt Arial font in the centre of a light grey
background under fixed illumination. Items were presented using the Experiment
Builder software associated with the EyeLink II eye-tracker (SR-Research, Ontario).
All items are presented in Appendix A. During the recognition test, pupil-size was
recorded using an EyeLink II head-mounted eye-tracker with a temporal resolution of
2ms and a spatial resolution of around 0.25 degrees. The stimuli were displayed on a
21 inch CRT monitor with a screen resolution of 1,280 x 1,024 pixels and a refresh
rate of 60Hz. Actual screen dimensions were 40cm horizontal and 30cm vertical.
Participants were seated approximately 70cm from the screen in an adjustable chair
that had been modified to prevent any rotational movement.
3.1.1.3.
Design and Procedure
The experiment comprised two conditions, an explicit recognition condition and an
implicit recognition condition, and in a between-subject design half of the participants
completed each condition. Both conditions contained a learning and recognition
phase. The learning phase was identical between the conditions. The 40 learning list
items were presented on screen for 3000ms. Before each item was presented,
participants saw a screen instructing them to process the following item at either a
surface level (“How many vowels in…”) or semantic level (“Give me a synonym
for…”). The same items were associated with the same LOP (shallow or deep) for all
participants across both conditions, and an equal number of items were processed at
the deep and shallow level. Participants were required to give an answer for each
question.
113
For the explicit recognition condition, the 80 recognition list items were presented on
screen for 2000ms after which the participant was prompted to say whether the item
was old (previously encountered in the learning phase) or new (not previously
encountered). The next screen required participants to estimate confidence in their
decision with a number between 1 and 5, where 1 represented a complete guess and
5 represented total confidence. This screen was then replaced by a drift-correction
dot in the centre of the screen before presentation of the next item. Old/new
judgements and confidence estimates were recorded on the computer after each
recognition item.
For the implicit recognition condition, the 80 recognition list items were present for two
monitor-refresh cycles (33.3ms at 60Hz), as determined by the eye-tracker software
(Experiment Builder, SR Research). In order that participants were looking at the
item during its brief presentation, participants were asked to blink whilst the driftcorrect dot was on screen and state when they were ready to proceed without
blinking for a few seconds. They were then required to read the item aloud and their
answers were noted by hand and entered into the computer at a later stage.
3.1.1.4.
Pupil Recording
Maximum pupil-size was recorded from the right eye during each recognition period –
the time during which the item was on screen for the explicit condition, and the time
which the item and the re-mask were on screen for the implicit condition. A Pupil
Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated expressing the
maximum pupil-size for each 2000ms recognition trial as a proportion of the maximum
pupil-size during that trial’s 200ms baseline.
114
3.1.2.
Results
3.1.2.1.
Behavioural Data: Old/New Responses
The proportion of correct responses to old and new items was calculated for implicit
and explicit conditions. A 2 x 2 ANOVA with within-subject factor of item-type (old vs.
new) and between-subject factor of condition (implicit vs. explicit) showed a
significant main effect of condition (F(1,48) = 21.1, MSE = 0.013, p <.001, ηp2 =.305)
– more correct responses were made in the implicit condition than the explicit
condition. This main effect was qualified by a significant item-type by condition
interaction (F(1,48) = 27.9, MSE = 0.012, p <.001, ηp2 =.368) – participants
responded correctly more often to old items (M =.947, SD = 0.074) than new items in
the implicit condition (M =.840, SD = 0.142, t(24) = 5.36, p <.001, r =.545), but
responded correctly more often to new items (M =.851, SD = 0.115) than old items in
the explicit condition (M =.723, SD = 0.112, t(24) = -3.22, p <.01, r =.302). The main
effect of item-type was not significant (F(1,48) = 0.220, MSE = 0.012, p >.05, ηp2
=.005; see Figure 3-1).
Figure 3-1: Proportion of correct responses in each condition. Error bars show standard error of
mean.
115
In order to determine the effect of the LOP manipulation on recognition memory, the
proportion of correct responses to surface and semantic old items were analysed in a
2 x 2 mixed ANOVA with LOP (surface vs. semantic) as a within-subject factor and
condition (implicit vs. explicit) as a between-subject factor. The main effect of LOP
was significant (F(1,48) = 135.7, MSE = 0.012, p <.001, ηp2 =.739) – more correct
responses were given to semantic items than to surface items. This main effect was
qualified by a significant LOP by condition interaction (F(1,48) = 113.6, MSE = 0.012,
p <.001, ηp2 =.703) – participants responded correctly much more often to semantic
items (M =.970, SD = 0.035) than surface items in the explicit condition (M =.476, SD
= 0.217, t(24) = 11.6, p <.001, r =.848) whereas in the implicit condition the proportion
of correct responses to semantic items (M =.958, SD = 0.064) and surface items (M
=.936, SD = 0.093) were more similar and only approached significance (t(24) = 1.90,
p =.07, r =.131). The main effect of condition was also significant (F(1,48) = 69.0,
MSE = 0.018, p <.001, ηp2 =.590; see Figure 3-2).
Figure 3-2: Proportion of correct responses to surface and semantic items in each condition. Error
bars show standard error of mean.
116
3.1.2.2.
Behavioural Data: Confidence
To determine the relationship between confidence and performance, the average
participant-reported confidence was calculated for the explicit condition (confidence
estimates for the implicit condition was not measured because participants were not
making a recognition judgement). Confidence ratings were analysed in a repeated
measures ANOVA with within-subject factors of item-type (old vs. new), and response
(old vs. new), which showed a significant main effect of item-type (F(1,17) = 98.1,
MSE = 0.094, p <.001, ηp2 =.852) – average confidence for old items was higher than
for new items. This was qualified by a significant item-type by response interaction
(F(1,17) = 118.4, MSE = 0.156, p <.001, ηp2 =.874) – average confidence was higher
for old items given an old response (M = 4.44, SD = 0.324) than a new response, and
average confidence was higher for new items given a new response (M = 3.71, SD =
0.794) than an old response. The main effect of response was not significant (F(1,17)
= 0.001, MSE = 0.441, p >.05, ηp2 <.001; see Figure 3-3).
Average Confidence Rating
p < .001
old
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
new
p < .001
Old
New
Item-Type
Figure 3-3: Average confidence rating for old and new responses to old and new items in the explicit
condition. Error bars show standard error of mean.
117
3.1.2.3.
Pupil-Size Data
Average PDR for old and new items was calculated for the implicit and explicit
conditions. As PDR is a function of baseline pupil-size, baseline pupil-sizes to old
and new items in both conditions were compared to ensure that any differences in
PDR were not due to baseline differences. The difference was not significant (F(1,48)
= 0.908, p >.05, ns, ηp2 =.019). A 2 x 2 ANOVA of PDR with within-subject factor of
item-type (old vs. new) and between-subject factor of condition (implicit vs. explicit)
showed that the main effect of item-type was not significant (F(1,48) = 0.63, MSE <
0.001, p >.05, ηp2 =.013), neither was the main effect of condition (F(1,48) = 0.74,
MSE = 0.011, p >.05, ηp2 =.015), however the interaction between item-type and
condition was significant (F(1,48) = 18.9, MSE < 0.001, p <.001, ηp2 =.282). As
predicted, average PDR was larger for old items (M = 1.160, SD = 0.069) than new
items (M = 1.150, SD = 0.065, t(24) = 2.71, p <.01, r =.234) in the explicit condition,
and was larger for new items (M = 1.180, SD = 0.081) than old items (M = 1.165, SD
= 0.080, t(24) = -3.40, p <.01, r =.325) in the implicit condition (see Figure 3-4).
old
p < .01
new
1.2
p < .01
Pupil Dilation Ratio
1.19
1.18
1.17
1.16
1.15
1.14
1.13
1.12
Implicit
Explicit
Condition
Figure 3-4: Pupil dilation ratio for old and new items in each condition. Error bars show standard
error of mean.
118
The ANOVA was repeated, but with the data averaged across only those trials to
which participants gave correct responses. An identical pattern of results was found
(perhaps unsurprisingly given the high level of accuracy with which both tasks were
completed). The main effect of item-type was not significant (F(1,48) = 2.49, MSE <
0.001, p >.05, ηp2 =.049), neither was the main effect of condition (F(1,48) = 0.37,
MSE = 0.011, p >.05, ηp2 =.008), however there was a significant interaction between
item-type and condition (F(1,48) = 34.6, MSE < 0.001, p <.001, ηp2 =.419).
To determine whether pupil-size was influenced by the LOP manipulation, a 2 x 3
ANOVA on mean PDR values for correct items, with within-subject factor of LOP (new
vs. surface vs. semantic) and between-subject factor of condition (implicit vs. explicit)
was performed. There was a significant main effect of LOP (F(1.75,82.3) = 8.99,
MSE = 0.001, p <.001, ηp2 =.161; Mauchly’s test indicated that the assumption of
sphericity had been violated (χ2(2) = 7.05, p <.05), therefore degrees of freedom were
corrected using Huynh-Feldt estimates of sphericity (ε = 0.97)) – average PDR for
semantic items was larger than for surface items or new items. This main effect was
qualified by a significant LOP by condition interaction (F(1.75,82.3) = 7.64, MSE <
0.001, p <.001, ηp2 =.140; Mauchly’s test indicated that the assumption of sphericity
had been violated (χ2(2) = 5.97, p <.05), therefore degrees of freedom were corrected
using Huynh-Feldt estimates of sphericity (ε = 0.98)) – average PDR for new items
was smaller than for surface and semantic items in the explicit condition, whereas in
the implicit condition, PDR for new items was larger than for surface or semantic
items. The main effect of condition was not significant (F(1,47) = 0.285, MSE =
0.016, p >.05, ηp2 =.006; see Figure 3-5).
119
New
Pupil Dilation Ratio
1.20
Surface
Semantic
p < .001
p < .01
p < .05
1.19
1.18
1.17
1.16
1.15
1.14
1.13
Implicit
Explicit
Condition
Figure 3-5: Pupil dilation ratio for new, surface and semantic items in each condition. Error bars show
standard error of mean.
A priori t-tests revealed that, as predicted, in the explicit condition average PDR for
correct semantic items (M = 1.173, SD = 0.078) was significantly larger than for
correct new items (M = 1.146, SD = 0.062; t(24) = 4.28, p <.001, r =.433), as was
average PDR for correct surface items (M = 1.169, SD = 0.068; t(24) = 2.33, p <.05, r
=.198), and average PDR for correct semantic items was larger than for correct
surface items at trend levels (t(24) = 1.95, p = .08). In the implicit condition, average
PDR for correct new items (M = 1.178, SD = 0.081) was larger than for correct
surface items (M = 1.158, SD = 0.088; t(24) = 3.00, p <.01, r =.272), however the
differences between correct new and correct semantic, and correct surface and
correct semantic items were not significantly different (ts < 1.6, ns).
3.1.2.4.
Pupil-Size Data: Confidence Analysis
Participants made higher confidence ratings on average to their correct “old”
judgments compared to their correct “new” judgements in the explicit condition. If the
increase in pupil dilation observed when participants view old compared to new items
reflects some kind of “arousal” associated with being correct, we might expect a
relationship between the average PDR for correctly identified old items and
confidence estimates for those correctly recognised stimuli. It is important therefore
120
to establish the extent to which the increase in PDR is associated with the increase in
confidence that is associated with giving an old compared to new response.
A 2 x 2 repeated measures ANOVA of mean PDR values for correct items, with
within-subject factors of item-type (old vs. new) and confidence (high vs. low) showed
a significant main effect of confidence (F(1,21) = 11.2, MSE = 0.004, p <.01, ηp2
=.347) – average PDR was larger for items ranked with high confidence (4 or 5) than
for items with low confidence (1-3). This was qualified by a significant item-type by
confidence interaction (F(1,21) = 1.87, MSE = 0.002, p <.05, ηp2 =.140) – despite
being overall slightly less confident in their correct rejections than their correct
recognitions, the increase in average PDR with increasing confidence was greater for
old items than new items. The main effect of item-type was not significant (F(1,21) =
1.09, MSE = 0.001, p >.05, ηp2 =.049). Analysis was restricted to the 22 participants
who had at least 5 high and low confidence correct old and new judgements.
3.1.3.
Discussion
The present experiment replicated the basic PONE effect when memory was tested
explicitly, but interestingly there was no PONE in the implicit condition. In addition,
overall PDR was larger in the implicit condition compared to the explicit condition.
Given the experimental design, it is not possible to say whether either of these effects
is due to differences in task requirements (reading vs. recognition) or duration of
stimulus presentation (33ms vs. 2000ms).
In an attempt to clarify the results of Experiment 1, a “control reading” condition was
carried out, which acted as an additional comparison. This experiment was included
in order to provide an estimate of the effect on pupil-size of simply reading items
121
presented onscreen for 2000ms during the recognition phase, without the
requirement of making an old/new judgment.
3.2. Experiment 1b – Reading Condition
3.2.1.
Method
3.2.1.1.
Participants
Twenty-five participants (6 male; age range: 18.92-41.33, M = 24.84, SD = 5.45), with
normal or corrected-to-normal vision were recruited from the psychology course-credit
and subject pools at the University of Sussex, and through personal contact.
Participants were briefed with a detailed consent form and verbal description of the
experiment, and invited to ask questions. Written consent was obtained prior to
testing and participants were fully debriefed at the end. The experiment was
approved by the relevant ethics committee.
3.2.1.2.
Materials/Apparatus
As for Experiment 1.
3.2.1.3.
Design and Procedure
In a within-subject design all participants completed a single ‘control’ reading
condition with a learning and recognition phase. The learning phase was identical to
that of Experiment 1. During the recognition phase, the 80 recognition list items were
presented on screen for 2000ms. The participant was then required to read the item
aloud and their answers were noted by hand and entered into the computer at a later
stage. The next screen required participants to estimate confidence in their decision
with a number between 1 and 5, where 1 represented a complete guess and 5
122
represented total confidence. This screen was then replaced by a drift-correction dot
in the centre of the screen before presentation of the next item. Confidence
estimates were recorded on the computer after each recognition item.
3.2.1.4.
Pupil Recording
As for Experiment 1.
3.2.2.
Results
3.2.2.1.
Behavioural Data
Participants performed at ceiling, correctly reading 100% of old and new items with
maximum confidence (see Figure 3-6).
Proportion Correct Responses
Old
New
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Control
Condition
Figure 3-6: Proportion of correct responses in each condition. Error bars show standard error of
mean.
3.2.2.2.
Pupil-Size Data
Average PDR for old and new items was calculated for the control reading condition.
As PDR is a function of baseline pupil-size, baseline pupil-sizes to old and new items
123
were compared to ensure that any differences in PDR were not due to baseline
differences. The difference was not significant (t(24) = 1.46, p >.05, ns, r =.082).
There was no PONE in the ‘control’ reading experiment, PDR for old (M = 1.119, SD
=.0.0535) and new items (M = 1.127, SD = 0.0628) were not significantly different
(t(24) = 1.44, p >.05, ns; see Figure 3-7). All items were included in the analysis as
none were read incorrectly.
old
new
Pupil Dilation Ratio
1.15
1.14
1.13
1.12
1.11
1.10
Control
Condition
Figure 3-7: Pupil dilation ratio for old and new items in each condition. Error bars show standard
error of mean.
3.2.3.
Discussion
Together, Experiments 1 and 1b sought to replicate the relative increase in pupil-size
that occurs when participants view previously learned items during a recognition test
compared to novel items, and to determine whether it also occurs when stimuli are
presented too briefly to evoke conscious recollection by participants, but may
nonetheless reveal effects of prior learning.
In Experiment 1 participants’ pupil-sizes increased to a greater extent when they
viewed old items compared to novel items in the explicit condition, a replication of the
PONE found by previous researchers. Items that had been deeply encoded
124
(semantic) or shallowly encoded (surface) produced a larger pupil-size than new
items, in the explicit condition, and semantic items were larger than surface items (at
trend level), similar to the findings of Otero et al. (2011).
If, as Võ et al. (2008) argue, pupil-size reflects cognitive effort, the PONE in the
explicit condition should be smaller for deeply encoded items (less effort needed
recollection). However, it was not, and the finding that the PONE is larger (at trend
level) for deeply encoded items supports Otero et al.’s (2011) suggestion that pupilsize reflects memory “strength” – the PONE is larger for deeply encoded items
because they evoke a stronger memory. The LOP manipulation influenced the
behavioural data: in the explicit condition, more semantic items were correctly
identified than surface items, whereas in the implicit condition there was no effect of
LOP on performance.
Importantly, the standard PONE was not present in the implicit condition, where pupilsize was larger for new items compared to old items. This pattern of results is similar
to those of Laeng et al. (2007) who looked at implicit memory in amnesic patients and
found larger pupil-sizes to novel stimuli. In the absence of the PONE, a “novelty”
response may have been visible instead, in the form of a larger pupil to novel stimuli
than non-novel stimuli (Laeng et al., 2007; Lynn, 1966; Pavlov, 1927; Sokolov, 1963).
An alternative explanation might be linked to presentation duration – the increased
difficulty of the task of reading novel stimuli only presented for 33ms compared to
learned stimuli that had been primed and would be easier to read even at brief
duration. The implicit condition had larger PDRs compared to the explicit condition
and the control reading task in Experiment 1b. This finding might be explained by an
element of increased cognitive effort in that the overall difficulty of the task has been
increased by the decreased presentation duration (see Chapter 1, section 1.2.2.1).
125
Experiment 1b was designed to explore the impact of task demands on pupillary
responses, by asking participants to read the stimuli rather than make an old/new
judgement. Interestingly, no PONE was found, suggesting that the PONE may occur
as a result of the requirement to make a recognition decision, rather than as an
automatic process resulting from the presentation of learned stimuli.
It is probably to be expected that trials that lead to a high level of confidence were the
same trials that had a “strong” memory and therefore a larger PDR. However,
although participants were more confident in giving old responses to old items than
new responses to new items, when only considering trials with a high confidence
rating (4 or 5), average PDR to correctly identified old items was still significantly
larger than for correctly identified new items.
Together, the results of Experiments 1 and 1b suggest that the increase in pupil-size
that occurs when participants encounter previously studied items, and recognise them
as old, reflects neurocognitive processes associated with explicit, but not implicit
recognition memory, and that the pupillary response is a function of task demands
(recognition memory test) as opposed to being the automatic consequence of being
exposed to items previously encountered during a learning phase (as in the control
reading condition), or an artefact of level of confidence.
There were a number of methodological issues that limit the extent to which further
inferences can be made. Whilst Experiment 1b was intended as a control task, it
differed to both the explicit and implicit conditions of Experiment 1 on both task
requirements and presentation duration, and therefore did not help to explain the
results of the implicit condition. Although there was an LOP effect, whereby a higher
number of semantically processed items were recalled than surface items in the
explicit condition, the implicit condition was too easy – participants had no difficulty
126
reading nearly all of the stimuli during the perceptual recognition test – old or new,
and as such it was not possible to tell whether more surface items than deep items
were also recalled in the implicit condition, as would be predicted by perceptual
fluency. Many participants also reported verbally that they had become aware that
the implicit condition was a memory test for the items they had just learned, so it is
unlikely that the short presentation time was a truly “implicit” test of memory. These
and other methodological issues are addressed in Experiment 2.
3.3. Experiment 2 – Short vs. Long Presentation Duration
One of the key issues with the design of Experiment 1 is that it did not allow the
effects of presentation duration (2000ms vs. 33ms) and task (reading vs. recognition)
on pupil-size to be separated. It was not clear whether the absence of the PONE in
the implicit condition arose because the manipulation had allowed old items to be
read more easily due to priming (implicit test of memory), or because participants had
to read stimuli rather than make an old/new judgement on them. In Experiment 2
these confounds were removed and the design strengthened by adopting a withinsubject approach and replacing the concepts of “explicit” and “implicit” tests of
memory with 2000ms exposure (long duration) and visual perceptual threshold
exposure (short duration) during the recognition phase, and for each exposure
duration asking participants to either read or identify the word as old or new.
Another methodological issue that arose in Experiment 1 was that in the “implicit”
condition most participants perceived all stimuli very clearly, whereas some couldn’t
read any of the stimuli at all. As a result, in Experiment 2, a thresholding program
was used to calculate individual presentation durations for each participant, by
increasing or decreasing presentation duration until approximately 60% of short
duration items could be correctly identified. Within the implicit literature, the
127
Cheesman and Merikle (1984) distinction differentiates an objective threshold for
presentation duration at which participants perform at chance because they are
genuinely guessing, and a slightly higher subjective threshold when presentation
duration produces the feeling of guessing at point of recognition, however participants
perform at levels above chance, but without consciously recollecting stimuli. This
type of marginally perceptible or “subliminal” (Cheesman, & Merikle, 1984)
presentation allows memory to be tested “implicitly” (e.g., Chan, 1992; Cheesman, &
Merikle, 1984; 1986; Dienes, Altmann, Kwan, & Goode, 1995; Dienes, & Berry, 1997;
Merikle, 1992).
A further methodological improvement was the introduction of an isoluminant visual
mask. Unless followed by a mask, briefly presented stimuli can leave an “afterimage”
created by temporary pigment changes in the photoreceptors of the retina, which
result in negative images of the stimuli persevering beyond the brief presentation.
Experiment 2 included a mask of 6 ampersands (“&&&&&&”) both before and after
stimuli in the same size and font. The mask also minimised any change in screen
luminance from a blank screen to one showing a stimulus (see Chapter 2, section
2.1.2.2). In order to reduce overall accuracy levels, items appeared either above or
below a fixation cross at random. This served to make it more difficult to read the
stimulus because it was not fixated and its position could not be predicted; it was
necessary because pilot testing of the thresholding program revealed that even at the
minimum presentation duration of a single monitor refresh (10ms at 100Hz), centrally
positioned stimuli could be read by most participants. Finally, in order to simplify the
design, the LOP manipulation and confidence measure were removed, allowing the
effects of task and duration manipulations to be seen more clearly.
128
Following on from the results of Experiments 1 and 1b, the PONE was predicted to be
present in the long-duration recognition memory condition (as per the explicit
condition), absent for long duration reading conditions (no difference in maximum
pupil-size for correctly identified new and old items as per the control condition), and
reversed for the short-duration condition (like the implicit condition). No prediction
was made for the short-duration recognition condition as it was not known whether it
was duration or task requirements impacting on pupil-size in Experiments 1 and 1b),
however if participants are able to make old/new judgements accurately at shortduration presentations, then the PONE may be observed. Consistent with
Experiment 1, due to perceptual fluency, more correct old responses than new
responses were expected in the short duration reading condition.
3.3.1.
Method
3.3.1.1.
Participants
Twenty-eight participants (2 male; age range: 18.3-49.4, M = 23.1, SD = 7.19), with
normal or corrected-to-normal vision were recruited from the psychology course-credit
and subject pools at the University of Sussex, and through personal contact.
Participants were briefed with a detailed consent form and verbal description of the
experiment, and invited to ask questions. Written consent was obtained prior to
testing and participants were fully debriefed at the end. The experiment was
approved by the relevant ethics committee.
3.3.1.2.
Materials/Apparatus
Using nouns selected from the MRC Psycholinguistic Database nine word lists were
created. Four learning lists each comprised 40 items, and the four recognition lists
contained the 40 nouns from the corresponding learning list plus 40 new items. An
129
extra 40 item list was created to determine individual thresholds for the short duration
conditions. All words were 6 letters long, and lists were matched for concreteness
(learning items range = 259-646, M = 486; new items range = 254-652, M = 489;
calibration items range = 267-643, M = 487), familiarity (learning items range = 277628, M = 509; new items range = 256-632, M = 512; calibration items range = 269634, M = 511), imageability (learning items range = 280-643, M = 506; new items
range = 289-643, M = 511; calibration items range = 301-646, M = 512) and
frequency (learning items range = 21-478, M = 54; new items range = 18-472, M =
56; calibration items range = 20-483, M = 54), according to the K-F norms.
The learning and recognition lists were presented in black 20pt Arial font in the centre
of a light grey background under fixed illumination. Items were presented using
Experiment Builder software (SR-Research, Ontario). All items are presented in
Appendix B. During the recognition test, pupil-size was recorded using an EyeLink
1000 tower-mounted eye-tracker with a temporal resolution of 2ms and a spatial
resolution of around 0.15 degrees. The stimuli were displayed on a 21 inch CRT
monitor with a screen resolution of 1,280 x 1,024 pixels and a refresh rate of 100Hz.
Actual screen dimensions were 40cm horizontal and 30cm vertical. Participants were
seated with their chin on a rest 70cm from the screen in an adjustable chair that had
been modified to prevent any rotational movement.
3.3.1.3.
Design and Procedure
The experiment comprised four conditions, two recognition memory conditions, one
short and one long-duration, and two reading conditions, one short and one longduration. In a within-subject design all participants completed all four conditions.
Before the main experiment started, participants completed a thresholding task using
the extra 40 noun list. List items replaced a mask (“&&&&&&”) either above or below
130
the fixation cross. This procedure was used because Experiment 1 showed that
without this unpredictability in location most people could read the items, even at the
lowest duration when looking directly at a single mask. Starting with a presentation
time of 100ms, the program shortened presentation duration by 10ms when
participants read an item correctly and lengthened it by 10ms when they failed to
identify the stimulus. After 40 items had been presented the proportion of correct
responses at the various presentation durations was displayed; the experimenter was
then able to select a presentation duration that was neither so short that participants
were unable to read the majority of items, nor so long that they were performing at
ceiling. The duration chosen for most participants was the shortest duration that
resulted in an equivalent number of correct and incorrect responses. Where there
was no duration at which numbers were equal, the duration at which correct
responses were greater than incorrect responses was chosen.
All conditions contained a learning and recognition phase. The learning phase was
identical across the conditions. The 40 learning list items were presented in the
centre of the screen for 2000ms and participants were asked to try to remember
them. For each recognition phase, list items replaced a 500ms mask either above or
below a central fixation cross. The order in which the four conditions were performed
was rotated across participants. For the long duration reading condition the 80
recognition list items were presented and left on screen for 2000ms. The participant
was asked to read them out loud as they were presented. This screen was then
replaced by a drift-correction dot in the centre of the screen before presentation of the
next item. The procedure was identical for the long duration recognition condition but
participants were instead asked to state whether the item was old (previously
encountered in the learning phase) or new (not previously encountered). For the
short duration reading condition the 80 recognition list items replaced the mask for the
131
brief length of time determined in the thresholding task for that participant. Items
were then remasked for 2000ms. Participants were asked to read the words out loud
(or give their best guess) as they were presented. This screen was then replaced by
a drift-correction dot in the centre of the screen before presentation of the next item.
For the short duration recognition condition the procedure was identical except
participants were asked to state whether the item was old (previously encountered in
the learning phase) or new (not previously encountered). In order that participants
were looking at the item during its brief presentation, in the short duration conditions
participants were advised to blink only whilst the drift-correct dot was on screen. In all
conditions, old/new judgements and correct/incorrect reading responses were
recorded on the computer after each item.
3.3.1.4.
Pupil Recording
Maximum pupil-size was recorded from the left eye during each recognition period –
the time for which the item was on screen for the long duration conditions, and the
time for which the item and the mask were on screen for the short duration conditions.
A Pupil Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated
expressing the maximum pupil-size for each 2000ms recognition trial as a proportion
of the maximum pupil-size during that trial’s 250ms baseline.
3.3.2.
Results
3.3.2.1.
Behavioural Data
The proportion of correct responses to old and new items was calculated for each
condition. In all four conditions participants performed significantly above chance
(50%) at correctly identifying old and new items (all ts > 2, ps <=.05), demonstrating
that the threshold measuring task worked – participants were not performing at ceiling
132
or floor levels (with the exception of the long reading condition where participants
achieved 100% correct responses). A 2 x 2 x 2 ANOVA on proportion of correct
responses, with within-subject factors of task (reading vs. recognition), presentation
duration (long vs. short), and item-type (old vs. new), showed a significant main effect
of presentation duration (F(1,27) = 64.3, MSE = 0.035, p <.001, ηp2 =.700) –
participants made more correct responses at long durations compared to short
durations. The main effect of task was also significant (F(1,27) = 91.1, MSE = 0.032,
p <.001, ηp2 =.770) – participants made more correct responses in the reading
conditions compared to the recognition conditions. The main effect of item-type was
not significant (F(1,27) = 1.98, p >.05, ns; see Figure 3-8).
Proportion Correct Responses
Old
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
New
p < .001
Short Reading
Short Recognition
Long Reading
Long Recognition
Condition
Figure 3-8: Proportion of correct responses for old and new items in each condition. Error bars show
standard error of mean.
These main effects were qualified by a number of interactions, including a significant
task by presentation duration interaction (F(1,27) = 61.9, MSE = 0.020, p <.001, ηp2
=.700) – increased presentation duration produced a larger improvement in
participants’ correct responses in the reading task (34.8%, SD = 13.9%) than in the
recognition task (5.04%, SD = 18.8%; t(27)= 7.86, p <.001, r =.696). The interaction
between task and item-type was also significant (F(1,27) = 11.0, MSE = 0.011, p
133
<.01, ηp2 =.290) – participants responded correctly to more old items (83.8%, SD =
7.36%) than new items (81.4%, SD = 8.05%) in the reading tasks (t(27)= 1.90, p
=.068, r =.118), but to more new items (63.3%, SD = 14.7%) than old items (56.6%,
SD = 12.7%) in the recognition tasks (t(27)= 2.56, p <.01, r =.195). Finally, the
presentation duration by item-type interaction also reached significance (F(1,27) =
10.9, MSE = 0.006, p <.01, ηp2 =.291) – participants responded correctly to old items
more often than new items at short duration (M = 1.25%, SD = 0.13%), and to new
items more often than old items at long duration (5.58%, SD = 5.0%; t(27)= 3.30, p
<.01, r =.287). The three-way interaction was not significant (F(1,27) = 1.64, p >.05
ns), however a priori t-tests showed that for the short reading condition, participants
were able to correctly read more old items (67.6%, SD = 14.7%) than new items at
trend level (62.8%, SD = 16.1%; t(27) = 1.90, p =.07, r =.118). Unlike the long
duration recognition condition (and the ‘explicit’ condition of Experiment 1), where
more new items were correctly identified than old items (t(27)= 4.34, p <.001, r
=.411), in the short duration recognition condition correct identification of old and new
items was very similar (t < 1, ns).
3.3.2.2.
Pupil-Size Data
Average PDRs for old and new items were calculated for all conditions. As PDR is a
function of baseline pupil-size, baseline pupil-sizes for old and new items in each
condition were compared to ensure that differences in PDR were not due to baseline
differences. The difference was not significant (F(1,27) = 1.21, p >.05, ηp2 =.043).
A 2 x 2 x 2 ANOVA on PDR, with within-subject factors of task (reading vs.
recognition), presentation duration (long vs. short), and item-type (old vs. new)
revealed a significant main effect of presentation duration (F(1,27) = 19.0, MSE =
0.002, p <.001, ηp2 =.413) – PDR values were greater when items were presented for
134
short durations. There was no main effect of item-type (F(1,27) = 0.224, p >.05, ns),
and no significant main effect of task (F(1,27) = 0.152, p >.05, ns). The main effect of
presentation duration was qualified by a significant interaction with task (F(1,27) =
6.72, MSE = 0.003, p <.01, ηp2 =.199) – there was a trend for stimuli displayed for a
short duration to produce a larger PDR in the reading task (M = 1.148, SD = 0.059)
than the recognition task (M = 1.134, SD = 0.059, t(27)= 1.93, p =.06, r =.122),
whereas stimuli displayed for a long duration produced a larger PDR in the
recognition task (M = 1.123, SD = 0.053) than in the reading task (M = 1.103, SD =
0.045, t(27)= -2.14, p <.05, r =.145). There was no significant interaction between
duration and item-type (F(1,27) = 0.863, p >.05, ns) or between task and item-type
(F(1,27) = 2.93, p =.098, ns), and the three-way interaction was not significant
(F(1,27) = 0.738, p >.05, ns; see Figure 3-9).
Old
New
Average Pupil Dilation Ratio
p = .06
1.16
p < .05
1.15
1.14
1.13
1.12
1.11
1.1
1.09
Short Reading
Short Recognition
Long Reading
Long Recognition
Condition
Figure 3-9: Pupil dilation ratio for old and new items in all four conditions. Error bars show standard
error of mean.
The analysis was repeated with PDRs calculated across only those trials in which old
and new items were correctly identified. A 2 x 2 x 2 ANOVA on PDR, with withinsubject factors of task (reading vs. recognition), presentation duration (long vs. short),
and item-type (old vs. new) now revealed a significant main effect of item-type
(F(1,27) = 14.3, MSE = 0.001, p <.001, ηp2 =.352) – average PDR was greater for old
135
items compared to new items. The main effect of presentation duration was also
significant (F(1,27) = 18.8, MSE = 0.003, p <.001, ηp2 =.410) – PDR values were
greater when items were presented for short durations than long durations. There
was no significant main effect of task (F(1,27) = 0.20, p >.05, ns). These main effects
were qualified by a significant task by presentation duration interaction (F(1,27) =
6.71, MSE = 0.002, p <.01, ηp2 =.203) – like before, stimuli displayed for a short
duration produced a larger PDR in the reading task (M = 1.150, SD = 0.058) than the
recognition task at trend level (M = 1.135, SD = 0.061, t(27)= 1.78, p =.086, r =.105),
whereas stimuli displayed for a long duration produced a larger PDR in the
recognition task (M = 1.123, SD = 0.052) than in the reading task (M = 1.103, SD =
0.045, t(27)= -2.06, p <.05, r =.136). There was also a significant interaction between
task and item-type (F(1,27) = 21.11, MSE < 0.001, p <.001, ηp2 =.44) – PDR for
correctly identified old items (M = 1.141, SD = 0.056) was significantly larger than for
correctly identified new items (M = 1.116, SD = 0.046) in the recognition tasks (t(27)=
5.31, p <.001, r =.510); however there was little difference in PDR for correctly
identified old (M = 1.120, SD = 0.047) and new items (M = 1.122, SD = 0.047) in the
reading tasks (t(27)= -0.361, p >.05, ns) – in other words there was a pupil old/new
effect in the recognition conditions but not in the reading conditions. There was no
significant interaction between duration and item-type (F(1,27) = 0.31, p >.05 ns) and
the three-way interaction was not significant (F(1,27) = 1.17, p >.05, ns; see Figure
3-10).
136
Old
New
Average Pupil Dilation Ratio
p < .01
1.16
p < .001
1.15
1.14
1.13
1.12
1.11
1.1
1.09
Short Reading
Short Recognition
Long Reading
Long Recognition
Condition
Figure 3-10: Pupil dilation ratio for correctly identified old and new items in all four conditions. Error
bars show standard error of mean.
3.3.3.
Discussion
Experiment 2 systematically contrasted the effects of presentation duration (short vs.
long) and task (reading vs. recognition) on pupil-size when participants were
presented with old and new items. As predicted, in the long duration recognition
condition participants’ maximum pupil-size was significantly larger for correctly
identified old items compared to correctly identified new items. Interestingly, the
PONE was also present at short duration presentations, even though recognition
performance was significantly worse. This old/new effect was not observed for the
reading conditions, either at short or long duration. Taken together these findings
suggest that the reversed effect in the implicit condition of Experiment 1 was most
likely due to differences in task demands (reading vs. recognition), not presentation
duration – the PONE occurs whenever participants are asked to make a recognition
judgement on a word, even when presented very briefly, but is not present when
participants are asked to read a word out loud without making a recognition judgment,
even when that word is present for a long duration.
Although recognition performance was worse than performance on the reading task,
recognition rates for old items were comparable across short and long durations, but
137
recognition rates for new items were impaired in the short duration compared to the
long duration recognition condition. However, overall rates of correctly recognising
old items as old were poor with 6 participants performing below chance in the short
recognition condition and, perhaps more surprisingly, 14 participants in the long
recognition condition, 3 of whom performed poorly in both recognition conditions.
There are many factors affecting recognition memory, including age, attention and
context. Although condition order was rotated across participants, one possible
explanation of the poor recognition performance in the present experiment could be a
build up of retroactive interference across word lists, as by the fourth condition
participants had been asked to remember a large number of items. This is a
disadvantage of the within-participants design employed.
Another interesting finding to emerge from Experiment 2 was that overall PDR was
greater when stimuli were presented for a short duration. A possible explanation for
this finding might be that the short durations increased the cognitive effort required to
perform the recognition/reading tasks. Increased “cognitive load” could also account
for the larger pupil-sizes seen in the implicit condition of Experiment 1, when stimuli
were presented for a short time. However it is important to note that the PONE is not
overwhelmed by effort-related increases in pupil-size due to increased task difficulty
because it is still present in the short duration recognition task.
3.4. General Discussion
The Experiments in Chapter 3 sought to replicate the PONE – the relative increase in
pupil-size that occurs when participants view previously learned items during a
recognition test compared to novel items, and to build on Otero et al.’s (2011) finding
of a pupil-size difference between recollection and familiarity ratings, by asking
138
whether the PONE still occurs when stimuli are presented too briefly to evoke
conscious recollection, but may nonetheless reveal effects of prior learning.
The evidence from Experiments 1 and 2 replicates the PONE demonstrated by
previous research (see Chapter 1, section 1.3.2.3), and shows that maximum pupilsize is larger when participants encounter previously studied items, compared to new
items, when carrying out an explicit recognition memory task. Importantly Experiment
2 extends this finding to show that this pupillary old/new effect is a function of task
demands (recognition memory), as opposed to being an automatic consequence of
being exposed to items previously encountered during a learning phase (as in the
reading conditions), and also occurs for stimuli presented for brief durations. The
problems identified in Experiment 1, such as participants performing at ceiling in the
implicit condition, were addressed in Experiment 2. A better comparison of task
demands at short and long durations was allowed by the introduction of a short
duration reading condition.
The LOP manipulation in Experiment 1 had the expected effect on performance,
increasing successful recognition for deeply encoded (semantic) items compared to
shallowly encoded (surface) items in the explicit condition, but having no effect on
performance in the implicit condition. When old responses were collapsed over
surface and semantic items, more new than old items were correctly identified in the
explicit condition, and more old than new items were correctly identified in the implicit
condition, suggesting that perceptual fluency may enhance recognition performance
(e.g., Jacoby, & Dallas, 1981). This was also the case in Experiment 2.
The results of the LOP manipulation suggest that, contrary to Võ et al.’s (2008)
cognitive effort explanation of the PONE, items which had been deeply encoded
produced a larger pupil-size than shallowly encoded items (albeit at trend level),
139
which in turn produced a larger pupil-size than new items in the explicit condition.
This pattern of results is similar to Otero et al.’s (2011) findings of a graded pupil-size
for LOP and supports their suggestion that instead of a difference in cognitive effort
between old and new items, pupil-size may reflect memory “strength” – the PONE is
larger for deeply encoded items because they evoke a stronger memory.
Interestingly, rather than the standard PONE, the implicit condition revealed the
reverse – larger pupil-size for new items compared to old items, a finding that is
similar to those of Laeng et al. (2007) who looked at implicit memory in amnesic
patients and found larger pupil-sizes to novel stimuli. The difference was driven by
surface items, which had the smallest pupil-size, with semantic items intermediate but
not significantly different to either new or surface items. A possible explanation of this
pattern of results is that a pupil orienting response to novel items (e.g., Lynn, 1966;
Nieuwenhuis et al., 2011a; Pavlov, 1927; Sokolov, 1963) was no longer masked by
the PONE when presentation was too brief to elicit conscious recollection of stimuli.
An alternative explanation is that the increased pupil-size to new items reflects the
increased difficulty of the short duration, relative to old items which were made easier
to read through perceptual fluency. As it was difficult to tell whether task demands or
presentation duration lead to the results in the implicit condition, a control reading
task was carried out with the duration of the explicit condition, but where participants
were simply required to read the items out loud as in the implicit condition. In this
task, performance and pupil-size was equal for old and new items, and it was not
possible to draw any conclusions. Experiment 2 was designed as a more complete
orthogonal comparison which allowed the effects of presentation duration (2000ms
vs. 33ms) and task (reading vs. recognition) to be separated.
140
The key finding from Experiment 2 was that the PONE was present in both duration
recognition conditions, but was absent from both duration reading conditions.
Therefore the PONE occurs whenever participants are asked to make a conscious
recognition judgement on a word, even when presented very briefly, but is not present
when participants are asked to read a word without making a recognition judgment,
even when that word is present for a long duration. These findings may provide more
evidence in support of Otero et al.’s (2011) memory strength explanation of the
PONE, rather than Võ et al.’s (2008) cognitive effort explanation, because the PONE
is not overwhelmed by effort-related increases in pupil-size due to increased task
difficulty in the short conditions – there were separate main effects of presentation
duration (effort) and item-type (memory strength) on PDR, with no duration by itemtype interaction. The decrease in presentation duration also had more of an effect on
pupil-size in the reading conditions than in the recognition conditions.
The results of Experiment 2 suggest that the reversed pupil effect in the implicit
condition of Experiment 1 was most likely due to task demands (reading), rather than
presentation duration, but it is unclear why this was not replicated in the short reading
condition of Experiment 2. An explanation may lie in part in the fact that the
presentation duration was customised for participants in Experiment 2 such that they
achieved ~60% correct identification of items, whereas in Experiment 1 participants
were able to correctly identify 89% of items. Taken together, Experiments 1 and 2
suggest that the changes in pupil-size that occur, when participants encounter
previously studied items, reflect neurocognitive processes associated with making a
recognition judgement that do not occur when simply reading, even within the context
of a ‘memory experiment’. Previous research has shown that pupil-size is affected by
explicitly making a decision in other types of study (see Chapter 1, section 1.2.3.6).
141
However at short duration the PONE was only apparent when analysis is limited to
correctly identified items. In this condition, it was not possible to determine which
items were presented at the subjective threshold of conscious awareness
(Cheesman, & Merikle, 1984) and their relationship with pupil-size. It is also not
possible to exclude the possibility that the perceptual recognition task was
contaminated with recollective processes in the short reading condition, with some
participants anecdotally reporting conscious recognition of stimuli.
To address these issues, Chapter 4 adopts a widely used implicit learning procedure
in order to establish whether the PONE occurs when participants encounter letter
strings that either do or do not conform to a learned artificial grammar. The
advantage of this approach is that whilst for half of the strings, the letter order follows
the same artificial grammar rules as the strings encountered in the learning phase, at
recognition all letter strings are novel and have not been previously encountered.
142
4. The Role of Conscious Awareness in the PONE
Implicit Grammar
4
As argued in the discussion of Experiments 1 and 2, limiting presentation to a brief
duration during recognition was not particularly successful for determining whether
the PONE is also present when memory is tested implicitly. Stimuli in Experiment 2
were presented for a duration derived for each participant individually as the number
of monitor refreshes at which ~60% of practice items could be read. It was hoped
that this duration would produce a significant proportion of words at the subjective
threshold (Cheesman, & Merikle, 1984), eliciting the feeling of guessing at the actual
point of recognition, but a performance significantly above chance. However, as
some participants reported explicit awareness of some of the recognition items, it was
not possible to draw conclusions about implicit recognition.
There is debate in the literature regarding the criteria for “implicitness”. Shanks and
St John (1994) argue that knowledge elicited by cued recall and forced choice tests is
not implicit, given that it is far too difficult to be sure that you have excluded explicit
influences on a task that implicitly tests explicitly learned knowledge. A more
“process pure” method than implicit tests of memory, which researchers have
employed, is to look at implicit learning using paradigms such as Artificial Grammar
Learning (AGL). The literature demonstrates that implicit learning is also very
complex and difficult to define simply as, for example, learning without conscious
awareness, because consciousness and awareness are also very complex and
subjective concepts that are problematic to define or measure (Cleeremans, &
Dienes, 2008), for example, Frensch (1998) describes eleven different definitions of
implicit learning.
143
Berry and Dienes (1993) argue that implicit learning is unintentional, happening
without conscious awareness, and that the resulting knowledge is relatively
inaccessible with free report. Explicit learning, in contrast, is usually hypothesisdriven (an attempt to define, test and refine rules or concepts whilst learning) and
happens with conscious awareness of both the learning experience and the
knowledge acquired (see also Cleeremans, & Dienes, 2008; Dienes, & Berry, 1997).
Research has shown that implicit learning is associated with attention to stimuli rather
than on underlying rules, and is more robust to neuropsychological impairment (e.g.,
Knopman, & Nissen, 1987; Nissen, & Bullemer, 1987; Nissen, Willingham, &
Hartman, 1989; Schacter et al., 1988). Knowlton, Ramus and Squire (1992) found
that amnesic participants were able to classify new items as well as control
participants (63% and 67%, respectively), but were less able than controls to correctly
recognise old items (62% and 72%, respectively). Individual differences, such as age
and IQ, have less of an effect on implicit learning than explicit learning (e.g., Cherry,
& Stadler, 1995; Frensch, & Miner, 1994; Howard, & Howard, 1989; 1992; Myers, &
Connor, 1992). Reber, Walkenfeld and Hernstadt (1991) showed that AGL
performance was less correlated with IQ than performance on a problem-solving task.
The first AGL experiments used to investigate implicit learning were carried out in the
1960s by Reber (1967; 1969; 1989). Participants were shown a series of letter
strings that obeyed a 5-letter finite-state artificial grammar (a network with a finite
number of rules/paths that can be followed from entry to exit, see Figure 4-1). They
were then asked to identify ‘grammatical’ from ‘non-grammatical’ strings in a
‘recognition’ phase. Importantly, the grammatical strings in the recognition phase
were not the same strings that participants were exposed to in the learning phase –
they simply followed the same artificial grammar rules. Participants correctly
144
identified 69% of the previously unseen grammatical strings, which Reber (1967)
proposed as evidence of implicit learning of the underlying grammatical rules. Reber
reported that participants emerged with some knowledge but were not able to explain
it fully. This was more formally tested by Dienes, Broadbent and Berry (1991) who
asked participants to describe rules or strategies that could be used by someone who
had not seen the stimuli. Using those rules three independent judges rated each
string as grammatical or ungrammatical to simulate a classification performance of
54%. This result was significantly less than actual classification performance (65%)
indicating that participants’ free-report of learned knowledge was impoverished
compared to their ability to apply it.
Figure 4-1: The two finite state grammars used to generate the strings (from Reber, 1969).
Since Reber’s original experiments, AGL has become a widely used technique for
investigating implicit learning, and his findings have been widely replicated. Several
different accounts have been proposed for the basis of above chance classification
judgements including: abstracting rules about the underlying grammar; memorising
whole strings or fragments of the training stimuli (chunking; Wickelgren, 1979); and
learning the statistical relationships between fragments of the strings such as in
connectionist models (see Pothos, 2007, for a comprehensive review).
Dienes and Berry (1997) argue that the learning of underlying correct classification of
grammatical strings in AGL experiments is implicit because, unless participants are
145
instructed to look for rules, learning occurs without intent or conscious awareness,
and participants perform at above chance whilst not being able to verbalise how they
are making their decisions. Confidence in grammatical judgements is also often
unrelated to performance, indicating an absence of metaknowledge (Chan, 1992;
Dienes et al., 1995; Dienes, & Altman, 1997).
Previous research and connectionist modelling have indirectly suggested that
subjective feelings of familiarity may be central to implicit learning (e.g. Norman,
Price, Duff, & Mentzoni, 2007; for a review see Cleeremans, & Dienes, 2008).
However, there had been no direct tests of this hypothesis; Scott and Dienes (2008)
provided the first direct evidence that structural similarities in grammatical as opposed
to non-grammatical letter strings (such as fragment frequency and repetition
structure) are experienced subjectively as familiarity, by examining how grammatical
judgements and confidence related to participants’ reports of familiarity. In a series of
AGL experiments they asked participants to state whether each test string was
grammatical, how certain they were in their decision, how they made their decision,
and how familiar the string felt to them. Participants were able to correctly identify
around 60% of the grammatical strings – significantly above chance. Importantly,
participants’ subjective familiarity ratings for strings predicted their grammaticality
judgements and the degree to which they were correct in their response – the more
familiar a string felt, the more likely participants were to rate it as grammatical, and
the more likely they were to be correct in their judgement, even in the absence of
confidence in their response (Scott, & Dienes, 2008).
As discussed in section 1.3.1.2 of the introductory chapter, some current models of
recognition memory argue that studied items can be recognized as old on the basis of
two separate processes – a conscious recollective process that involves retrieval of
146
specific contextual information concerning the original learning experience
(sometimes represented by a “remember” or “R” response) and a familiarity based
process that provides a context-free sense that an item has been previously
encountered (a “know” or “K” response; e.g. Yonelinas, & Jacoby, 1996). Many
studies show that the two processes can be differentiated behaviourally and these
findings are used to support the argument that they have different underlying neural
mechanisms (see Yonelinas, 2002, for a review). A further line of evidence
supporting dual-process models is the ERP old/new effect in which specific
neuroelectric signals appear to be linked to R/K responses (e.g., Curran, 2004;
Duarte et al., 2004; Düzel et al., 1997; Rugg, Schloerscheidt, & Mark, 1998; Trott,
Friedman, Ritter, Fabiani, & Snodgrass, 1999; Wolk et al., 2006; see Chapter 6,
section 6.1.1).
An obvious question is therefore whether similar differentiation occurs for the PONE –
in other words does it differ in magnitude between items that are consciously
recollected compared to those that are recognised on the basis of familiarity alone. If
it does, then any explanation of the PONE in recognition memory must therefore take
account of both recollection and familiarity processes and the ways in which they
dissociate experimentally. Previous research into the PONE (see Chapter 1 section
1.3.2.3 and Chapter 3 section 3.4) has not yet answered this question. Using a
remember-know paradigm Otero et al. (2011) found that the PONE was larger when
participants responded “remember” (recollection), compared to when they responded
“know” (familiarity). Critically, the PDR for “know” items was greater than the PDR
when participants viewed novel items (although this difference was only significant at
a trend level). Otero et al.’s (2011) suggestion that the PONE reflects an aggregate
strength of memory signal, based on both familiarity and recollective processes, was
prompted by recent accounts of recognition memory that reject the idea that
147
recollection, when it occurs, is ‘all or nothing’, and that familiarity processes are only
important in the absence of recollection (Wixted, 2007a; Wixted, & Stretch, 2004).
Instead these accounts suggest that both recollective and familiarity signals vary on a
continuum, and items are judged to be old when an aggregate signal exceeds a
certain threshold. Items that are recognized on the basis of familiarity, in the absence
of conscious recollection, would still have a greater overall strength of memory signal
than new items, but the strength of memory signal would be considerably weaker
than the signal associated with items that are recollected.
Experiment 3 aims to replicate the previous findings of increased PDR for old
compared to new words in an explicit recognition condition. Additionally, Experiment
3 aims to further investigate the mnemonic processes associated with pupil dilation by
measuring pupil-size in an implicit recognition condition using AGL, proposing that for
the implicit condition, conscious recollection will not be available as both the
“grammatical” and “nongrammatical” strings presented in the recognition phase will
be different to those presented in the learning phase. Previous research (Reber,
1967; 1969; Scott, & Dienes, 2008; 2009) has indicated that implicit learning of
grammatical letter strings evokes a greater sense of familiarity than non-grammatical.
This would suggest that familiarity alone is enough to elicit a memory signal
exceeding recognition threshold. It was predicted that a PONE would occur in the
implicit condition, reflecting familiarity signal strength, but that this effect would be
smaller than for the explicit test of memory.
148
4.1. Experiment 3 – Implicit Grammar
4.1.1.
Method
4.1.1.1.
Participants
Twenty-four participants (8 male; age range: 19.0-36.2, M = 24.6, SD = 5.29) with
normal or corrected-to-normal vision were recruited from the psychology course-credit
and subject pools at the University of Sussex, and through personal contact.
Participants were briefed with a detailed consent form (specific to the condition to
which they were allocated) and verbal description, and invited to ask questions. To
gain informed consent to participate, without deception or revealing the artificial
grammar, participants were informed that they would be presented with letter strings
in the learning phase of that test, and would have to make ‘yes/no’ decisions on 60
letter strings based on the strings presented in the learning phase. Participation was
on a voluntary basis, and participants were thanked and debriefed with a verbal
description of the aims of the study and the opportunity to ask any further questions
after completing the study. The experiment was approved by the relevant ethics
committee.
4.1.1.2.
Materials/Apparatus
For the explicit condition two 30 item word lists (A and B; see appendix C) were
created using nouns selected from the MRC Psycholinguistic Database. All items
were 7 letters long, and lists were matched for familiarity (list A range = 420-613, M =
521; list B range = 442-598, M = 524), imageability (list A range = 274-593, M = 425;
list B range = 258-591, M = 410) and frequency (list A range = 21-96, M = 53; list B
range = 22-98, M = 52), according to the K-F norms. Words likely to elicit a strong
emotional response were removed. Items were presented using the Experiment
149
Builder software associated with the EyeLink II eye-tracker (SR-Research, Ontario) in
black 20pt Arial font in the centre of a light grey background under fixed illumination.
During the learning phase participants saw either list A or B, and during the
recognition phase participants saw all items from both lists in a randomised order.
Stimuli for the implicit condition were taken from Scott and Dienes (2008) who used
two finite-state grammars from Reber (1969) to generate 45 grammatical strings, 5-9
letters long, from each grammar using the letters M, R, T, V, and X and the same set
of valid starting bigrams and final letters. 15 strings from each grammar were
repeated three times in a random order to create two 45 item learning lists (A and B;
see appendix C). The remaining 30 items from each grammar were combined in a
random order to create the recognition list. Lists were matched for string length and
Scott and Dienes (2008) carried out statistics to ensure the structural similarity of
learning strings to recognition strings. Items were presented using Experiment
Builder (SR-Research, Ontario) in black 20pt Arial font in the centre of a light grey
background under fixed illumination. During the learning phase participants saw
either list A or B, and during the recognition phase participants saw new strings, 30
items generated from each grammar, in a randomised order, a total of 60 items.
Throughout the recognition phase of both conditions, pupil-size was recorded using
an EyeLink II head-mounted eye-tracker with a temporal resolution of 2ms and a
spatial resolution of around 0.25 degrees. The stimuli were displayed on a 21 inch
CRT monitor with a screen resolution of 1,280 x 1,024 pixels and a refresh rate of
60Hz. Actual screen dimensions were 40cm horizontal and 30cm vertical.
Participants were seated approximately 70cm from the screen in an adjustable chair
that had been modified to prevent any rotational movement. Responses during the
150
recognition phase were made using the left and right triggers of a computer gamepad
and recorded by the host PC.
4.1.1.3.
Design and Procedure
The study employed a 2 (recognition test: explicit vs. implicit) x 2 (item-type: old vs.
new) within-subjects design. All participants completed both the implicit and explicit
conditions, which each contained a learning and recognition phase. Participants
completed either the implicit or explicit condition first, and learned either list A or B
(independently) in each condition. This gave a total of 8 different combinations (AA,
AB, BB, AA x explicit first, implicit first) which were rotated across participants to
avoid practice and list effects. Participants were briefed on arrival with general
instructions and asked to complete a consent form. Further instructions specific to
condition were displayed on a computer screen as they completed the tasks.
Explicit Condition
The experimenter selected which list (A or B) should be used for the learning phase.
On-screen instructions informed participants that they would see a set of 30 words
and asked them to try to remember them. The 30 learning items were presented one
at a time in the centre of the screen for 2000ms. Participants were then fitted with the
eye-tracker and performed a short 3 point calibration task. During the recognition
phase, participants were presented with 60 items (lists A and B in a randomised
order), 30 of which had been presented in the learning phase (old items), the other 30
were from the list that had not been learned (new words). At the start of each trial, a
drift-correction dot was presented in the centre of the screen, followed by a mask
“&&&&&&&” of 7 ampersands for 500ms, and then a recognition item for 2000ms.
Masks and recognition items were presented in the same size and font as items in the
151
learning phase. Participants were asked to state whether the item was old
(previously encountered in the learning phase) or new (not previously encountered),
prompted by a screen with the word “old” on the left and “new” on the right,
corresponding to the buttons on the gamepad. This screen remained visible until the
participant made a decision, at which point the screen was replaced by a driftcorrection dot before presentation of the next item. Old/new judgements were
recorded on the computer after each recognition item.
Implicit Condition
The experimenter selected which list (A or B) should be used for the learning phase.
Participants were not informed that there were rules to the letter strings during the
learning phase. Onscreen instructions informed participants that they would see 45
letter strings one at a time, each on screen for 5000ms with a 5000ms gap in between
to encourage attention to the unfamiliar stimuli, during which they should write down
as much of the string as they could remember. No indication was given that they
would be learning an artificial grammar. The eye-tracker was then fitted followed by a
3 point calibration task. During the recognition phase, participants were presented
with 60 new letter strings, 30 from grammar A and 30 from grammar B. No strings in
the recognition phase had been presented in the learning phase so could not be
explicitly recognised. At the start of each trial, a drift-correction dot was presented in
the centre of the screen, followed by a mask “&&&&&&&” of 7 ampersands for 500ms,
and then a recognition item for 2000ms. Masks and recognition items were
presented in the same size and font as items in the learning phase. Participants were
asked to state whether the item was grammatical, prompted by a screen with the
word “yes” on the left and “no” on the right, corresponding to the buttons on the
gamepad. This screen remained visible until the participant made a decision, at
152
which point the screen was replaced by a drift-correction dot before presentation of
the next item. Yes/no judgements were recorded on the computer after each
recognition item. For ease of analysis, strings that obeyed the artificial grammar are
referred to as ‘old’ items and non-grammatical strings are referred to as ‘new’ items.
4.1.1.4.
Pupil Recording
Maximum pupil-size was recorded from the right eye during each recognition period.
A Pupil Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated
expressing the maximum pupil-size for each 2000ms recognition trial as a proportion
of the maximum pupil-size during that trial’s 250ms baseline.
4.1.2.
Results
4.1.2.1.
Behavioural Data
The proportion of correct responses to old and new items was calculated for each
condition. In the explicit condition, average hit rate (correctly identified old items) was
71.6%, average correct rejection rate (correctly identified new items) was 71.6%,
false alarm rate (incorrectly identified new items) was 28.4% and miss rate
(incorrectly identified old items) was 28.4%. Due to a programming error,
participants’ responses in the implicit condition were not recorded and the
percentages could not be calculated.
4.1.2.2.
Pupil-Size Data
Average PDR for old and new items was calculated for both conditions. As PDR is a
function of baseline pupil-size, baseline pupil-sizes for old and new items in each
condition were compared to ensure that differences in PDR were not due to baseline
differences. The difference was not significant (F(1,23) = 0.801, p >.05, ηp2 =.034).
153
A 2 x 2 repeated measures ANOVA of PDR with within-subject factors of item-type
(old vs. new) and condition (implicit vs. explicit) revealed no significant main effect of
item-type (F(1,23) = 1.236, MSE <.001, p >.05, ηp2 =.05), but a significant main effect
of condition (F(1,23) = 68.00, MSE =.001, p <.001, ηp2 =.75) – across both old and
new items participants’ pupils dilated to a greater extent in the explicit condition (M =
1.132, SD = 0.046) than in the implicit condition (M = 1.071, SD = 0.028, t(23) = 8.25,
p <.001, r =.747). The interaction between condition and item-type failed to reach
significance (F(1,23) = 2.184, MSE < 0.001, p =.153, ηp2 =.087), however, as is clear
from Figure 4-2, there appears to be a difference between PDR for old and new items
in the explicit but not implicit condition. A priori t-tests show that in line with previous
findings PDR to old items (M = 1.137, SD = 0.048) in the explicit condition is larger
than PDR to new items (M = 1.127, SD = 0.048, t(23) = 2.024, p < 0.05, r =.151;
Figure 4-2).
Old
New
Average Pupil Dilation Ratio
p = .028
1.16
1.14
1.12
1.10
1.08
1.06
1.04
Explicit
Implicit
Condition
Figure 4-2: Pupil dilation ratio for old and new stimuli in both conditions. Error bars show standard
error of mean.
4.1.3.
Discussion
Experiment 3 replicated the PONE in the explicit recognition condition – participants’
pupil-sizes increased to a greater extent when they viewed old items compared to
154
new items. The other key finding from this experiment is that there was no PONE in
the implicit condition – average pupil dilations for grammatical and ungrammatical
strings were the same. This suggests either that the PONE may be associated with
recollection but not familiarity, or that insufficient levels of familiarity were generated
by the grammatical strings to produce a measurable increase in PDR compared to
non-grammatical strings.
However, the absence of behavioural data for the implicit condition makes further
interpretation of the data difficult as it is not possible to restrict the analysis to
correctly identified old and new items, which are presumed to be the best attended to
stimuli and most likely to be engaging recognition memory processes. It is also not
possible to determine whether the experimental design was effective and implicit
learning took place, because we don’t know whether “old” grammars were correctly
judged to be grammatical at a level significantly above chance. Therefore the
experiment was re-run with 23 participants, using the same stimuli and design.
4.2. Experiment 4 – Implicit Grammar Replication
4.2.1.
Method
4.2.1.1.
Participants
Twenty-three participants (3 male; age range: 18.9-49.1, M = 21.9, SD = 6.06) with
normal or corrected-to-normal vision were recruited from the psychology course-credit
and subject pools at the University of Sussex, and through personal contact.
4.2.1.2.
Materials/Apparatus/Design/Procedure
As detailed for Experiment 3.
155
4.2.2.
Results
4.2.2.1.
Behavioural Data
The proportion of correct responses to old and new items was calculated for each
condition. In the explicit condition, average hit rate (correctly identified old items) was
74.6%, and average correct rejection rate (correctly identified new items) was 72.8%.
Examination of implicit data reveals that participants performed significantly above
chance when correctly judging old strings as old (61.9%, t(22) = 3.24, p <.01, r =.32)
and new strings as new (64.3%, t(22) = 4.10, p <.001, r =.43), suggesting that
participants were able to learn the artificial grammar to some extent (see Table 4-1).
Explicit
Implicit
M
SD
M
SD
Hits
.746
.115
.619
.176
Correct Rejections
.728
.136
.643
.168
False Alarms
.272
.136
.355
.167
Misses
.254
.115
.381
.176
Table 4-1: Proportion of stimuli judgments for both conditions.
The proportion of correctly identified items was calculated for each condition. A 2 x 2
repeated-measures ANOVA with within-subject factors of item-type (old vs. new) and
condition (implicit vs. explicit) showed a significant main effect of condition (F(1,22) =
8.54, MSE = 0.03, p <.01, ηp2 =.28) – on average, participants identified more items
correctly in the explicit than in the implicit condition – but no significant main effect of
item-type (F(1,22) = 0.011, MSE = 0.016, p >.05, ηp2 =.001) and no significant
interaction (F(1,22) = 1.24, MSE = 0.009, p >.05, ηp2 =.054; see Figure 4-3).
156
Proportion of Correct Responses
Old
New
0.80
0.75
0.70
0.65
0.60
0.55
Explicit
Implicit
Condition
Figure 4-3: Proportion of correct responses. Error bars show standard error of mean.
4.2.2.2.
Pupil-Size Data
Average PDR for old and new items was calculated for each condition. As PDR is a
function of baseline pupil-size, baseline pupil-sizes for old and new items in each
condition were compared to ensure that differences in PDR were not due to baseline
differences. The difference was not significant (F(1,22) = 0.943, p >.05, ηp2 =.041).
A 2 x 2 repeated-measures ANOVA of PDR for all trials with within-subject factors of
item-type (old vs. new) and condition (implicit vs. explicit) revealed a significant main
effect of item-type (F(1,22) = 5.34, MSE <.001, p <.05, ηp2 =.195) – participants’
pupils dilated to a greater extent to old items (M = 1.066, SD = 0.034) than to new
items (M = 1.061, SD = 0.033, t(22) = 2.31, p <.05, r =.195). The main effect of
condition was also significant (F(1,22) = 28.2, MSE =.001, p <.001, ηp2 =.562) –
participants’ pupils dilated to a greater extent in the explicit condition (M = 1.083, SD
= 0.046) than in the implicit condition (M = 1.044, SD = 0.026, t(22) = 5.31, p <.001, r
=.562).
157
Average Pupil Dilation Ratio
Old
New
p = .021
1.11
1.10
1.09
1.08
1.07
1.06
1.05
1.04
1.03
1.02
Explicit
Implicit
Condition
Figure 4-4: Pupil dilation ratio to old and new items in each condition. Error bars show standard error
of mean.
The interaction between condition and item-type just failed to reach significance
(F(1,22) = 3.13, MSE < 0.001, p =.09, ηp2 =.125). However, t-tests show that in the
explicit condition PDR to old items (M = 1.089, SD = 0.048) is larger than PDR to new
items (M = 1.076, SD = 0.046, t(22) = 2.49, p <.05, r =.220) whereas this difference is
not significant in the implicit condition (old: M = 1.044, SD = 0.027; new: M = 1.045,
SD = 0.029; t(22) = -0.193, p >.05, r =.002; see Figure 4-4).
The analysis was repeated, but restricted to items that were correctly identified. The
main effect of condition remained significant (F(1,22) = 25.2, MSE = 0.001, p <.001,
ηp2 =.534) – average PDR was larger in the explicit condition than in the implicit
condition. The main effect of item-type also remained significant (F(1,22) = 5.04,
MSE < 0.001, p <.05, ηp2 =.186) – participants’ pupils dilated to a greater extent to old
stimuli than for new stimuli.
Importantly the item-type by condition interaction reached significance (F(1,22) =
9.03, MSE < 0.001, p <.01, ηp2 =.291) – the difference in PDR between correctly
identified old and new items in the explicit condition (M = 0.022, SD = 0.033) was
larger than the difference in PDR between correctly identified old and new items in
158
the implicit condition (M = -0.003, SD = 0.022, t(22) = 3.00, p <.01, r =.290) – the
PONE was present in the explicit but not the implicit condition (Figure 4-5).
Average Pupil Dilation Ratio
Old
New
p = .007
1.12
1.11
1.10
1.09
1.08
1.07
1.06
1.05
1.04
1.03
1.02
Explicit
Implicit
Condition
Figure 4-5: Pupil dilation ratio to correctly identified old and new items in each condition. Error bars
show standard error of mean.
4.3. General Discussion
Experiments 3 and 4 sought to determine whether the relative increase in pupil-size
that occurs when participants view previously learned items during a recognition test,
compared to novel items, also occurs when participants make “old/new” decisions
that reflect implicit learning as opposed to explicit recognition of previously
encountered items. The key finding is that whilst the PONE was again observed in
the explicit recognition conditions in both experiments, there was no evidence for any
increase in pupil-size when participants were exposed to grammatical compared to
ungrammatical letter strings.
Behavioural measures showed that in both conditions participants were attending to
and learning stimuli reasonably well. Across the explicit conditions of Experiments 3
and 4 participants were able to correctly identify 72.8% of old words, and to correctly
judge 72.4% of new words, consistent with the literature (e.g., Yonelinas, 1994).
Behavioural measures in Experiment 4 showed that participants were attending to the
159
stimuli and had been able to learn the artificial grammar to some extent, they
performed above chance when correctly judging 61.9% of old strings and 64.3% of
new strings, consistent with the degree of learning shown using these stimuli in recent
studies (cf., Scott, & Dienes, 2008; 2009).
The PONE was not present in the implicit condition of either experiment – average
pupil dilations for grammatical and ungrammatical strings were the same. This was
the case even when the analysis was restricted in Experiment 4 to those items that
participants successfully identified as grammatical. Artificial Grammar Learning
(AGL) was used in order to investigate the PONE in a recognition situation in which
familiarity, but not recollective processes were involved. Scott and Dienes (2008)
showed that participants make “old” decisions in implicit grammar learning tasks
based on a sense familiarity, and that stronger perceived familiarity ratings lead to
more correct hits. They suggest this may reflect unconscious recognition of the
stimuli’s features as a result of implicit learning during the learning phase. The strings
presented in the recognition phase of AGL tasks were all new, therefore no recall of
individual whole strings should occur (although the possibility that small fragments are
explicitly recognised remains).
According to the signal-detection unequal-variance theory, both familiarity and
recollection can vary in strength and the greater their combined strength, the greater
the memory signal. Recognition results only if this signal exceeds a recognition
threshold (Wixted, 2007a; Wixted, & Stretch, 2004; Otero et al., 2011). If, as Otero et
al. (2011) have argued, the PONE reflects an aggregate of both recollection and
familiarity signals, the PDR should be greater for grammatical strings than nongrammatical strings, due to a greater sense of familiarity in the absence of conscious
recognition (Scott, & Dienes, 2008) for these items. The absence of the PONE in the
160
implicit condition suggests that it may reflect recollective but not familiarity processes.
This suggests that genuine amnesiacs may not show the PONE in the absence of
explicit memory. Alternatively, the PONE might reflect both recollection and
familiarity signals, but in the absence of recollection in the implicit condition the level
of familiarity achieved was insufficient to produce a measureable effect on pupil-size.
As discussed in Chapter 3, pupil response has already been found to indicate intact
implicit memory in amnesic patients. Our results suggest that the increase in pupilsize that occurs when participants encounter correctly recognised old items reflects
neurocognitive processes associated with explicit, but not implicit, recognition
memory. This could indicate that the pupillary response is a function of task demands
(recognition memory test) as opposed to being the automatic consequence of being
exposed to items previously encountered during a learning phase (confirming the
findings of Experiments 1 and 2). The larger PDR which occurs to new items in the
explicit condition compared to ‘new’ (ungrammatical) items in the implicit condition
may occur because the grammatical decision in the implicit condition does not involve
recognition memory processes, whereas the correct rejection of a new item in the
explicit condition may involve an effortful memory search, leading to effort-related
increases in pupil-size (e.g. Kahneman, 1973), and studies have found that rejection
typically takes longer than correct recognition (e.g., Ratcliff, & Murdock, 1976).
In Chapter 5, the issue of whether the PONE reflects automatic or voluntary
mnemonic processes is explored further in a series of experiments that seek to
establish the extent to which it is associated with item-type (old vs. new) or
participant response (old vs. new).
161
5. Malingering and the Old/New Response
Is the PONE Under Voluntary Control?
5
Building on the findings of Experiments 1 and 2, the results of Experiments 3 and 4
suggest that the increase in pupil-size that occurs when participants encounter
previously studied items, and recognise them as old, reflects neurocognitive
processes associated with explicit, but not implicit recognition memory. This could
indicate that the PONE represents processes involved in making an explicit
recognition memory judgement, for example a memory strength signal, which is larger
for old items than new items, and suggests that conscious awareness is an important
factor in the PONE. Therefore, in Chapter 5 this idea is extended to explore whether
the PONE reflects voluntary mnemonic processes, in three experiments that seek to
establish the extent to which it is associated with item-type (old vs. new) or participant
response (old vs. new). Experiments 5 and 6 follow the logic of studies that have
attempted to exploit the ERP old/new effect as an index of feigned amnesia
(malingering), to investigate the effect of asking participants to deliberately give wrong
answers during the recognition phase. Experiment 7 attempts to genuinely impair
recognition memory performance in healthy participants by dividing attention.
5.1. Malingering and Deception
5.1.1.
Background
The nature of deception and how to detect it has long concerned scientists,
philosophers, clinicians and forensic professionals. There is also a significant
commercial interest in creating a simple, reliable, and accurate evidence-based
method of lie-detection that could be used to identify individuals attempting to deceive
162
or malinger for personal or financial gain, or to evade criminal prosecution. The
ultimate goal of lie-detection is a method that doesn’t rely on subjective assessment,
and is not subject to the same interpersonal manipulations as people.
A particularly complex type of lying is malingering, the deliberate fabrication or
exaggeration of physical or psychological symptoms for secondary gain, such as
financial compensation, evading a penalty or conviction, gaining drugs, or avoiding
military service (Hutchinson, 2001; Tugcu, 2010). Adults, adolescents and children
may feign a wide range of conditions including movement disorders, sensory issues,
epilepsy, loss of consciousness, and neurological deficits (Kasikci, & Bek, 2010). It
has been estimated that the incidence of malingering is approximately 12% in patient
populations, rising to 40% in patients seeking financial compensation (see
Hutchinson, 2001; Larrabee, 2003; Mittenberg, Patton, Conyock, & Condit, 2002).
There may be elaborate motivations behind malingering and, as Hutchinson (2001)
observes, the boundaries between exaggeration and fabrication, and deliberate and
unintentional malingering are not clear. For this reason, some believe that within a
biopsychosocial model, use of a value-laden term like “malingering” is prejudicial,
preventing best practice, and that patients with such “disorders of simulation” have a
genuine need for help (Hutchinson, 2001; Ucar, & Atac, 2010).
Malingering does not have well-established diagnostic criteria, primarily due to
clinicians’ reliance on patients’ self-reported symptoms (Hutchinson, 2001). This is
particularly difficult in cognitive neuropsychology, where there may be no objectively
measurable physical symptoms, and patients may fabricate language impairments
(e.g., Cottingham, & Boone, 2010), posttraumatic stress disorder (e.g., Merten, Thies,
Schneider, & Stevens, 2009), low intelligence, psychosis, amnesia, or affective
disorders (e.g., Slick, Sherman, & Iverson, 1999). Most conditions are assessed
163
through clinical interview and/or the administration of neuropsychological tests
designed to measure the degree of difficulty or impairment. Feigned poor
performance is difficult to detect, and presents a real challenge to practicing clinicians
(see Hutchinson, 2001). According to Binder and Rohling (1996), patients seeking
financial compensation following closed head injury perform more poorly on
neuropsychological assessments than do patients who are not seeking financial
reward by around one half of a standard deviation. It has been shown that effort and
cooperation, rather than brain injury, accounts for up to 50% of the variance in results
(Rohling, Green, Allen, & Lees-Haley, 2000). Mittenberg et al. (2002) have argued
that survey-based studies which attempt to identify the base-rate of malingering are
likely to underestimate the true rate, as the ability of clinicians to detect malingerers is
not 100%.
In response to these issues, psychologists have developed instruments which index
the likelihood of malingering, for example the Fake Bad Scale (FBS; Lees-Haley,
English, & Glenn, 1991) which relies on malingerers’ desire to appear healthy and
honest, and is sensitive to “illogical symptom histories” (Greiffenstein, Baker, Gola,
Donders, & Miller, 2002) or insufficient cognitive effort in litigants with mild traumatic
head-injury (Ross, Millis, Krukowski, Putnam, & Adams, 2004; Slick et al., 1996).
5.1.1.1.
Malingered Memory-Impairment
A commonly malingered cognitive deficit is memory-impairment (Rüsseler, Brett,
Klaue, Sailer, & Münte, 2008; Samuel, & Mittenberg, 2005). Memory is a key
cognitive function, and disruption to the memory systems impacts massively on a
person’s quality of life, which has implications for the types of service they receive
and, in the case of accidents, the amount of compensation. It is widely known to the
general public that memory-impairment is a frequent outcome of a head injury such
164
as concussion (Gouvier, Prestholdt, & Warner, 1988), and even patients with genuine
memory-impairment may exaggerate their symptoms when seeking compensation
(Yochim, Kane, Horning, & Pepin, 2010).
Therefore existing memory measures have also been adapted to measure the
likelihood of malingering, such as the expanded version of the Auditory Verbal
Learning Test (Barrash, Suhr, & Manzel, 2004) and the Wechsler Memory Scale
Third Edition (WMS-III; Wechsler, 1997). In addition specific malingering tests have
been developed, such as the Test of Memory Malingering (TOMM; Tombaugh, 1996),
Word Memory Test (Green, & Astner, 1995) and the Victoria Symptom Validity Test
(VSVT; Slick, Hopp, Strauss, & Thompson, 1997). These rely on patterns of
performance that are inconsistent with the performance of genuinely memoryimpaired individuals, for example poorer recognition performance than delayed recall,
and absence of memory effects such as the primacy effect (Barrash et al., 2004).
The WMS-III has a Rarely Missed Index in its scoring system, of items that are
unlikely to be forgotten by participants with genuine memory-impairment, classifying
the genuinely neurocognitively-impaired from those fabricating memory difficulties
with an accuracy of 98% (Wechsler, 1997), and head injured patients from those with
substance abuse with an accuracy of 95% (Miller, Ryan, Carruthers, & Cluff, 2004).
However, subsequent studies including genuine patients, participants exaggerating
symptoms, and head-injury litigants, found less desirable overall classification
accuracies of 75-89% (Lange, Senior, Douglas, & Dawes, 2003; Langeluddecke,
2004), specificities of 87-91% (correctly identifying non-malingers as genuine; Lange
et al., 2003; Lange, Sullivan, & Anderson, 2005; Swihart, Harris, & Hatcher, 2008),
and much lower sensitivities of only 18-25% (correctly identifying a malingerer as
such; Lange et al., 2005; Swihart et al., 2008). Axelrod, Barlow and Paradee (2010)
165
indicate that in the case of memory-impairment severe enough to cause random
responses, 69% of respondents would fall within the range of suboptimal effort and
may be falsely accused of malingering.
The task of assessing memory-impairment is even more of a challenge among
forensic populations (Denney, 2007), because when neuropsychological measures,
which have been standardised on a civil population, are used to assess criminal
populations, they show a negative bias toward a more impaired performance than
would be predicted by their general cognitive ability (Ardolf, Denney, & Houston,
2007; Franzen, & Iverson, 2000). It is estimated that 70% of forensic patients are
thought to adapt their presentation when assessed by a clinical neuropsychologist
(Heilbrun, Bennett, White, & Kelly, 1990), with 25-50% of murder and manslaughter
suspects claiming crime-related amnesia (Pujol, & Kopelman, 2003). Although some
genuine memory losses are reported for the period in which the crime was alleged,
due to intoxication, seizure, or sleep disorder (Bourget, & Whitehurst, 2007), in a
study of over 300 convicted prisoners, Cima, Nijman, Merckelbach, Kremer and
Hollnack (2004) found none whom they considered to have genuine trauma
dissociation amnesia, stating that such claims should be approached with caution.
Clearly, the severe implications of being wrongly convicted, or acquitted, of a crime
require accurate methods of detecting crime-related amnesia (Bourget, & Whitehurst,
2007; Merckelbach, & Christianson, 2007). A concerning example of the flaws in
neuropsychological tests comes from Galappathie and Vakili (2009), who report a
case study of a man attempting to evade a conviction for murder by feigning memoryimpairment. Mr X initially “passed” the TOMM, but six months’ observation showed
his abilities to learn new information and recall past information to be functioning well
above the level indicated by neuropsychological tests. He later “failed” a second
166
administration of the TOMM, was deemed to be making insufficient effort and his trial
was resumed. O’Bryant and Lucas (2006) showed that although the sensitivity of the
TOMM is 98%, its specificity is only 78%, meaning that only 3 out of 4 people who
“pass” the TOMM are genuinely making enough effort.
Unfortunately neuropsychological tests may also be vulnerable to countermeasures
(attempts by the individual to distort the results in their favour), such as coaching,
where participants are given detailed information about the symptoms of memoryimpairment or strategies to avoid detection (e.g., Bauer, & McCaffrey, 2006; Jelicic,
Merckelbach, Candel, & Geraerts, 2007). Although layperson ideas about amnesia
may be inaccurately based on examples from films (Baxendale, 2004), as tests
advance, so too does the sophistication of malingering strategies. In an arms race
with test developers, malingerers can use the internet to access up-to-date
information such as pass/suspicious/fail cut-off scores, details of the scoring system,
and example stimuli (Bauer, & McCaffrey, 2006). Differences in performance levels,
reaction times and responses to feedback between genuinely memory-impaired, nonmemory-impaired and malingering populations on a range of different tests can also
be researched. In this way potential malingerers can determine situations and tests
on which memory-impaired patients would actually be expected to perform well, such
as on the Amsterdam Short Term Memory test (ASTM; Schagen, Schmand, de
Sterke, & Lindeboom, 1997; Bauer, & McCaffrey, 2006; Jelicic, Merckelbach, Candel,
& Geraerts, 2007). Bauer and McCaffrey (2006) found that the TOMM was
particularly well documented on the internet and suggest that its simpler format may
be more conducive to malingering than tests that measure multiple domains or
include reverse-scored items of bizarre or atypical symptoms not usually associated
with genuine impairment, such as in the Structured Inventory of the Malingered
Symptomatology (SIMS; Smith, & Burger, 1997).
167
In short, most tests of malingering based on participants’ responses rely on
malingerers making too many errors and taking too long to respond compared to
genuinely memory-impaired individuals. As such, careful research about symptoms
and how the test works could result in an individual feigning believable poor
performance. Although the administration of several tests increases the likelihood of
detection, this is time consuming. Peter, Merten, Merckelbach and Oswald (2010)
found that after warning malingering simulators not to exaggerate symptoms, despite
the improved sensitivity and specificity of using multiple validated instruments, three
out of twenty malingerers were able to evade detection.
5.1.1.2.
Psychophysiological Detection of Deception
Memory is subjective, and behavioural responses rely on the ability and desire to
respond accurately. An obvious literature for neuropsychological professionals to turn
to when seeking solutions to problems of memory malingering is that of the
psychophysiology of deception detection. Psychophysiological techniques may offer
a more suitable way of assessing memory performance without reliance on self-report
(see Chapter 1, section 1.3.2).
The attempt to identify measurable psychophysiological markers of deception has a
long history. As long ago as 1000 BC physiological responses to stress were
documented as being used to identify liars (for a historical overview of the nature of
deception and lie-detection, see Ford, 2006). The first four-channel polygraph device
was developed in 1932 to assess heart rate, blood pressure, respiratory rate and
GSR (Saxe, Dougherty, & Cross, 1985). Typically this approach operates on the
basis that, for the majority of people, deception is generally more cognitively
demanding, and more anxiety provoking, than telling the truth. This results in an
involuntary difference in the psychophysiological response to incriminating stimuli
168
compared to non-incriminating stimuli in a guilty person. The polygraph works on the
assumption that changes in these autonomic responses indicate dishonesty. It does
not take into account other influences, such as emotional upset, fatigue or
medication, and is open to manipulation by individuals who understand the process
and use countermeasures; for example, thinking anxiety-provoking thoughts whilst
their baseline levels are being measured (Ford, 2006). Individuals with antisocial
personality disorder or psychopathy may be more able to pass polygraph tests whilst
lying (Verschuerea, Crombeza, Kostera, & De Clercq, 2007). This is proposed to be
due to autonomic hyporesponsivity in psychopathic individuals, in particular a reduced
skin conductance, and a reduced resting heart rate (Verschuerea et al., 2007).
Contemporary polygraph lie detection, which could be seen to represent the state of
the art, still measures autonomic nervous system activity via breathing rate and
pattern, GSR, blood pressure, and heart rate, but its validity and reliability are
contested (Lykken, 1998). Reviews suggest that accuracy varies between 50% and
90% (Brett, Phillips, & Beary, 1986; Stern, 2003) with a converging approximation of
around 75% sensitivity (correctly identifying guilty persons) and 65% specificity
(correctly identifying innocent persons) (Brett et al., 1986). Polygraph evidence is
usually not permitted in a court of law (Ford, 2006). A recent report by the United
States National Research Council (NRC) concluded that there was insufficient
evidence that the polygraph is able to accurately detect deception, and that,
‘‘Countermeasures pose a serious threat to the performance of polygraph testing
because all the physiological indicators measured by the polygraph can be altered by
conscious efforts through cognitive or physical means’’ (National Research Council,
2003, p. 4), highlighting the need for alternative methods. For these reasons, modern
interest in deception-detection has moved on from measures of peripheral nervous
169
system activity, such as polygraphs, to more direct assessment of the central nervous
system, using techniques such as fMRI or EEG.
Crucial to detecting deception through physiological changes, is understanding of the
neurological underpinnings of deliberate deception. The immense progress that has
been made in psychophysiological techniques in the last few decades has begun to
untangle these processes and direct future research. Psychophysiological
techniques measure the physiological responses generated by cognitive activities,
using markers such as blood flow in the brain, electrical or magnetic field changes,
and pupil dilation. Multiple techniques have been developed to assess subtle taskinduced changes in psychophysiological markers in the laboratory. These
approaches share goals, such as gaining understanding of ‘hidden’ psychological
processes, but they differ in aspects such as methodology, equipment, resolution,
which structures, signals and processes can be measured, how invasive participation
may be, and the types of participants who are suitable for analysis. Different
techniques bring complementarities to the topic through their strengths and
weaknesses, and different populations, such as healthy adult, brain injured, elderly,
and developmental, contribute different insights.
The majority of techniques used to assess central nervous activity can be broadly
divided into real time recording techniques that directly tap the activity of the neurons,
such as ERPs and MEG, and indirect techniques that monitor cognitive activity by the
haemodynamic and metabolic changes within the brain, such as fMRI and Positron
Emission Tomography (PET). Diffuse Optical Tomography (DOT) techniques such as
Near Infrared Spectroscopy (NIRS) and Event-Related Optical Signal (EROS) are
non-invasive, cheap and relatively portable, and may have the potential to measure
170
directly and indirectly (Irani, Platek, Bunce, Ruocco, & Chute, 2007), but have not yet
been extensively applied to the topic of deception and won’t be considered here.
In conducting empirical tests of the efficacy of psychophysiological indices of
deception, it is important to consider the experimental protocols involved. Studies in
the detection of deception have used various experimental deception protocols. One
common paradigm involves asking a participant to commit or observe a small crime in
another room, and then conceal their involvement in the incident from the
experimenter (e.g., Berrien, & Huntington, 1943; Mohamed et al., 2006). This Guilty
Knowledge Test (GKT), or Concealed Information Test (CIT), has been used
extensively with polygraph tests and in applied forensic settings (Ben-Shakhur, &
Elaad, 2003). It involves forced-choice answers to questions about the crime under
investigation, where one “relevant” answer contains information about the incident,
and several “neutral” answers act as control items. Neutral answers are chosen so
that they are indistinguishable from the relevant answer to innocent individuals, for
example: “You stole money from my wallet. Was it £20, £100, £10, £50, £30, £70?”
(Engelhard, Merckelbach, & van den Hout, 2003). Participants are asked to respond
“no” to each item, and guilt is inferred if individuals consistently display larger
psychophysiological responses (for example GSR or ERPs) when giving the relevant
answer compared to neutral answers (Ben-Shakhur, & Elaad, 2003). A variation of
the GKT uses three different item-types in a task similar to a standard recognition
task: ‘target’ items, provided for the participant to learn prior to the test (e.g., farm
animals) and which require a “yes” response, and two types of non-learned items that
require a “no” response - ‘probe’ items, representing knowledge relevant to the
investigation that only the guilty person would know (e.g., “ring” when the mock crime
was stealing a ring), and ‘irrelevants’, neutral items not relevant to the investigation
(e.g., 5 other items of jewellery) (Ford, 2006). An innocent person will respond to
171
probes and irrelevant items in the same way as probes are no more meaningful than
the irrelevant items, however guilt is inferred if psychophysiological responses to
probes are similar to targets, which are meaningful (Ford, 2006). Reviews of the GKT
have shown it to have a sensitivity of 70%-85% (Ben-Shakhur, & Elaad, 2003) and
specificity of 80%-82% (MacLaren, 2001).
As the NRC point out, measures such as breathing rate and GSR can be manipulated
consciously. In an attempt to find measures that are not so susceptible to
countermeasures, researchers have used ERPs to try to detect deception.
Abootalebi, Moradi and Khalilzadeh (2006) carried out a single-probe GKT whilst
recording ERPs. In two tasks participants chose whether they were ‘guilty’ or
‘innocent’. Prior to being questioned, guilty participants looked in a box containing an
object whilst the examiner was outside the room. The test contained five pictures of
objects, one a target object that had been shown to all participants, one a probe (the
object in the box) and three objects that had not been seen by any of the participants.
Participants then had to indicate whether they had seen the object in the picture and
guilty participants had to hide their knowledge of the probe by responding “no”.
Abootalebi et al. (2006) measured the P300, an ERP thought to represent higher
cognitive responses, to cognitively salient, distinct, learned or unexpected stimuli (see
Polich, 2007, for a review; Molnar, 1994; Sutton, Braren, Zubin, & John, 1965). Due
to the fact that each set of stimuli were presented 30 times, giving a single average
P300 ERP per item-type per participant, the authors used a bootstrapping
(Wasserman, & Bockenholt, 1989) algorithm to estimate the distribution of average
P300 waves and create a “guilt threshold” with which to assess an individual’s
innocence. Using bootstrapped amplitude differences to classify participants they
were able to correctly detect 74% of guilty participants, but noted that a guilty
172
individual with a poor P300 response to probe stimuli would remain undetected and
be classified as innocent.
Whilst the psychophysiology of deception literature is primarily concerned with
forensic rather than neuropsychological applications, some studies have looked
directly at malingering using psychophysiological methods. For example, Wu, Allen,
Goodrich-Hunsaker, Hopkins and Bigler (2010) showed with fMRI scans that, in spite
of considerable brain damage, the same brain regions were activated during the
WMT as in controls, whereas different patterns were seen for participants who were
asked to simulate incomplete effort, suggesting that fMRI could be used to
differentiate malingered and genuine performance.
ERPs have also been used to investigate feigned memory-impairment. Rosenfeld,
Ellwanger and Sweet (1995) used an oddball paradigm in which autobiographical
items such as participants’ dates of birth, phone numbers and mothers’ maiden
names were presented on a screen among eight unrelated non-oddball stimuli of the
same type. Participants were required to repeat the item out loud to ensure that
“malingering” was not achieved simply by participants avoiding looking at the items.
They found an enhanced P300 in response to these uncommon, personally relevant
items, regardless of overt recognition response. Participants were classed as guilty if
their oddball to average non-oddball amplitude ratio was larger than 1.5, and no
individual non-oddball amplitude was more than 20% larger than the average nonoddball amplitude. Using this method the authors were able to identify 92% of
malingerers for birthdates and phone-numbers and 77% for mothers’ maiden names.
Whilst ERP based techniques show some promise, they are costly in terms of both
equipment and time. In addition, as with standard polygraph techniques,
countermeasures to defeat P300 as an index of deception are easily learned, such as
173
moving fingers or toes, or visualising being slapped, whilst responding to irrelevant
stimuli, to generate a P300 similar to those evoked by relevant and probe stimuli
(Rosenfeld, Soskins, Bosh, & Ryan, 2004). Rosenfeld et al. (2008) report
development of a new, countermeasure-resistant P300-based method for detecting
concealed information, The Complex Trial Protocol, which identifies attempted
countermeasure use from extended reaction times (Winograd, & Rosenfeld, 2011).
Lykken (1959; 1960) devised the GKT to test recognition memory, rather than
deception-related stress, and the neuroelectric changes associated with mnemonic
processes have been extensively documented in the ERP literature. For example a
larger frontal negative going component (FN400) and larger parietal late positive
component (LPC) have been found in response to the presentation of old learned
items compared to new items during a recognition memory test (Warren, 1980; van
Hooff et al., 1996; Friedman, & Johnson, 2000; see Chapter 6, section 6.1.1 for more
detailed discussion of these components). This ERP old/new effect provides
researchers with a more direct strategy with which to attempt to identify people who
feign memory loss (e.g., Browndyke et al., 2008; Tardif, Barry, Fox, & Johnstone,
2000; van Hooff, Sargeant, Foster, & Schmand, 2009).
For example, Tardif et al. (2000) reasoned that if the ERP old/new effect is not under
conscious control then it should be detectable in people feigning amnesia when they
claim that previously encountered old information is actually “new”. In their
experiment, participants learned a set of words before completing a recognition
memory test. Half the participants were given standard test instructions to respond
“old” to the learned words and “new” to new words. The other half were asked to
perform deliberately poorly on the test. Although the behavioural performance was
worse for the malingering group, the ERPs revealed an old/new effect comparable in
174
magnitude and topography to the control group, suggesting that the malingering
group did in fact have intact recognition of the learned words. In addition, the
malingering group did not show differences in response latency to correct and
incorrect responses, and demonstrated an early (190-320ms) old/new P2 difference
that was not present in the truthful group – which may represent processing
differences involved in the act of malingering (Tardif et al., 2000). Classification was
carried out using a direct discriminant function analysis of differences in reaction time,
LPC old/new effect and P2 old/new effect, and correctly identified 84.2% of
malingerers and 78.9% of the control group. However, Allen and Mertens (2009)
suggest that memory distortion limits the accuracy of this type of approach. Their
study showed that under certain conditions, patterns of neural activity for true and
genuine false recognition are indistinguishable.
5.1.1.3.
Pupillometry Studies
As discussed in Chapter 1, section 1.2.3.7, despite being under autonomic control,
relatively little research has been conducted into the activity of the pupil during lying
(Janisse, 1977). However, the eyes in general have long been viewed as non-verbal
indicators of truth-telling. In their meta-analysis of 158 different cues to deception in
120 independent samples spanning 60 years of research, DePaulo et al. (2003) found
several aspects of the eyes had been associated with lying, such as the amount of
eye contact (avoidance), gaze direction, blink rate (linked to arousal and cognition,
see Chapter 1, section 1.2.4.2) and pupil dilation (also linked to arousal and cognitive
effort, see Chapter 1, sections 1.2.2 and 1.2.2.1). Of these, only pupil-size
demonstrated a statistically significant effect size (d =.39, p <.05; DePaulo et al.,
2003), yet the relationship between pupil-size and deception has received
comparatively little attention, possibly because such subtle changes are much harder
175
to detect and measure accurately or easily in real world situations, and therefore may
be less consciously linked with dishonesty.
Elaad (2009) investigated male police officers’, prisoners’ and laypersons’ beliefs
regarding their own and others’ nonverbal abilities at detecting and telling lies. The
questionnaire included perceived cues (e.g., “Liars are more gaze aversive”) and
actual cues to deception (e.g., “Lies are accompanied by pupil dilation”) and found
that prisoners (and in particular older prisoners) were more aware of pupil dilation as
a cue to deception (50%) than lay persons (17%), or even police officers (21%).
Around half the participants in all three groups were aware that, “Lies contain more
negative statements”. Elaad (2009) suggests that the increased awareness among
prisoners may be due to increased exposure to and corrective feedback from lies.
In an early study Berrien and Huntington (1943) asked half of their participants to
commit a small monetary theft as part of the experiment, and to later lie when
questioned about it. They compared changes in their pupil-size with a group of
participants with no knowledge of the crime, instructed to answer questions truthfully.
Measurements were made using a telescope focussed on the eye and moved left and
right by an observer as the pupil dilated or constricted; movements were transmitted
to, and amplified by, a capillary pen on a polygraph. All participants were asked
unrelated baseline questions, and questions relating to the crime; deceivers were told
that they would be able to keep the money if their dishonesty remained undetected.
Lying was associated with a greater incidence of slow dilations followed by quick
constrictions and increased instability in pupil-size. They also found that pupil dilation
appeared to be more specific to the liars than concurrent increases in blood pressure,
and attributed this to the emotion evoked by being dishonest. However, in the
176
absence of any statistical analysis or objective quantification of the changes in pupilsize, this finding clearly requires replication with more modern techniques.
Later attempts using photographic methods confirmed increased pupil-size when
participants lied about demographic information, were asked about “guilty
information”, or viewed photographs of items from the scene of the “crime” (e.g.,
Heilveil, 1976; Janisse, 1973; Bradley, & Janisse, 1979; 1981; Janisse, & Bradley,
1980; Lubow, & Fein, 1996). Given the large body of evidence demonstrating that
pupil-size increases with “arousal” and is correlated with other psychophysiological
measures of arousal such as heart rate and GSR (see Chapter 1, section 1.2.6),
researchers have often used the logic behind polygraph recording to investigate the
potential of pupil-size as an index of deception in laboratory settings.
Lubow and Fein (1996) combined a GKT with pupil-size measurements to index
processing load, arousal and emotional state. Participants were randomly allocated
to either an innocent group or a group instructed to carry out a mock crime.
Photographs of the crime-scene and aspects of the crime acted as probes, and
participants were told that their pupil-size, eye movements and GSR would be
measured during the ‘interrogation’. Guilty participants had larger pupil-sizes in
response to probe items than to control items, whereas innocent participants’ pupilsizes were comparable between probe and control items. Participants with guilty
knowledge also had larger pupil-sizes to control items compared to innocent
participants. The experimenters were able to correctly classify 50% guilty and 100%
innocent participants using only differences in pupil-size, comparable to electrodermal
data (55% and 93%).
Webb et al. (2009) examined pupil diameter using a mock crime experiment where
half the participants stole money in the lab and all participants were later given a
177
Comparison Question Test (CQT). Like the GKT, the CQT asks questions relevant to
the crime (e.g., "Did you take any of the missing money?"), but also probable-lie
questions that are vague and difficult to answer when trying to appear honest (e.g.,
"Before the age of 30, did you ever take something that did not belong to you?"). The
rationale is that innocent participants will react more to the probable-lie questions
because they can honestly answer the relevant questions, whereas guilty participants
will react more strongly to the relevant questions. The CQT is controversial, because
results may reflect surprise, anxiety or stress as much as deception, requiring
subjective interpretation by the investigator (see Ben-Shakhar, & Furedy, 1990; Honts
et al., 2005; Iacono, & Lykken, 2005), however it is used internationally in actual
forensic cases (Raskin, & Honts, 2002). Webb et al. (2009) found that innocent
participants showed larger increases in pupil-size to probable-lie than to relevant
questions, whereas guilty participants showed similar increases in pupil-size to both
question types. Regression analyses revealed that pupil-size was a significant
predictor for deception, improving the adjusted R2 from .39 to .46, however this
increase only approached significance. The authors concluded that pupil-size could
be used in place of blood pressure measurement in traditional polygraphy, but did not
increase detection rates very much in addition to existing measures.
Using a different approach, Dionisio et al. (2001) asked participants to answer
episodic and semantic questions relating to general knowledge (“What are the colours
of the American flag?”) or specific vignettes (“What was the name of the person in the
story?”). They were required to answer twice, once honestly and once deceptively.
In 92% of participants they observed significantly larger pupil dilation for both types of
question when participants were lying. The researchers proposed that greater
cognitive processing was associated with creating convincing deceptive responses
than with genuine recall, thus leading to larger pupil-sizes (Dionisio et al., 2001).
178
Wang (2010) argues that deception studies often lack realism because their nature
dictates participants’ responses, and real-world motivations to malinger, such as
strong financial or emotional incentives, are largely absent (Bauer, & McCaffrey,
2006). Wang, Spezio, and Camerer (2010) attempted to address these drawbacks in
a biased-transmission game eye-tracking study, where the sender communicates
biased information and gets more points for greater exaggerations if they successfully
mislead their receiver. They found that senders’ pupils dilate when sending deceptive
messages compared to sending accurate messages, and that dilation increases with
the magnitude of the deception. They suggested that this was because figuring out
how much to deceive another player is cognitively difficult (Wang et al., 2010).
However psychophysiological approaches that rely on indices of increased “cognitive
effort” and “arousal” are still vulnerable to countermeasures such as counting
backwards in sevens or thinking anxiety-provoking thoughts whilst baseline levels are
being measured, and attempting to ignore relevant items (Rosenfeld et al., 2004). In
addition, if a person is able to lie easily without increased stress, then these
measures would not be appropriate – for example, individuals with antisocial
personality disorder or psychopathy are thought to pass polygraph tests whilst lying,
due to the hyporesponsivity of their autonomic nervous system (Verschuerea,
Crombeza, Kostera, & De Clercq, 2007). It remains to be seen whether pupil-size is
also susceptible to countermeasures in relation to deception. Ekman, Poikola,
Mäkäräinen, Takala and Hämäläinen (2008a; Ekman, Poikola, & Mäkäräinen, 2008b)
have designed a computer game that responds to player pupil-size and have had
modest success in training participants to manipulate their own pupil-size by holding
their breath, hurting themselves, reflecting on an emotional event, performing mental
arithmetic or changing their point of focus in a (slow) paced task.
179
As yet the pupillometry research has not yet been extended to look at the malingering
of memory-impairment. Given the comparative ease with which pupil measurements
can now be made, it is important to establish whether an approach similar to the ERP
old/new malingering studies (e.g., Tardiff et al., 2000; van Hooff et al., 2009) might be
feasible using the PONE. Some early evidence suggests that, like the ERP old/new
effect, the PONE is not under voluntary control during a memory task. Clark and
Johnson (1970) informed participants that their pupil would increase or decrease in
size during a short term memory task, or, in a control condition, did not mention pupilsize at all. They found that pupil-size increased to a similar extent in each condition,
not only when participants had been told it would increase. If the PONE is not under
voluntary control, and represents an automatic consequence of successful
recognition, this could provide a method of detecting deception that is independent of
emotional stress levels and cognitive effort.
This chapter presents a series of three experiments which artificially reduce
recognition memory performance to explore the role of conscious awareness in the
PONE, and the extent to which it may relate to conscious control. Experiment 5 looks
at whether the PONE is under voluntary control by asking participants to simulate
memory malingering to determine whether the pupil responses are aligned with item
status (old or new) or participant response (old or new). In an attempt to further
establish the effect of particular strategies on the PONE, Experiment 6 looks at
different types of malingering strategy, including asking participants to provide
incomplete effort, or to think their answer without indicating behaviourally to the
experimenter. Finally Experiment 7 looks at whether dividing attention at learning
and/or recognition reduces recognition performance (simulating memory-impairment)
without participants using a malingering strategy, and how genuinely reduced
memory performance affects the PONE.
180
5.2. Experiment 5 – Malingering and the PONE
Experiment 5 set out to replicate the PONE and to investigate its relationship to
participants’ responses, employing the same basic task as the explicit recognition
conditions in Experiments 1 to 4 in which participants are asked to learn a list of
words and then state whether items on a second list are old or new. If, like the ERP
old/new effect, the PONE is not under voluntary control, pupil-size should increase for
old items compared to new items even when participants feign amnesia and pretend
not to recognise learned stimuli in the “malingering” condition (e.g. falsely respond
“new” to items that are actually “old”). As pupil-size has been shown to increase in
relation to cognitive load (see Chapter 1, section 1.2.2.1), and it is generally assumed
that for most participants deception involves more cognitive effort than telling the
truth, a third “single response” condition was included, in which participants answered
“new” to all items. It was predicted that pupil-size would also increase for old items
compared to new items in this condition.
5.2.1.
Method
5.2.1.1.
Participants
Twenty-six participants (6 male; age range: 19.5-30.3, M = 23.1, SD = 3.3), with
normal or corrected-to-normal vision in at least one eye were recruited from the
student psychology participation pool at the University of Sussex, and through
personal contact. Participants were briefed with a detailed information sheet and
verbal description of the task, and invited to ask questions. Written consent was
obtained prior to testing and participants were fully debriefed. The experiment was
approved by the relevant ethics committee.
181
5.2.1.2.
Materials/Apparatus
Three study lists were created for the learning phase, each list comprising 40 nouns
selected from the MRC Psycholinguistic Database. For the recognition test, three
lists were constructed, each containing the 40 items that were on the corresponding
study list and 40 new nouns that were not. All nouns were 7 letters long, and the old
and new items were matched for familiarity and imageability, according to the K-F
norms (familiarity range = 301-646, M = 493; imageability range = 261-630, M = 497).
The three parallel sets of study lists and recognition tests formed blocks A to C, and
were presented in black 20pt Mono-spaced font on a light grey background under
fixed illumination. Words were presented using Experiment Builder software (SRResearch, Ontario) on a 21” CRT monitor. Participants viewed the monitor from a
distance of 70cm and the visual angle subtended by the words was approximately 3 o.
Eye movements were recorded with an EyeLink II (SR-Research, Ontario), with a
sampling rate of 500Hz. All items are presented in Appendix D.
5.2.1.3.
Design and Procedure
In a within-subject design each participant completed three separate recognition
memory tests. Each test contained a learning phase and a recognition phase but the
instructions given differed across the tests. At the start of the experiment, participants
were asked to imagine that they had recently been involved in a car accident and as a
result were unconscious for 15 minutes and had to spend one night in hospital for
observation. They were told that their condition had gradually improved over the
following months and they had now made a full recovery. They were asked to
imagine that the purpose of the test that they were are about to undertake was to
determine whether the accident had produced any long-term memory-impairments
due to brain damage. This scenario was adapted from van Hooff et al. (2009).
182
In the standard instructions condition participants were asked to perform as
accurately as possible during the recognition test. In the malingering condition,
participants were asked to produce responses that would convince an examiner that
they still had a memory-impairment. They were advised their responses should be
presented in a “believable” manner, and major exaggerations, such as not
remembering anything, should be avoided. To simulate real-world compensation
participants were told that £10 worth of book vouchers would be awarded to the
individual who best managed to simulate a believable memory deficit. In the “single
response” control condition, participants were instructed to simply answer “new” to all
items, regardless of whether they knew them to be old or new. This condition was
intended to mimic a simple strategy that might be used by people feigning amnesia,
and also allowed us to rule out any potential confounding influences on pupil-size that
might result from the increased cognitive effort required to generate incorrect
responses in the malingering condition.
During the learning phase, 40 study list target items were presented on screen for
2000ms with 1000ms between words, and participants were asked to remember the
items. During the testing phase, 80 recognition list items (40 old targets and 40 new
distracters) were presented for 1750ms, each following a 250ms mask (“&&&&&&”).
The mask reappeared after 2000ms and participants stated whether the word was old
(previously encountered in the learning phase) or new (not previously encountered).
Participants were then presented with a screen prompting them to estimate their
confidence in their decision with a number between 1 and 5, where 1 represented a
complete guess and 5 represented total confidence. This screen was then replaced
by a drift-correction dot in the centre of the screen in preparation for the next trial.
183
To control against list and order effects, the condition order was rotated across
participants. To determine whether any effects on pupil-size differed as a function of
condition order, this variable was added as a factor to all initial statistical analyses.
There were no main effects of order nor did it interact with any other factors, so, for
ease of interpretation, it is not included in the analyses reported in the results section.
To prevent the recognition phase instructions influencing behaviour during the
learning phase (i.e. participants may not have concentrated on the study items if they
knew they were going to be saying “new” to all items), instructions for the recognition
phase were provided after the learning phase in each condition. Old/new judgements
and confidence estimates were recorded on the computer after each recognition item.
5.2.1.4.
Pupil Recording
Maximum pupil-size was recorded from the right eye during each recognition period.
A Pupil Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated
expressing the maximum pupil-size for each 1750ms recognition trial, as a proportion
of the maximum pupil-size during that trial’s 250ms baseline.
5.2.2.
Results
5.2.2.1.
Behavioural Data: Old/New Responses
The proportion of old responses to old and new items was calculated for standard and
malingering conditions. No old responses were made in the single response control
condition. A 2 (item-type: old vs. new) by 2 (condition: standard vs. malingering)
repeated measures ANOVA showed a significant main effect of item-type – in general
participants responded “old” more often to old items than new items (F(1,25) = 162.5,
MSE = 0.027, p <.001, ηp2 =.87), a significant main effect of condition – in general
participants responded “old” more often in the standard condition (F(1,25) = 4.16,
184
MSE = 0.012, p <.05, ηp2 =.14), and a significant interaction between item-type and
condition (F(1,25) = 30.95, MSE = 0.03, p <.001, ηp2 =.55) – participants responded
“old” to old items significantly more in the standard condition (M = 0.79, SD = 0.12)
than in the malingering condition (M = 0.56, SD = 0.15; t(25) = 5.55, p <.001, r =.552),
whereas participants responded “old” to new items significantly more in the
malingering condition (M = 0.34, SD = 0.15) than in the standard condition (M = 0.20,
SD = 0.15; t(25) = 3.79, p <.001, r =.365; see Figure 5-1).
Standard
Malingering
Proportion "Old" Responses
p < .001
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
p = .001
Old
New
Item Type
Figure 5-1: Proportion of old responses to old and new items for standard and malingering
conditions. Error bars show standard error of mean.
5.2.2.2.
Behavioural Data: Confidence
Confidence ratings were analysed with a 3 (condition: standard, malingering, single
response) by 2 (item-type: old vs. new) repeated measures ANOVA. The main effect
of item-type was significant (F(1,25) = 27.06, MSE = 0.062, p <.001, ηp2 =.52) with
average confidence levels for old words (3.33) higher than for new words (2.96). The
main effect of condition was significant (F(2,50) = 69.58, MSE = 0.523, p <.001, ηp2
=.74) with average confidence levels close to ceiling in the single response condition
(4.84) and lowest in the malingering condition (3.17). Average confidence in the
standard condition was 3.90. A significant condition by item-type interaction (F(2,50)
185
= 15.64, MSE = 0.058, p <.001, ηp2 =.39) arose because confidence ratings were
significantly higher for old (M = 4.36, SD = 0.473) compared to new items in the
standard (M = 3.72, SD = 0.745; t(25) = 6.30, p <.001, r =.613) and malingering
conditions (old: M = 3.32, SD = 0.623 vs. new: M = 3.09, SD = 0.595; t(25) = 2.58, p
<.05, r =.210) but not in the single response condition (old: M = 4.86, SD = 0.368 vs.
new: M = 4.86, SD = 0.632; t(25) = 0.56, p >.05, r =.012; see Figure 5-2).
Old
5.00
New
p < .001
Average Confidence
4.50
p = .02
4.00
3.50
3.00
2.50
2.00
1.50
1.00
Standard
Malingering
Single Response
Condition
Figure 5-2: Average confidence rating for correct old and new items in each condition. Error bars
show standard error of mean.
5.2.2.3.
Pupil-Size Data
Average PDR for old and new items was calculated for each condition. As PDR is a
function of baseline pupil-size, baseline pupil-sizes to old and new items in the three
conditions were compared to ensure that any differences in PDR were not due to
baseline differences. The difference was not significant (F(1.63,40.76) = 1.90, p >.05,
ns, ηp2 =.07; Mauchly’s test indicated that the assumption of sphericity had been
violated (χ2(2) = 6.17, p <.05), therefore degrees of freedom were corrected using
Huynh-Feldt estimates of sphericity, ε = 0.98).
Average PDR for old and new words in the three conditions was compared with a 2 x
3 ANOVA with item-type (old vs. new) and condition (standard vs. malingering vs.
186
single response) as within subject factors. There was a main effect of item-type
(F(1,25) = 47.02, MSE < 0.001, p <.001, ηp2 =.65) – the PDR was larger for old items
compared to new items regardless of whether people were instructed to respond
veridically, feign amnesia or identify all items as new. The main effect of condition
was also significant (F(2,50) = 24.37, MSE = 0.001, p <.01, ηp2 =.49). Average PDRs
to old and new items were higher in the standard condition compared to the
malingering condition and higher again in the malingering condition compared to the
single response condition. These differences were significant for both old and new
items (all ts > 2.6, ps <.05). The main effects were, however, qualified by a significant
item-type by condition interaction (F(2,50) = 5.17, MSE < 0.001, p <.01, ηp2 =.17).
The interaction arises because the average increase in pupil-size is smaller in the
single response condition (M = 0.009, SD = 0.017) than in the standard (M = 0.025,
SD = 0.021, t(25) = 3.34, p <.01, r =.31) or malingering conditions (M = 0.018, SD =
0.021, t(25) = 2.07, p <.05, r =.15) (see Figure 5-3).
Old
New
Average Pupil Dilation Ratio
p < .001
1.14
1.13
1.12
1.11
1.1
1.09
1.08
1.07
1.06
1.05
1.04
1.03
p < .001
p = .02
Standard
Malingering
Single Response
Condition
Figure 5-3: Pupil dilation ratio for old and new items in standard, malingering and single response
conditions. Error bars show standard error of mean.
As participant response (old vs. new) was not meaningful in the malingering
condition, it was not included as a factor in the analysis above. However, it is
important to establish whether, in the standard condition, pupil-size increases for old
187
items that are not correctly recognised (misses), as a patient with genuine memory
problems might show poor explicit recognition memory but an increase in pupil-size
when targets are presented. Four participants made fewer than 5 misses, and were
therefore excluded from this analysis. Average PDR to missed old items was 1.09 –
the same PDR as was observed for correct rejections (t(20) =.02, p >.05, ns).
5.2.2.4.
Pupil-Size Data: Confidence Analysis
Participants made higher confidence ratings on average to their correct “old”
judgments compared to their correct “new” judgements in the standard and
malingering conditions. It is important to establish the extent to which the increase in
PDR that occurs when participants view old items is associated with the increase in
confidence that is associated with giving an old, compared to new, response. PDR
was significantly higher in the standard condition for high confidence (4 or 5; M =
1.13, SD = 0.055) compared to low confidence (< 4; M = 1.10, SD = 0.048) correct old
judgements (t(15) = 3.41, p <.01, r =.44). This analysis was restricted to the 16
participants who had at least 5 high and low confidence correct old judgements, and
to the standard condition because confidence judgements were not meaningful in the
malingering condition (it is impossible to determine whether reduced confidence
reflects a genuine uncertainty as to the correctness of their response or an
understandable attempt by participants to give the impression that they have a poor
memory).
Despite being overall slightly less confident in their correct rejections than their
correct recognitions, participants made significant numbers of high confidence correct
rejections. In order to further explore the relationship between confidence and PDR
we compared PDR for correctly identified old and new items to which participants
gave high confidence responses. PDR to old items that were correctly identified with
188
a high degree of confidence was greater than the PDR to new items that were
correctly identified with high confidence for all three conditions (standard: t(24) = 5.43,
p <.001, r =.55; malingering: t(23) = 4.03, p <.001, r =.41; single response: t(25) =
2.14, p <.05, r =.15).
5.2.3.
Discussion
Experiment 5 sought to replicate the PONE and determine its relationship with
participants’ responses. The size of participants’ pupils increased to a greater extent
when they viewed old items compared to novel items in a standard recognition test;
critically, this effect was also observed when participants were instructed to feign
amnesia, or even just to give a “new” response to all items.
The finding that under standard recognition memory instructions, participants’ relative
increase in pupil-size is greater when they view old items compared to new items,
replicates the findings demonstrated in experiments reported earlier in this thesis as
well as previous published research (see Chapter 1, section 1.3.2.3) and
demonstrates that the PONE is a robust phenomenon.
As discussed in section 1.3.2.3, it has been suggested that the PONE reflects
cognitively demanding recollective processes that occur during the recognition of old
items but not the correct rejection of new items (Võ et al., 2008). This interpretation
builds on an extensive body of work demonstrating that increases in pupil-size occur
as processing demands or cognitive load increase (see e.g., Kahneman, 1973).
However, it is not clear why recognition of previously presented items should
necessarily be more cognitively demanding than the correct rejection of novel items –
for example, correct rejection may involve an effortful memory search, particularly
when participants are using strategies such as “recall-to-reject” (Rotello, & Heit, 1999;
189
2000; Rotello, Macmillan, & Van Tassel, 2000) or recollection rejection (Brainerd, &
Reyna, 2002; Brainerd, Wright, Reyna, & Mojardin, 2001), where they recall a similar
or related item and know that the stimulus was not on the learning list (e.g., Jones, &
Jacoby, 2005; Leding, & Lampinen, 2009). Studies have found that it typically takes
longer to make a correct rejection than a correct recognition (e.g. Ratcliff, & Murdock,
1976).
The finding that the PONE was also observed in a single response condition (in which
participants simply had to respond “new” to all items) may also seem problematic for
an interpretation of the PONE based on cognitive effort – it could be argued that it
takes the same amount of cognitive effort to respond “new” to a word during a
recognition test when that word is old as it does when the word is new. It is possible,
however, that despite the lack of any requirement for a genuine old/new decision to
be made in the single response condition, recognition (and accompanying mnemonic
processes) still occurred when people encountered old items. If it is assumed that it
is these mnemonic processes themselves (as opposed to the cognitive effort they
may involve) that are associated with the increase in pupil-size, then the present
pattern of results would be expected. The PONE was greatest when participants
were given standard instructions to make a genuine old/new decision for each word
and diminished somewhat in the malingering and single response conditions. In the
absence of any requirement to respond accurately in the malingering condition
participants may have “preloaded” either an old or new response. This preloading
strategy was required in the Single Response condition (and is explored further in
Experiment 6). As a result in both malingering conditions less genuine
recognition/recollection may have occurred, with a resulting reduction in the
magnitude of the overall PONE effect when averaged across trials.
190
Another possibility is that the increase in pupil-size that occurs when participants view
old items during a recognition test somehow reflects differences in confidence
associated with correct recognition of old items compared to the correct rejection of
new items. Participants did indeed give higher confidence ratings on average to their
correct “old” judgments than their correct “new” judgements in both the standard and
malingering conditions. If (as suggested above) the PONE reflects the operation of
mnemonic processes during recognition then the extent of the pupil-size increase
might be expected to be associated with confidence. Recent models of recognition
memory have moved away from the idea that recollection is an ‘all or nothing’
process, instead suggesting that, like familiarity, the recollection signal may vary
along a continuum (e.g., Wixted, & Stretch, 2004). If the aggregate “strength of
memory” signal exceeds a certain threshold the item is identified as old (Wixted,
2007a; Wixted, & Stretch, 2004). If confidence ratings are taken as a reflection of
participant’s subjective experience of the strength of this aggregate signal, and the
pupil-size increase reflects the cognitive processes that drive this signal, then pupilsize increases should be greater for high compared to low confidence judgments, as
was indeed the case.
Despite this relationship between confidence and successful recognition, the PONE
does not simply reflect the difference in confidence between correct recognition of
targets and the correct rejection of distracters. Participants can, of course, be highly
confident that an item was not on the study list. When PDR was compared between
only the highly confident (4 or 5) correctly identified old and new items, the PONE
remained significant in all conditions. Similarly, participants were significantly more
confident when making correct rejections than false alarms, but there were no
differences in pupil-size. These findings suggest that whilst confidence may be
191
related to the magnitude of the PONE, the increase in pupil-size that occurs when
participants view old items does not simply reflect a “confidence signal”.
The key finding of Experiment 5 is the demonstration that the PONE occurs even
when participants are deliberately giving incorrect answers under instructions to
malinger, and when they are instructed to simply identify all items as new. These
results support Clark and Johnson’s (1970) finding that the PONE is not under
voluntary control and show that it is independent of participants’ actual response. A
similar argument has been made concerning the ERP old/new effect, and has been
used to support its potential use as an index of malingering (Tardif et al., 2000; van
Hooff et al., 2009). In a recent study, however, the ERP old/new effect was not
observed in a group of participants instructed to malinger (Vagnini, Berry, Clark, &
Jiang, 2008). Differences in procedure, in particular whether participants were asked
to feign amnesia before or after learning the word list, may account for the different
findings. Experiment 8 (Chapter 6) explores the potential relationship between the
PONE and the ERP old/new effect.
In conclusion, this study confirms and extends previous research demonstrating that
pupil-size increases more for previously encountered stimuli than for new items
during a recognition memory test. Critically, this increase appears to be independent
of the veracity of the behavioural responses and may have potential as a
comparatively simple and easy tool with which to detect patients feigning amnesia.
One thing that Experiment 5 did not address is types of malingering strategy,
participants were only asked informally and retrospectively what type of strategy they
used. Research has shown different strategies used in malingering including
providing incomplete effort (see section 5.1.1), therefore Experiment 6 was designed
to investigate the effects of the different types of malingering strategy that a
192
participant might use on the PONE, using three different instructions explaining how
to generate poor behavioural data.
5.3. Experiment 6 – Methods of Malingering
Experiment 5 sought to determine whether the PONE was still present when
participants gave deliberately incorrect responses, and found that the PONE is
independent of the veracity of the behavioural response. Experiment 6 seeks to
investigate different strategies that a malingering participant might use by giving
specific instructions on how to generate poor behavioural data. In addition to a
standard recognition condition, two conditions used strategies which were based on
malingering behaviours that many neuropsychological tests were designed to detect
(see section 5.1.1) – to make incomplete effort and not pay attention to stimuli during
the learning phase (Incomplete Effort condition), and to answer randomly by
preloading an old/new response before the item appeared on screen (Random
condition).
Bradley and Janisse (1975, cited in Janisse, 1977) conducted an experiment in which
participants were asked to select a numbered card, then either respond honestly
(neutral condition), lie out loud to all questions by saying “no” (overt condition), or
think “no” without responding (covert condition). Pupil-size was largest when
responding to the critical card than the other cards in all three conditions, suggesting
that due to deception and/or a response to a salient stimulus, participants were
unable to hide their true response even when giving a false response or not
responding out loud at all. Clark (1975, cited by Janisse, 1977) found similar results
in overt and covert response conditions, and was able to detect 80% of lies based on
pupil-size, compared to 85.8% using GSR and 63% using heart rate, and giving a
combined accuracy of 96.7%. A fourth condition therefore required participants to not
193
respond behaviourally at all, but to remain silent and think “old” or “new” (Quiet
condition).
In line with previous research (see Chapter 1, section 1.3.2.3) and the results of
Experiments 1 to 5, it was predicted that the PONE would be present for correct
responses in all conditions where a recognition decision was being made (Standard,
Incomplete Effort and Quiet). This was because even when some stimuli were
ignored during learning in the Incomplete Effort condition, for successfully encoded
stimuli, the pupil would still reflect mnemonic processes associated with recognition.
It was predicted that the PONE would not occur in the Random condition where
answers were preloaded rather than a recognition decision being made. This was
based on the findings of the reading conditions of Experiments 1 and 2 where no
PONE occurred in the absence of a recognition decision.
5.3.1.
Method
5.3.1.1.
Participants
Seventy-six participants (24 male; age range: 19.4-48.0, M = 24.5, SD = 5.27), with
normal or corrected-to-normal vision were recruited from the psychology course-credit
and subject pools at the University of Sussex, and through personal contact.
Participants were briefed with a detailed information sheet and verbal description, and
invited to ask questions. Written consent was obtained prior to testing and
participants were fully debriefed. The experiment was approved by the relevant
ethics committee.
194
5.3.1.2.
Materials/Apparatus
One study list was created for the learning phase, comprising 40 nouns selected from
the MRC Psycholinguistic Database. For the recognition test, another list was
constructed, containing the 40 items that were on the study list and 40 new nouns
that were not. All items were 6 letters long, matched for familiarity and imageability,
according to the K-F norms (familiarity range = 436-632, M = 556.15; imageability
range = 368-643, M = 548.85). The study list and recognition test were presented in
black 20pt Monospaced font on a light grey background under fixed illumination.
Words were presented using Experiment Builder software (SR-Research, Ontario) on
a 21” CRT monitor. Participants viewed the monitor from a distance of 70cm and the
visual angle subtended by the words was approximately 3 o. Eye movements were
recorded with an EyeLink II (SR-Research, Ontario), with a sampling rate of 500Hz.
All items are presented in Appendix E.
5.3.1.3.
Design and Procedure
In a between-subject design nineteen participants completed one of four conditions.
Each condition comprised a learning phase and a recognition phase. The four
conditions were standard instructions (Standard), instructions not to concentrate fully
during the learning phase but perform genuinely in the recognition phase (Incomplete
Effort), instructions to concentrate during the learning phase but randomly preload an
old/new response before the item appeared on screen during the recognition phase
(Random), and instructions to concentrate during the learning phase but to only think
old/new without a behavioural response during the recognition phase (Quiet).
Prior to the start of the experiment, instructions appeared informing the participant
that they would now see a list of 40 words. In the Standard, Random and Quiet
conditions the instructions informed them that they would have to try and remember
195
these words for a later task. In the Incomplete Effort condition, the instructions stated
that they should not try their best to remember the words as they were to fake a
convincingly poor performance at recognition.
During the learning phase of all four conditions the same 40 study list items were
presented on screen one at a time for 2000ms each in a randomised order. During
the recognition phase, an instruction screen then appeared informing the participants
that they were about to be presented with a list of 80 words, 40 of which they had
seen before (old) and 40 of which they hadn’t (new). In the Standard and Incomplete
Effort conditions, participants were asked to indicate whether each word was old or
new using the left and right trigger keys of the gamepad which served as the
response box on the EyeLink II system. In the Random condition participants were
asked to decide on their answer before the item was presented on screen, and
respond with that answer using the gamepad, even if it was wrong. In the Quiet
condition participants were asked to think to themselves whether the word was old or
new but not to indicate their answer verbally or via the gamepad.
At the start of each trial participants saw a drift correct dot, then a mask (“&&&&&&&”)
in the centre of the screen which lasted for 250ms. The mask was replaced by an
item from the recognition list for 2000ms. The next screen asked participants to
decide whether the word was old or new using a computer gamepad. This screen
was then replaced by a drift-correction dot in the centre of the screen before
presentation of the next trial, until all 80 items had been presented in a randomised
order.
196
5.3.1.4.
Pupil Recording
Maximum pupil-size was recorded from the right eye during each recognition period.
A Pupil Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated
expressing the maximum pupil-size for each 2000ms recognition trial as a proportion
of the maximum pupil-size during that trial’s 250ms baseline.
5.3.2.
Results
5.3.2.1.
Behavioural Data
The proportion of correct responses to old and new items was calculated for the
Standard, Incomplete Effort and Random conditions (no responses were made in the
Quiet condition). A 2 x 3 mixed-design ANOVA with a within-subjects factor of itemtype (old vs. new) and a between-subject factor of condition (Standard vs. Incomplete
Effort vs. Random) showed a significant main effect of item-type (F(1,54) = 134.26,
MSE = 0.018, p <.001, ηp2 =.713) – in general participants responded correctly more
often to new items than to old items, a significant main effect of condition (F(2,54) =
3.30, MSE = 0.014, p <.05, ηp2 =.109) – in general participants responded correctly
more often in the Standard condition than the Incomplete Effort or Random
conditions, and a significant interaction between item-type and condition (F(2,54) =
27.31, MSE = 0.018, p <.001, ηp2 =.503). The interaction occurred because
significantly more old items were correctly identified in the Standard condition (M =
0.729, SD = 0.138) than in the Incomplete Effort condition (M = 0.571, SD = 0.180;
t(36) = 3.03, p <.01, r =.204), whereas the difference in correctly identified new items
was not significant (Standard: M = 0.787, SD = 0.120; Incomplete Effort: M = 0.734,
SD = 0.146; t(36) = 1.21, p >.05, ns). Additionally significantly more new items were
correctly identified in the Incomplete Effort condition (M = 0.734, SD = 0.146) than in
197
the Random condition (M = 0.546, SD = 0.069; t(25.7) = 5.07, p <.001, r =.500;
Levene’s test indicated unequal variances (F = 7.19, p =.01), so degrees of freedom
were adjusted from 36 to 25.7), whereas the difference in correctly identified old items
was not significant (Incomplete Effort: M = 0.571, SD = 0.180; Random: M = 0.512,
SD = 0.066; t(22.7) = 1.35, p >.05, r =.07; see Figure 5-4).
Proportion Correct Responses
Standard
0.90
0.85
0.80
0.75
0.70
0.65
0.60
0.55
0.50
0.45
0.40
Effort
Random
p < .001
p = .004
Old
New
Condition
Figure 5-4: Proportion of correct responses to old and new items for Standard, Incomplete Effort and
Random conditions. Error bars show standard error of mean.
5.3.2.2.
Pupil-Size Data
Average PDR for old and new items was calculated for each condition. As PDR is a
function of baseline pupil-size, baseline pupil-sizes for old and new items in each
condition were compared to ensure that any differences in PDR were not due to
baseline differences. The difference was not significant (F(1,72) = 1.16, p >.05, ns,
ηp2 =.016).
A 2 x 4 mixed-design ANOVA with a within-subjects factor of item-type (old vs. new)
and a between-subject factor of condition (Standard vs. Incomplete Effort vs. Random
vs. Quiet) showed a main effect of item-type (F(1,72) = 12.76, MSE < 0.001, p <.001,
ηp2 =.151) – the PDR was larger for old items compared to new items regardless of
instructions. There was no main effect of condition (F(3,72) = 1.83, MSE = 0.002, p
198
>.05, ηp2 =.071) and the condition by item-type interaction was not significant (F(3,72)
= 1.63, MSE < 0.001, p >.05, ηp2 =.064), however planned contrasts revealed that
PDR was significantly larger for old items compared to new items in the Standard,
Incomplete Effort and Quiet conditions (all ts > 2, ps ≤ .05), but was not significantly
larger for old items than new items in the Random condition (t(18) = 0.055, p >.05, r
<.001); see Figure 5-5).
Average Pupil Dilation Ratio
Old
1.10
New
p = .05
p = .006
p = .031
1.09
1.08
1.07
1.06
1.05
1.04
Standard
Effort
Random
Quiet
Condition
Figure 5-5: Pupil dilation ratio for old and new items in Standard, Incomplete Effort, Random and
Quiet conditions. Error bars show standard error of mean.
5.3.3.
Discussion
Experiment 6 sought to investigate the effect on the PONE of different malingering
strategies that a participant might use, by giving specific instructions on how to
generate poor behavioural data. Instructions appear to have been effective because
participants in the Standard condition performed in line with previous experiments
(old: 72.9%, new: 78.7%) and in the Random condition performed at chance (52.9%).
Participants instructed not to learn the items in the Incomplete Effort condition
performed significantly lower at correctly recognising learned old items (57.1%), whilst
their performance in identifying new items was relatively intact (73.4%), as this
199
strategy would impair their ability to know if an item is old, but not that they have not
seen an item before.
As for Experiment 5, the size of participants’ pupils increased to a greater extent
when they viewed old items compared to novel items under standard recognition
instructions, further demonstrating the robustness of the PONE. The key finding of
the present study is the demonstration that the PONE occurs even when participants
are deliberately using malingering strategies. Whilst the main effect of item-type in
the absence of a significant interaction could be used to statistically argue that the
PONE is equivalent across conditions, suggesting that the PONE is robust to all three
malingering strategies, looking at the conditions independently using planned
comparisons showed that as predicted, the PONE was present in the Standard,
Incomplete Effort and Quiet conditions, but was absent when participants preloaded
an answer in the Random condition. The finding that the PONE exists when there is
no behavioural verbal response at all is fascinating, and suggests that similar to the
ERP technique, pupil-size provides a window on cognitive processes even in the
absence of an overt behavioural response.
In section 5.2.3, it was suggested that in the Malingering condition of Experiment 5
participants may have “preloaded” an old or new response at random, in the absence
of any requirement to respond accurately, reducing the magnitude of the overall
PONE when averaged across trials. It was also suggested that this preloading
strategy was a requirement in the Single Response condition (in which participants
replied “new” to all items), resulting in both conditions eliciting less genuine
recognition processes. However in the present experiment, explicitly asking
participants to use this strategy appears to eliminate the PONE, suggesting that
random preloading was not the explanation for the pattern of results in Experiment 5.
200
Further possible explanations for the reduction in PONE found in the Malingering and
Single Response conditions of Experiment 5 are considered in the General
Discussion in section 5.4.4.
As in the Random condition, the pupil-size for old and new items was also not
statistically different in the long and short duration reading conditions in Experiment 2
(Chapter 3, section 3.3), when participants were asked to simply read the words out
loud during the recognition phase, without the requirement to make an actual old/new
recognition decision. If it is assumed that it is the mnemonic processes themselves
(as opposed to the cognitive effort they may involve) that are associated with the
increase in pupil-size, then we would not expect to see a difference in these
conditions. It seems less likely that these processes occur unless a recognition
decision is being made. Alternatively, as participants were given instructions at the
start of the task, they may have employed two strategies together, firstly not paying
attention during the learning phase as their knowledge was not going to be utilised,
and then randomly preloading a response. While this strategy may have potential to
defeat a PONE-based test of malingered amnesia, it would produce chance-level,
and therefore clinically inappropriate, performance and could still identify a person
who is malingering as clinical cut-offs for tests of malingered memory are well above
chance (Bauer, & McCaffrey, 2006). Whilst this may catch a naïve malingerer, as
discussed in section 5.1.1, the advancement of tests of malingering advance is met
by improvements in strategies to defeat them.
As in Experiment 5, the two malingering conditions in which the PONE occurred,
Incomplete Effort and Quiet, show an overall smaller pupil-size. Whilst it can be
argued that less ‘recognition’ may be happening in the Incomplete Effort condition,
this is not the case in the Quiet condition where participants learned items under
201
standard instructions and were still asked to make the recognition decision in their
head. Research has shown that even the small amount of physical effort involved in
making a behavioural response amplifies pupil-size (e.g., Simpson, & Paivio, 1968;
see section 1.2.5), so pupil-sizes in the Quiet condition may have been smaller due to
the absence of a behavioural response. In addition, because the Quiet condition had
no behavioural data, it was not possible to analyse only correct trials. Including
missed old items in the analysis would have reduced the average PDR for old items
because misses, which can be interpreted to be items with a memory strength
insufficient to reach recognition threshold, are associated with a smaller pupil-size
than correctly identified old items, and statistically equivalent to that of correct
rejections (see Experiment 5).
All trials were also analysed in the Random condition because ‘correct’ trials would
have been an arbitrary half of the trials rather than the trials in which items were
correctly recognised or rejected. Therefore, the inclusion of misses in the analysis of
old items, and the inclusion of false alarms (which can be associated with an
intermediate pupil-size – Otero et al., 2011) in the analysis of new items, may have
masked the presence of a small PONE had it been present. Although this is not
ideal, it was not possible to restrict analysis to correct items in these conditions, and it
was necessary to collect data in this way for the purposes of testing potential
strategies. Whilst an (extra) response could have been added to the Quiet and
Random conditions, which was outside the period of the trial used for analysis of
pupil-size, and asked for an accurate assessment of each item’s old/new status, this
might have delayed the mnemonic processes we were attempting to measure or
distracted the participant from the task. However, this idea could be developed in the
future.
202
Building on strategies attempting to feign memory-impairment, Experiment 7 will
investigate the effects of genuinely impairing memory performance in healthy
participants using a divided attention task. By asking participants to perform a
secondary task at encoding and retrieval, this procedure should interfere with
mnemonic processing of list items, with the aim of reducing overall memory
performance and observing the resulting effect on the PONE.
5.4. Experiment 7 – Emulating Memory-impairment
Experiment 5 sought to determine whether the PONE was still present when
participants gave deliberately incorrect responses, and found that the PONE is
independent of behavioural response. Experiment 6 sought to investigate different
strategies that a malingering participant might use by giving specific instructions on
how to generate poor behavioural data, and found that not paying attention during the
learning phase decreased the pupil-size overall but did not diminish the magnitude of
the PONE.
Experiment 7 was designed to more realistically emulate the effects of a genuine
memory-impairment by dividing participants’ attention at learning and recognition.
Rather than asking participants to actively feign poor performance, it was hoped that
reduced attention to the stimuli at encoding, retrieval, or both, would lessen genuine
performance without participants having to use a strategy ‘on-line’ (which has its own
cognitive demands and may produce slower reaction times). For example,
performing a secondary task during the study phase has been shown to impair
performance at recognition, however there is asymmetry in the memory system
whereby divided attention at retrieval usually has only minimal effects on performance
(e.g., Craik, Govoni, Naveh-Benjamin, & Anderson, 1996; Naveh-Benjamin, Craik,
Guez, & Dori, 1998). It was therefore predicted that performance would be much
203
worse when the dual-task occurred during the learning phase than during recognition,
and that the best performance would occur when no secondary task took place. As
the PONE is thought to represent mnemonic processes, it was predicted that the
PONE would be reduced by dividing attention at encoding during the learning task
compared to performing a single task at encoding, and irrespective of whether a
single or dual task was performed at recognition. This was because the memory
strength for each item would be reduced due to reduced attentional resources
devoted to learning during encoding and fewer deep, elaborative strategies being
used. As retrieval is relatively robust to simultaneous tasks, it was predicted that
dividing attention during retrieval would only reduce the PONE when attention had
also been divided during the learning phase.
An effective secondary task is that of target detection (e.g., Naveh-Benjamin et al.,
1998) where the participant remains alert to intermittent and irregularly timed stimuli
and is asked to respond behaviourally when a target is detected. As this experiment
already utilised the computer monitor and the primary task was visual, for the
secondary task an auditory tone was presented. Participants were required to
respond verbally when they detected a tone, whilst also either learning stimuli during
the learning task or using the gamepad to indicate whether or not they recognised
stimuli in the recognition task.
5.4.1.
Method
5.4.1.1.
Participants
Seventy-two participants (28 male; age range: 18.8-66.1, M = 26.9, SD = 9.45), with
normal or corrected-to-normal vision were recruited from the psychology course-credit
and subject pools at the University of Sussex, and through personal contact.
204
Participants were briefed with a detailed information sheet and verbal description, and
invited to ask questions. Written consent was obtained prior to testing and
participants were fully debriefed. The experiment was approved by the relevant
ethics committee.
5.4.1.2.
Materials/Apparatus
One study list was created for the learning phase, comprising 40 nouns selected from
the MRC Psycholinguistic Database. For the recognition test, a second list was
constructed, containing the 40 items that were on the study list and 40 new nouns
that were not. All items were 7 letters long, matched for familiarity and imageability,
according to the K-F norms (familiarity range = 293-646, M = 491.86; imageability
range = 357-630, M = 544.91). The study list and recognition test were presented in
black 20pt Mono-spaced font on a light grey background under fixed illumination.
Words were presented using Experiment Builder software (SR-Research, Ontario) on
a 21” CRT monitor. Participants viewed the monitor from a distance of 70cm and the
visual angle subtended by the words was approximately 3 o. Eye movements were
recorded with an EyeLink II (SR-Research, Ontario), with a sampling rate of 500Hz.
All items are presented in Appendix F.
A podcast of an episode of Radio 4 programme The Archers was downloaded from
the BBC website. Using Audacity, open source software for editing audio tracks, a
100Hz tone 100ms in length was inserted into the Archers audio track on average
every 5s. The positioning of the tone within a 5s bin was random, using a list of
random numbers between 0 and 5 generated in Excel. The gain of the tone was 13Db in relation to the main audio track to make it difficult to detect to encourage
participants to attend to it. The file was converted into mp3 format using LAME MP3
205
Encoder, and then played on a Sony mp3 player through speakers positioned either
side of the monitor.
5.4.1.3.
Design and Procedure
In a between-subject design eighteen participants completed each of four conditions:
under standard instructions (Single/Single), under instructions to perform an auditory
task during the learning phase (Dual/Single), under instructions to perform an auditory
task during the recognition phase (Single/Dual), and under instructions to perform an
auditory task during both learning and recognition phases (Dual/Dual). Each
condition contained a learning phase and a recognition phase.
Prior to the start of the experiment, instructions appeared informing the participant
that they would now be seeing a list of 40 words. In all conditions, instructions
informed them that they would have to try and remember these words for a later task.
Both the Dual/Single and Dual/Dual condition instructions proceeded to say that there
would be a secondary task in which a recorded radio programme with embedded
tones would be played. Participants were asked to say “tone” when the tone
sounded. They were then played the first 10s of the recording so that they knew what
sound to identify. During the learning phase of all conditions the same 40 study list
items were presented on screen one at a time for 2000ms each in a randomised
order.
During the testing phase, an instruction screen then appeared informing the
participants that they were about to be presented with a list of 80 words, 40 of which
they had seen before (old) and 40 of which they had not (new). In the Single/Dual
and Dual/Dual conditions, participants were also informed about the secondary
auditory task and those in the Single/Dual condition (who had not completed the
206
secondary task in the learning phase) were then given a 10s clip of the sound file. At
the start of each trial participants saw a drift correct dot, then a mask (“&&&&&&&”) in
the centre of the screen which lasted for 250ms. The mask was replaced by an item
from the recognition list for 2000ms. The next screen asked participants to decide
whether the word was old or new using a computer gamepad. This screen was then
replaced by a drift-correction dot in the centre of the screen before presentation of the
next trial, until all 80 items had been presented in a randomised order.
5.4.1.4.
Pupil Recording
Maximum pupil-size was recorded from the right eye during each recognition period.
A Pupil Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated
expressing the maximum pupil-size for each 2000ms recognition trial as a proportion
of the maximum pupil-size during that trial’s 250ms baseline.
5.4.2.
Results
5.4.2.1.
Behavioural Data
The proportion of correct responses to old and new items was calculated for all
conditions. A 2 x 2 x 2 mixed ANOVA with a within-subjects factor of item-type (old
vs. new) and between-subject factors of learning task (single vs. dual) and recognition
task (single vs. dual) showed a trend effect of item-type (F(1,68) = 3.08, MSE =
0.015, p =.08, ηp2 =.043) – in general participants responded correctly more often to
new items than to old items (see Figure 5-6). There was also a significant main effect
of learning task (F(1,68) = 5.98, MSE = 0.014, p <.05, ηp2 =.081) – in general
participants responded correctly more often when their learning phase contained a
single task than a dual task. There was no main effect of recognition task (F(1,68) =
0.824, MSE = 0.014, p >.05, ηp2 =.012).
207
Proportion Correct Responses
Old
New
0.85
0.80
0.75
0.70
0.65
0.60
0.55
0.50
0.45
0.40
Single-Single
Single-Dual
Dual-Single
Dual-Dual
Condition
Figure 5-6: Proportion of correct responses to old and new items for all conditions. Error bars show
standard error of mean.
However, the learning by recognition task interaction was significant (F(1,68) = 3.81,
MSE = 0.015, p <.05, ηp2 =.053) – when the recognition phase contained a single
task, a dual task learning phase lead to poorer performance (M = 0.691, SD = 0.068)
than a single task learning phase (M = 0.778, SD = 0.079; t(34) = 3.56, p <.001, r
=.271). Additionally, whilst memory performance might be expected to be worst in the
Dual-Dual condition, when the recognition phase contained a dual task, performance
was decreased regardless of whether participants carried out a single or dual task at
learning (M = 0.722 and M = 0.712, SD = 0.105 and SD = 0.078; t(34) = 0.314, p
>.05, r =.003; see Figure 5-7).
Single Learning
Dual Learning
Proportion Correct Responses
p = .001
0.80
0.78
0.76
0.74
0.72
0.70
0.68
0.66
0.64
Single
Dual
Recognition Phase
Figure 5-7: Proportion of correct responses to all items for the four combinations of single and
learning task conditions. Error bars show standard error of mean.
208
5.4.2.2.
Pupil-Size Data
Average PDR for old and new items was calculated for each condition. As PDR is a
function of baseline pupil-size, baseline pupil-sizes for old and new items in each
condition were compared to ensure that any differences in PDR were not due to
baseline differences. The difference was not significant (F(1,68) = 0.751, p >.05, ns,
ηp2 =.011).
A 2 x 2 x 2 mixed-design ANOVA with a within-subjects factor of item-type (old vs.
new) and between-subject factors of learning task (single vs. dual) and recognition
task (single vs. dual) showed a significant main effect of item-type (F(1,68) = 35.55,
MSE < 0.001, p <.001, ηp2 =.343) – in general the PDR was larger for old items
compared to new items. There was no main effect of learning task, no main effect of
recognition task, and none of the interactions were significant (all Fs < 1, ps >.05) –
carrying out a secondary task did not affect the PONE. When the analysis was
restricted to those items correctly identified, there was still only a significant main
effect of item-type, no effects of learning or recognition task and no interactions (see
Figure 5-8).
Average Pupil Dilation Ratio
Old
New
1.16
1.15
1.14
1.13
1.12
1.11
1.10
1.09
1.08
Single-Single
Single-Dual
Dual-Single
Dual-Dual
Condition
Figure 5-8: Pupil dilation ratio for old and new items in all conditions. Error bars show standard error
of mean.
209
5.4.3.
Discussion
Experiment 7 sought to determine what the effect of dividing participant attention at
learning and recognition would be on the PONE. The key finding was that there was
still a main effect of item-type on pupil-size, but no effect of condition, even when
secondary task and learning and recognition were analysed separately. This result
suggests that the PONE is a robust effect that is not diminished by dividing attention.
It is less likely that the secondary task simply was not sufficiently distracting enough
from the main task to impact on the PONE, since performance measures were
affected by the manipulation.
It was hoped that reduced attention to stimuli at encoding, retrieval, or both, during
the divided attention conditions, would lessen genuine performance relative to the
Single-Single condition, without participants having to devise and apply a strategy
‘on-line’ (which has its own cognitive demands and may produce slower reaction
times) as in Experiments 5 and 6. There was a significant main effect of learning task
on performance, where participants responded correctly more often when their
learning phase contained a single task than a dual task, but there was no main effect
of dividing attention during recognition. This is consistent with the literature, which
states that performance is reduced by divided attention at encoding but not at
retrieval (Craik, Govoni, Naveh-Benjamin, & Anderson, 1996; Naveh-Benjamin, Craik,
Guez, & Dori, 1998).
5.4.4.
General Discussion
Experiments 5, 6 and 7 sought to further elucidate the circumstances in which the
PONE occurs by manipulating participant responses and secondary task demands,
and found that when a recognition decision is made, the PONE occurs even if the
210
behavioural response is deliberately false or absent, and is present for correct items
even when a secondary task reduces genuine performance levels.
In Experiment 5 it was suggested that preloading as a strategy might explain why in
the Malingering and Single Response conditions, the PONE was attenuated. It was
proposed that in the absence of the requirement to give an accurate response, less
genuine recognition/recollection may have occurred, reducing the overall PONE when
averaged over several trials. However, in the results of the Random condition of
Experiment 6, where participants were asked to preload an answer ahead of stimulus
presentation as an active strategy, no PONE was found. Therefore, whilst in the
Single Response condition of Experiment 5 participants were giving a predetermined
answer, it is possible that they were still making an old/new judgement on some of the
stimuli prior to answering “new”, and resulting in a PONE. For example, participants
were still required to respond to the prompt to make a confidence judgement after
each item, which may have kept participants focussed on the task as one of item
recognition. In the Random condition of Experiment 6, however, participants
concentrated on generating a random old/new response in the time before the next
stimulus was presented, and attention may have been focussed on this task rather
than the stimulus on the screen.
In the Malingering condition of Experiment 5, it would have been necessary for
participants to decide themselves whether the item was old or new in order to ensure
that they gave a performance that was below their best but above chance (as per the
instructions for “believable” feigned memory-impairment). Therefore, some trials with
new items, but which the participants identified as “old”, would have been averaged
into the PDR for old items, making it smaller, and vice versa with new items becoming
larger – the result being an attenuated PONE. The Random condition in Experiment
211
6 suggests that preloading prevents the PONE occurring, and similar to the reading
conditions in Experiment 2, may be due to the absence of a requirement to make an
explicit recognition decision on presented stimuli.
A notable result was that in Experiment 7, reducing the attention paid to items at
encoding and retrieval affected behavioural measures but not the PONE, in a similar
manner to the Incomplete Effort condition of Experiment 6, where participants were
asked non-specifically to pay less attention to stimuli during the learning phase but
still demonstrated a PONE. Whereas more new items are usually correctly identified
than old items, in the dividing attention tasks this was not the case, similarly in the
short duration recognition condition of Experiment 2, equivalent numbers of old and
new items were correctly identified. This might suggest that increasing task difficulty
has more of an effect on the processes involved in correct rejection than on those
involved in correct recognition.
The finding in Experiment 5 that the PONE can be reliably detected even when
participants are feigning amnesia and are reporting that they believe the items to be
new, or when they are remaining silent as in the Quiet condition of Experiment 6,
might have implications for individuals and organisations who administer
neuropsychological recognition memory tests in clinical or forensic settings. These
findings are similar to those of Tardif et al. (2000) who demonstrated an intact ERP
old/new effect in participants asked to feign amnesia. The absence of a significant
difference in PDR between old items missed, and correct rejections of new items, in
the standard condition, suggests that if a patient with legitimate memory problems
makes a genuine miss they would not be incorrectly identified as a malingerer on the
basis of their pupil-size. Clearly it will be important to establish how pupil-size
changes in genuinely memory-impaired populations when they perform this type of
212
recognition memory test. Laeng et al. (2007) recently investigated the pupil old/new
effect in three patients with amnesia resulting from hippocampal lesions. They found
that a larger pupil response occurs for new words compared to old words in these
patients, similar to the findings of Experiment 1.
Given more time it would have been interesting to replicate Experiments 6 and 7
using a within-subject design to give them more statistical power and further draw out
these effects. It would be interesting to look at how the PONE responds under these
experimental conditions in genuinely memory-impaired populations, to further
elucidate the contexts in which the PONE occurs. For example whether the PONE
for items that amnesic participants correctly identify is of the same magnitude as in
healthy participants, just occurring to fewer items as in the Incomplete Effort condition
of Experiment 6 and the dual task conditions of Experiment 7, or whether the overall
character of the PONE is diminished. Determining these parameters would indicate
whether a PONE by itself implies intact recognition memory for learned items, and
therefore whether the presence of a PONE in the absence of a correct behavioural
response indicates malingering.
It would also be important to establish whether the PONE can be diminished by
countermeasures other than the random preloading of responses seen in the
Random condition of Experiment 6. Techniques have been used by Ekman et al.
(2008a; 2008b) to train participants to increase and decrease the size of their pupil
with the aim of using this to control aspects of a computer game, including holding
their breath, hurting themselves, thinking about an emotional event, performing
mental arithmetic or changing their point of focus. As so many psychological and
physical events cause pupil dilation, it is possible that someone could train
213
themselves to, for example, perform an effortful cognitive task only when responding
to new items, in order to increase pupil-size to that evoked by new items.
Although technology has become more sophisticated (and more complicated), the
question remains of whether society is any better equipped to identify when someone
is lying (Wolpe, Foster and Langleben, 2005). Drob (2004, cited in Ford, 2006)
considers that almost any finding from current lie detection techniques could be
accounted for “by something other than lying or deception” (p. 169) – current
techniques are not definitive and, on their own, should not be taken as proof of lying
(Ford, 2006). Most psychophysiological techniques, including pupil-size measures,
require data to be averaged across multiple trials, increasing the signal to noise ratio,
but also increasing the costs and time involved, making it difficult assess individual
responses. An interesting extension of the experiments reported here would be to
adapt the design to perform a classification analysis using bootstrapping comparison
data to attempt to identify participants who are feigning memory-impairment.
Performed using a sample including genuinely memory-impaired participants,
participants simulating memory-impairment, and healthy controls, this would allow
cut-off scores for performance and pupil-size to be established for the three
categories of participants (e.g., Rogers, & Bender, 2003).
214
6. ERP and Pupil Old/New Effects
A Comparison
6
Previous chapters demonstrated that the PONE is a robust phenomenon that
accompanies explicit recognition decisions, whereby participants’ pupils are larger
when correctly judging old items than new items. In this chapter, Experiments 8 and
9 try to establish whether there is a relationship between the PONE and another
psychophysiological index of recognition memory, the ERP old/new effect.
6.1. Introduction
6.1.1.
Background to ERPs and Recognition Memory
For nearly 100 years researchers have known that external events produce
measurable electrical changes in the brain, with the first unambiguous experiments
being conducted in the 1930s and the development and proliferation of modern
Electroencephalographic (EEG) based Event-Related Potential (ERP) techniques
from the 1960s onwards (Luck, 2005). Small electrical voltage differences (relative to
a reference electrode) produced by the neurons of the brain are measured by scalp
electrodes whilst a participant carries out a task, and amplified, digitised and stored
on a computer. Whilst on individual trials consistent activity may not be visible within
the continuous EEG recording, when stimulus-linked sections are averaged over a
large number of trials, an ERP signal representing consistent neural activation can be
distinguished from random background noise (Luck, 2005). Individual ERPs are
identified as positive or negative deflections of the EEG voltage, conventionally
named after their polarity, and either the approximate time at which they peak, or their
215
ordinal position (for example the first positive deflection is P1, the second P2 and so
on; Luck, 2005).
Data from ERP recording has poor spatial sensitivity, due to the way that electricity
spreads out through the conductive medium of the brain (Luck, 2005). When meeting
the skull, which has a high electrical resistance, activity spreads sideways to reach
the point of least resistance (Luck, 2005). The local voltage recorded by the
electrode may relate to activity occurring in a distant part of the brain (Luck, 2005). In
addition, the mathematical ‘inverse problem’ means that for a given voltage
distribution it is not possible to definitively determine the sources of the underlying
activity (generators; Helmholtz, 1853; Nunez, 1981; Plonsey, 1963). For these
reasons ERPs alone cannot be used with confidence to localise cognitive processes;
instead experiments should be designed to play to the strengths of the ERP
technique (Luck, 2005). ERPs have millisecond time-resolution and can help to
determine the time-course of neural activation in response to cognitive activity
(Handy, 2004; Luck, 2005; van Hooff, Brunia, & Allen, 1996; see Chapter 2 section
2.2 for information about the recording, processing and analysis techniques used in
this thesis). The excellent temporal resolution of ERPs complements poor
temporal/good spatial resolution techniques which rely on slower metabolic
processes, such as glucose uptake in PET (see Bailey, Townsend, Valk, & Maisey,
2005), or blood flow in fMRI (see Huettel, Song, & McCarthy, 2004) to localise neural
activity with millimetre spatial-resolution.
During ERP data collection, participants are able to sit up to complete tasks in a more
realistic situation than may be possible with other forms of neuroimaging where
participants must lie horizontally in a scanner. Additionally, unlike behavioural
measures, ERPs are measured directly from the scalp, and can be utilised with
216
participants who are unable to speak or press buttons in response to stimuli, such as
young children, and in tasks where the process of interest is not measurable
behaviourally, such as aspects of language processing (Luck, 2005). For wellcharacterised ERPs, carefully designed studies can help determine which processing
stage(s) are influenced by an experimental manipulation (Luck, 2005).
Since the 1970s researchers have recorded ERPs that accompany the recognition of
a previously learned item (see Donaldson, Allan, & Wilding, 2002; Fabiani, Gratton, &
Coles, 2000; Friedman, & Johnson, 2000; Johnson, 1995; Rugg, & Allan, 1999; 2000,
for reviews). Using the old/new paradigm, different patterns of brain activity have
been observed for items recognised as old, compared to unseen new items.
Specifically, correctly identified old items tend to evoke a more positive-going ERP
occurring approximately 300-800ms post-stimulus onset compared to new items,
misses and false alarms (Karis et al., 1984; Sanquist et al., 1980). This shift,
sometimes referred to as a recognition positivity, is larger for better remembered
items (Smith, 1993) and occurs later than priming positivity (a broad positivity from
250-700ms which occurs in response to repeated stimuli; e.g., Bentin & Peled, 1990;
Rugg et al., 1994; Rugg, & Nagy, 1989), leading some researchers to interpret it as a
confidence-related enhancement of P3 (also known as P300) – a ubiquitous positive
going ERP which responds to a variety of task manipulations and overlaps spatially
and temporally with memory ERP effects (Johnson, 1986; Rugg, & Nagy, 1989; Rugg
et al., 1994). However other researchers have demonstrated that the old/new effect
and P3 are differently affected by manipulations such as probability and previous
exposure (Smith, & Guster, 1993), and when confidence is held constant the ERP
old/new effect is enhanced by factors such as low word frequency because infrequent
words are better remembered than common words (Rugg et al., 1995).
217
More recent research suggests there are two ERP old/new effects, the first concerns
a frontal N400 (or FN400) wave occurring 300-500ms after stimulus presentation that
is more negative for new items than old items (Wiese, & Daum, 2006), and which is
also known as the MTL-N4 (Smith, Stapleton, & Halgren, 1986), medial frontal
(Friedman, & Johnson, 2000), early frontal (Mecklinger, 2000) or mid-frontal old/new
effect (Tsivilis, Otten, & Rugg, 2001; Curran et al., 2006). This frontal old/new effect
bears similarities to the N400 evoked by visual or auditory word stimuli in the
semantic processing literature (Kutas, & Hillyard, 1980), but differs functionally and
topographically (Curran, Tucker, Kutas, & Posner, 1993; Curran et al., 2001), and is a
sensitive index of the degree of mismatch between a word and a previously
established semantic context (semantic priming; e.g., Bentin, & McCarthy, 1994;
Bentin, McCarthy, & Wood, 1985; Holcomb, 1998). The N400 responds to word
frequency (which affects familiarity), being larger for less frequent words (Van Petten,
& Kutas, 1990; 1991), and to stimulus repetition (Van Petten, Kutas, Kluender,
Mitchiner, & McIsaac, 1991), which may explain its similarity to the frontal old/new
effect, where new words are a mismatch with the experimental learning context.
However, the FN400 effect has also been reported for pictures (Curran, & Cleary,
2003), faces (Nessler, Mecklinger, & Penney, 2005) and objects (Mecklinger, von
Cramon, & Matthes-von Cramon, 1998).
The second old/new effect appears as a parietal positive-going wave from around
400-800ms, that is more positive for old items than new items (see Johnson, 1995, for
a review; Allan, Wilding, & Rugg, 1998; Friedman, & Johnson, 2000; Mecklinger,
2000; Rugg, 1995; Wilding, & Sharpe, 2003). As is common in ERP research (Luck,
2005), this component overlaps with the P300 component (Bentin, & McCarthy, 1994;
Spencer, Vila Abad, & Donchin, 2000; ERP waveform shapes reflect the sum of
underlying positive and negative going latent ERP components, which may be
218
independent yet difficult to isolate, leading to issues with interpretation – see Luck,
2005), and is referred to as the P300 old/new difference (Johnson, 1995), the parietal
old/new effect (Allan et al., 1998), the late ERP old/new effect (Rugg, 1995), the MTLP3 (Smith et al., 1986), the Late Positive Complex (LPC) (Olichney et al., 2000), and
the P600 old/new when studying sentences (Rugg, & Doyle, 1992; Curran, 1999;
Curran et al., 2006).
The parietal LPC old/new effect is suggested to index recollective processes (see
Allan et al., 1998, for a review; Friedman, & Johnson, 2000; Mecklinger, 2000;
Wilding, & Sharpe, 2003), whereas the FN400 old/new effect is thought to index
familiarity (Rugg et al., 1998a). These two old/new effects have been used to provide
electrophysiological evidence in support of dual-process models of recognition
memory, owing to the fact that they respond differently to experimental manipulations
designed to differentiate recollection and familiarity. For example, when using a
remember/know paradigm, “remember” responses produce a larger parietal old/new
effect than “know” responses (Curran, 2004; Düzel et al., 1997; Friedman, 2004;
Rugg et al., 1998b; Smith, 1993; Trott et al., 1999). Functional imaging has also
shown different patterns of brain activation associated with recollection and familiarity,
which are congruent with ERP topography (e.g., Wheeler, & Buckner, 2004).
Consistent with its association with recollective processes, the LPC old/new effect
has been associated with the retrieval of contextual “source” information about the
original learning experience, such as temporal source – whether an item was on the
first or second of two learning lists (Trott, Friedman, Ritter, & Fabiani, 1997), voice –
whether the speaker of an auditory word stimulus was male or female (Rugg et al.,
1998b; Wilding, & Rugg, 1996; 1997a), or stimulus modality – whether the item was
read or heard (Wilding, Doyle, & Rugg, 1995; Wilding, & Rugg, 1997b).
219
Levels of processing manipulations, which are believed to affect recollection more
than familiarity, have a greater effect on the parietal old/new effect – which shows
enhanced positivity for correctly identified deeply encoded words than for shallowly
encoded words (Paller, & Kutas, 1992; Paller, Kutas, & McIsaac, 1995; Rugg et al.,
1995) – than on the mid-frontal old/new effect, which does not differentiate between
deeply and shallowly encoded items (Rugg et al., 1998a). Similarly, the LPC old/new
effect is larger for old and new words than for old and new pseudo-words, however
the mid-frontal old/new effect is similar for old and new words and old and new
pseudo-words (Curran, 1999). Instead, the mid-frontal old/new effect is more
negative for old pseudo-words than for old words, possibly due to its sensitivity to
contextual mismatch (Curran 1999).
Curran (2000) used a plurality recognition task (see Chapter 1, section 1.3.1.2 for a
description of this manipulation) to differentiate familiarity and recollection in an ERP
study. Participants completed three blocks, in each of which they learned 40 words
and were tested on 60 words (20 old, 20 new and 20 similar lures of reversed
plurality). As expected, participant responses showed a higher rate of false alarms to
lures than to new items due to increased familiarity. The mid-frontal old/new effect at
300-500ms was more negative-going for correctly identified new items than correctly
identified old items and false alarms to similar lures, reflecting a difference in
familiarity, whereas the parietal LPC at 400-800ms was larger for old items than for
new or similar lures, reflecting recollection. Based on the qualitative difference in
topographical distribution of difference waves for familiarity (similar-new) and
recollection (old-new) in the two time windows, Curran (2000) concluded that their
findings produced strong evidence of dissociation between familiarity and recollection.
This pattern of results was replicated for old, new and reversed picture stimuli among
participants who performed well at discriminating between old and similar word items
220
(Curran, & Clearly, 2003). Participants who were poor at discriminating word stimuli
showed the same frontal familiarity effect, but did not differentiate between correct old
items and false alarms at parietal electrodes (Curran, & Clearly, 2003).
However, some researchers have shown posterior old/new effects associated with
familiarity and anterior old/new effects associated with recollection, when using
previously unknown faces (MacKenzie, & Donaldson, 2007; Yovel, & Paller, 2004).
Yovel and Paller (2004) used faces rather than word stimuli because they observe
that words have pre-existing levels of familiarity, which are a potential confound and
mean that learned items are not compared with truly novel stimuli. By pairing
unknown faces with spoken occupations during the learning phase, they were able to
isolate recollection and familiarity by asking participants to qualify an “old” judgement
with whether or not occupation, or other detail, could also be retrieved (indicating
recollection), or whether no additional detail could be retrieved (indicating familiarity;
Yovel, & Paller, 2004). As none of the faces had been previously seen, there could
be no pre-existing familiarity or recollection.
Recollected items evoked a parietal old/new effect in the 500-700ms time window,
consistent with previous studies (e.g., Curran, 2004; Düzel et al., 1997; Friedman,
2004; Rugg et al., 1998b; Smith, 1993; Trott et al., 1999). Familiar items also
demonstrated a parietal old/new effect between 500-700ms, but no early mid-frontal
old/new effect (Yovel, & Paller, 2004). Yovel and Paller (2004) somewhat
controversially suggest that for faces, familiarity is represented by the parietal old/new
effect, rather than the frontal old/new effect, which they attribute to the ‘conceptual
priming’ associated with the use of lexical stimuli (MacKenzie, & Donaldson, 2007).
In addition, the topographical distribution of the old/new effects for recollection and
familiarity were statistically equivalent and they concluded a shared set of underlying
221
neural generators for both types of recognition judgement (Yovel, & Paller, 2004),
which could be interpreted to support single-process models of recognition memory
(MacKenzie, & Donaldson, 2007). However, MacKenzie and Donaldson (2007)
extended this finding using faces devoid of hair, ears or background, paired with
names, in shorter test blocks and recording using a higher density of electrodes (61
vs. 21 scalp locations). In addition to Yovel and Paller’s (2004) posterior familiarity
old/new effect, they found a novel anterior recollection old/new effect, demonstrating
dissociation of recollection and familiarity by means of distinct ERP components
(MacKenzie, & Donaldson, 2007). The authors interpreted their findings from a dualprocess perspective and suggested that recognition old/new effects may differ under
different experimental settings.
6.1.2.
Pupil Responses and ERPs
As seen in Chapter 1, section 1.2.6, some researchers have recorded concurrent
pupil-size with other psychophysiological measures. Interestingly, some pupil
responses appear to have specific parallels in the ERP literature, for example in
vigilance or reaction time tasks, a pupil dilation has been observed beginning 10001500ms prior to the presentation of an expected stimulus or the requirement to make
a response, and which varies with the force of anticipated movement (e.g., Bradshaw,
1969; Klingner, 2010; Richer, Silverman, & Beatty, 1983). This response dilation has
been likened to the ERP components known as Contingent Negative Variation (CNV;
Richer, & Beatty, 1985; Richer et al., 1983; Rohrbaugh, Syndulko, & Lindsley, 1976)
and the Lateralised Readiness Potential (LRP; Becker, Iwase, Jürgens, & Kornhuber,
1976; Luck, 2005), both of which precede responses by around 1000-1500ms and
may represent pre-motor preparation to react to a stimulus (Beatty, & LuceroWagoner, 2000). Response preparation has also been shown to occur in other
222
psychophysiological measures such as heart rate (Coles, & Strayer, 1985) and
electromyogram (Brunia, & Vingerhoets, 1980).
Other researchers have shown that stimulus probability is inversely related to the size
of both task-evoked pupil dilations and the P300, whereby rare, low-frequency or
unlikely stimuli evoke the largest pupil dilations and P300 amplitudes (Bock, 1976,
cited by Janisse, 1977; Friedman et al., 1973; Qiyuan, Richer, Wagoner, & Beatty,
1985; Steinhauer, 1982). Some of these studies have looked at ERP and pupil-size
measures simultaneously. For example, after observing that P300 and pupil dilation
behaved in a similar manner in response to stimulus probability, Friedman et al.
(1973) measured them concurrently, and found that both the P300 and pupil-size
were inversely and monotonically related to stimulus probability in a guessing game.
Steinhauer (1982) found that the P300 and pupil-size both increased in relation to bet
value, event uncertainty and the absence of expected feedback in a gambling task,
and were much larger when participants selected bets rather than when the computer
chose for them.
Just et al. (2003) concluded that the correspondence between ERPs, pupillometry,
and also fMRI responses, to the same cognitive tasks, indicate a common underlying
“construct”, which they believe to be cognitive load (see Chapter 1, section 1.2.2.1).
Nieuwenhuis, Aston-Jones and Cohen (2005a; Nieuwenhuis et al., 2011a) proposed
that the P300 corresponds to the neuromodulatory Locus Coeruleus Norepinephrine
(LC-NE) system, reacting to perceptual decision-making in stimulus evaluation (the
role of the LC-NE system in stimulus evoked pupil-size change was discussed in
Chapter 1, section 1.1.2.7). Several researchers have recently explored this model,
utilising links between pupil-size, LC activity and task exploitation (e.g., Gilzenrat et
al., 2010; Jepma, & Nieuwenhuis, 2011; Murphy et al., 2011).
223
Murphy et al. (2011) used an extended auditory oddball paradigm to see whether, on
a trial by trial basis, P300 also indexed fluctuations in task performance predicted by
Adaptive Gain Theory (AGT; Aston-Jones, & Cohen, 2005; see Chapter 1, section
1.2.4.1), and how it related to tonic and phasic changes in pupil diameter. Twentyfour participants were asked to respond to 1000Hz target tones presented on 20% of
trials and ignore the remaining 500Hz standard tones presented the other 80% of the
time. They found that both P300 and pupil-size reflected changes in task
engagement as described by AGT. On trials with an intermediate prestimulus pupilsize, and large stimulus-evoked pupil dilations, P300 amplitudes were found to be
large, and task performance better, than when prestimulus pupil-size was larger or
smaller, and stimulus-evoked dilations were smaller. This pupillary behaviour was
assumed to reflect intermediate tonic LC activity interspersed with phasic bursts of LC
activity, consistent with the operation of the “phasic” LC mode thought to promote
task engagement. Murphy et al. (2011) concluded that, in addition to pupil-size, the
P300 may also index LC exploration/exploitation mode.
Kuipers and Thierry (2011) recorded concurrent pupil-size and ERPs to investigate
the relationship between semantic integration, reflected by the N400 component, and
phasic pupil dilation influenced by the LC. Maximal pupil-size usually occurred at
least 1000ms after stimulus presentation onset (Beatty, 1982b), however, Steinhauer
and Hakerem (1992) observed an initial peak dilation beginning 200ms after stimulus
onset, reaching maximum amplitude between 500–600ms, only slightly later than the
N400. Kuipers and Thierry (2011) investigated this early pupil peak in conjunction
with ERPs by presenting participants with semantic matching/non-matching spoken
word-picture and picture-spoken word pairs, and asking them to passively attend
rather than to engage in a task. In the word-picture condition, the N400 amplitude
was larger for matching than non-matching pairs, and pupil dilations were larger for
224
non-matching pairs than matching pairs. In the picture-word condition, the N400
showed larger amplitude for matching than non-matching pairs, but pupil dilations did
not differ between conditions. Despite these findings, the authors focussed on
spurious positive correlations between pupil-size and ERP waveforms at 16ms
intervals throughout an 850ms epoch, and interpret results from the entire window in
relation to the N400. Later significant differences in pupil-size (from 366ms onward)
in the word-picture condition extended to the end of the 850ms epoch (and likely
extended beyond it) but were not analysed.
The interpretation of Kuipers and Thierry’s (2011) data is difficult. As is clear from the
reported waveforms, there were large differences between the conditions before even
comparing matching vs. non-matching item pairs, but Kuipers and Thierry (2011)
performed separate ANOVAs for each condition disallowing any test for a significant
effect of condition. The analysis also failed to provide a sense of the topographical
distribution of the effects found by the authors. Major light-reflex confounds were
introduced because display brightness decreased from high to low in the word-picture
condition and increased from low to high in the picture-word condition. Additionally
the pupil data in the picture-word condition were incorrectly baselined. There was
also a more subtle confound of stimulus repetition (each pair was repeated twice in
each condition) which was not included as a factor in the analyses – repeated
mismatched pairs might be more memorable than matching pairs, and therefore be
encoded more strongly (e.g., Otero et al., 2011), thus leading to a larger pupil dilation
to non-matching pairs than matching pairs; alternatively these items may cause
repetition suppression (e.g., Schacter, & Buckner, 1998), leading to a smaller pupil
dilation. The authors also over-interpreted their findings in line with LC-NE influences
on task performance, stating that there is no functional connection between the
auditory orienting response and pupil dilation, despite acknowledging that auditory
225
neurons are known to respond to NE (only released by the LC) in the monkey (Foote
et al., 1975) and that pupil dilation can be triggered by auditory input (Beatty, 1982a).
They also ignored any possible link between small pupil dilation and decreased
phasic LC firing due to task disengagement, when there was no task, and participants
were asked to passively look at the screen (Gilzenrat et al., 2010).
The role of the N400 as a sensitive measure of context mismatch was also omitted.
Instead the authors concluded that changes in pupil-size in their study were due to
accommodation, and that decreased phasic LC firing increased the “effort” involved in
semantic integration (as measured by the N400), which decreased pupil dilations.
The continuation of this line of argument is that larger phasic LC input, which would
increase semantic integration efficiency, would also reduce effort (and the N400) but
increase pupil-size. We have only to look at the wealth of literature spanning the last
six decades, demonstrating increases in pupil-size associated with cognitive effort, to
question both the results and the conclusion (for a review see Beatty, & LuceroWagoner, 2000; Beatty, 1982; Granholm, & Steinhauer, 2004; Hess, & Polt, 1964;
Janisse, 1977; Kahneman, 1973).
Few studies measure concurrent ERP and pupil-size. Van Droof et al. (2010)
indirectly tested recognition memory for words by looking at receptive vocabulary
knowledge in nonverbal autistic participants and found that peak dilation was larger
for known words than unknown words and that the N400 was enhanced for
mismatched known words. Stone and Rothenheber (1992) added EEG and
pupillometry to traditional polygraph measures during an oddball experiment where
participants were asked to count instances of a known photograph among a series of
unknown photos. They concluded that: “Although these results were encouraging our
226
findings at this point are inconclusive thus warranting additional study” (Stone, &
Rothenheber, 1992, p.73).
6.2. Experiment 8 – Strength of Memory Effect
Experiment 8 employs a novel approach to understanding recognition memory,
concurrently measuring pupil-size and ERPs to study the effects of a memory
strength manipulation. The procedure and design were based on an ERP study by
Finnigan et al. (2002) who recorded continuous EEG whilst participants performed an
old/new recognition test on items that were unstudied (new), studied once (weak) or
studied three times (strong) during learning. Consistent with Van Petten et al. (1991),
they found a graded FN400 component which had a more negative-going amplitude
for new items than weak, and for weak than strong items. Like Yovel and Paller
(2004), their early old/new effect was maximal over parietal electrode sites. They
found larger amplitude of the LPC component (between 500-700ms) for strong items
than weak and for weak than new items. The LPC amplitude was also larger for
correct than incorrect decisions, with maximal amplitude at centro-parietal electrodes.
This design was selected because of the graded effect of the memory manipulation
on the psychophysiological responses, and it was hoped that it would also produce a
graded pupil response in that PDR for strong items would be larger than for weak
items, which would be larger than for new items. Comparable memory strength
effects in the two measures would suggest they index the same underlying events
and provide support for the idea that the PONE reflects mnemonic processes. The
manipulation was expected to work because stimulus repetition has been shown to
enhance memory performance on behavioural measures (e.g., Leding, & Lampinen,
2009; Yonelinas, 2002).
227
6.2.1.
Method
6.2.1.1.
Participants
Twenty-two right-handed native English speaking participants (9 male; age range:
19.3-50.6, M = 26.0, SD = 1.67), with normal or corrected-to-normal vision in at least
one eye and no self-reported psychiatric or neurological conditions, were recruited
from the student psychology participation pool at the University of Sussex and
through personal contact. Participants were briefed with a detailed consent form and
verbal description, and invited to ask questions. Written consent was obtained prior
to testing and participants were fully debriefed at the end. The experiment was
approved by the relevant ethics committee. Four participants failed to contribute
more than thirty artefact-free correct ERP trials to all three item-types and were
excluded from the analysis.
6.2.1.2.
Materials/Apparatus
Three study lists were created for the learning phase, each list comprising 60 nouns
selected from the MRC Psycholinguistic Database, half of which were included three
times in the respective learning list. For the recognition test, three lists were
constructed, each containing the 30 items that were presented once on the
corresponding study list (“weak”), the 30 items that were presented three times on the
study list (“strong”), and 30 new nouns that had not previously been seen (“new”).
All items were 5 letters long, matched for frequency, familiarity and imageability,
according to the K-F norms (frequency range = 10-40, M = 20.3; familiarity range =
351-618, M = 515; imageability range = 293-632, M = 507). The three parallel sets of
study lists and recognition tests formed blocks A to C, and were presented on a
computer monitor in white 20pt Monospaced font on a black background under fixed
228
illumination. Words were presented using E-Prime 2.0 software (Psychology
Software Tools Inc, Pennsylvania) on a 17” CRT monitor, which participants viewed
from a distance of 50cm and the visual angle subtended by the words was ~3o. Eye
movements were recorded with desk mounted EyeLink 1000 (SR-Research, Ontario),
with a sampling rate of 500Hz. All items are presented in Appendix G.
The experiment took place inside a Faraday cage. Continuous EEG recordings were
acquired from the scalp by a Net Amp and 128 electrode dense-array Geodesic
Sensor Net (Tucker, 1993), in conjunction with Net Station software package
(Electrical Geodesics Inc, Oregon), filtered online by bandpass 0.01-100Hz and
digitized at a sampling rate of 500Hz. Both the EEG and eye movement recordings
were triggered simultaneously by E-Prime; Net Station commands were sent via an
Ethernet cable by the E-Prime Net Station extension, and EyeLink commands via a
modified parallel cable and a custom E-Prime script that turned the cable pins on and
off to stop and start eye-tracker recording. Messages indicating the beginning and
end of each trial, and the onset and offset of stimuli presentation were also sent to
both Net Station and EyeLink in order that pupil and EEG trials could be aligned. Net
Station also received additional trial messages including item-type and participant
responses, which were made using a button box.
6.2.1.3.
Design and Procedure
In a within-subject design participants completed 3 recognition blocks under standard
instructions. Each block contained a learning phase and a recognition phase. During
the learning phase, 120 study list target items (30 items presented once, 30 items
presented three times) were displayed on screen for 1000ms with 200ms of blank
screen between words, and participants were asked to remember the items. During
the recognition phase, 90 list items (30 new, 30 weak, 30 strong) followed a 500ms
229
fixation cross and a 1000ms mask of “HHHHH”, and were presented on screen for
1000ms before being remasked for 1000ms. Participants were then presented with a
sign indicating that they could blink, and after 400ms a response prompt appeared
that remained on screen until they responded (see Figure 6-1).
OLD NEW
(response)
HHHHH
(1000ms)
WORDS
(1000ms)
HHHHH
(1000ms)
+
(500ms)
Figure 6-1: Experimental procedure.
Participants were asked to wash and brush their hair before the application of the
electrodesic net to their head. Once seated in the Faraday cage they were required
to use a chin rest to enable accurate eye-tracking. Participants were reminded at the
start of each recognition block that their eye movements and brain waves were being
recorded and to remain still and blink only when prompted by the blink screen; they
were also shown the impact on the EEG traces of blinks and eye-movements.
Participants were prompted to press a button to indicate whether the word was old
(target previously encountered in the learning phase) or new (not previously
encountered). This response screen was replaced by a fixation cross in the centre of
the screen before presentation of the mask followed by the next item.
230
Using standard recognition instructions, participants were asked to perform as
accurately as possible. In a within-subject design all participants viewed all study and
recognition lists once. To control against list and order effects, items were
randomised within blocks, and blocks were rotated across participants. Instructions
were repeated at the beginning of each block and participants were able to take a
break or initiate the next block when they felt ready with a verbal response. A
response device with two buttons corresponding to “old/new” answers was provided
and button configuration was counterbalanced across participants. Old/new
judgements were recorded, by the computer running E-Prime, after each recognition
item. Maximum pupil-size was recorded by the EyeLink host computer during the
time the item was on screen during the recognition test, and EEG activity was
recorded continuously by the Net Station host. Preparation and experimental
procedure lasted approximately 1 hour, with the task lasting around 29 minutes.
6.2.1.4.
Pupil Recording
Maximum pupil-size was recorded from the right eye during each recognition period.
A Pupil Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated
expressing the maximum pupil-size for each 2000ms recognition trial as a proportion
of the maximum pupil-size during that trial’s 200ms baseline.
6.2.1.5.
Electrophysiological Recording and Analysis
EEG was continuously recorded with a vertex reference. Vertical and horizontal eyemovements were monitored by using two bipolar ocular electrodes. Impedance was
kept below 50kΩ. Sampling rate was 500Hz and an online 0.01-100Hz band-pass
filter was used. Offline the continuous EEG was segmented into epochs from 200ms
before to 1000ms after stimulus onset. Segments with artifacts exceeding +/- 75μV
were automatically rejected and electro-oculogram (EOG) artifacts were detected
231
using Net Station Waveform Tools software package (Electrical Geodesics Inc,
Oregon). Manual eye artifact rejection was also used because Net Station algorithms
eliminated too many trials; in total the average number of trials lost to artifact rejection
per participant per condition were: 9.47 new, 4.76 weak, and 7.79 strong. Final
average trial numbers per participant per condition were: 63.61 new, 46.72 weak, and
57.83 strong. Bad channels were replaced by spline-constructed data from adjacent
channels using Waveform Tools for creation of topographical maps (average 1.59
channels per participant); however reconstruction was not performed for channels
included in the statistical analysis. Segments were baseline-corrected over the
200ms pre-stimulus interval and re-referenced to the average mastoid electrode.
Separate grand-average ERPs were computed for strong hits and misses, weak hits
and misses, correct rejections and false alarms. Only correct trials were included in
statistical analyses. Data were discarded from participants for whom ERP averages
did not comprise at least 30 artifact-free trials (4 participants in total). Grand-averages
were low-pass filtered at 40Hz prior to plotting and after statistical extraction.
6.2.2.
Results
6.2.2.1.
Behavioural Data
The proportions of correct responses to new, weak (presented once at learning) and
strong (presented three times at learning) items were calculated and averaged 81.2%
(SD = 11.9%) correct new items, 57.2% (SD = 14.3%) correct weak items and 72.9%
(SD = 12.0%) correct strong items (see Figure 6-2). A one-way repeated-measures
ANOVA revealed a significant main effect of item-type (F(1.17,19.8) = 18.1 MSE =
0.025, p <.001, ηp2 =.516). Mauchly’s test indicated that the assumption of sphericity
had been violated (χ2(2) = 20.1, p <.001), therefore degrees of freedom were
corrected using Greenhouse-Geisser estimates of sphericity (ε = 0.583). Bonferroni-
232
corrected subsidiary t-tests revealed that, in general, participants correctly identified
significantly fewer weak items than strong (t(17) = 9.65, p <.001, r =.846) or new
Proportion Correct Responses
items (t(17) = 4.81, p <.001, r =.577).
p < .001
0.90
0.85
0.80
0.75
0.70
0.65
0.60
0.55
0.50
0.45
0.40
p < .001
New
Weak
Strong
Item type
Figure 6-2: Proportion of correct responses to new, weak and strong items. Error bars show standard
error of mean.
6.2.2.2.
Pupil-Size Data
Average PDR was calculated for correctly identified new, weak (presented once at
learning) and strong items (presented three times at learning). As PDR is a function
of baseline pupil-size, baseline pupil-sizes for old and new items in each condition
were compared to ensure that any differences in PDR were not due to baseline
differences. The difference was not significant (F(2,34) = 1.78, p >.05, ns, ηp2 =.095).
A one-way repeated measures ANOVA, with within-subject factor of item-type (new
vs. weak vs. strong) showed a main effect of item-type (F(2,34) = 9.02, MSE < 0.001,
p <.001, ηp2 =.347). Planned t-tests revealed that in general participants’ pupils
dilated more to correctly identified weak old items (M = 1.138, SD = 0.077) than to
correctly identified new items (M = 1.118, SD = 0.0717, t(17) = 5.51, p <.001, r =.641).
Counter to predictions, participants’ pupils also dilated more to correctly identified
weak old items than to correctly identified strong old items (M = 1.127, SD = 0.0817,
233
t(17) = -2.09, p <.05, r =.204), and only approached significance for new and strong
items (t(17) = 1.78, p =.09, r =.157; see Figure 6-3).
p = .05
Average Pupil Dilation Ratio
p < .001
1.16
1.15
1.14
1.13
1.12
1.11
1.10
1.09
New
Weak
Strong
Item type
Figure 6-3: Pupil dilation ratio for correctly identified new, weak and strong items. Error bars show
standard error of mean.
6.2.2.3.
Event-Related Potentials
Figure 6-4 shows grand-average ERP waveforms at midline frontal, central and
parietal electrodes for correctly identified new, weak and strong items. Waveforms
diverge at around 300ms post-stimulus, and last for the rest of period of interest (the
following 700ms). Inspection of the grand-average ERPs reveals a negative-going
waveform between 300-500ms with a centro-parietal distribution, which is larger in
amplitude for new and strong items compared to weak items, and maximal at P7.
This component is similar in polarity and timing to the mid-frontal old/new effect (e.g.,
Mecklinger, 2000; Wiese, & Daum, 2006) but has a more posterior distribution. This
is followed by a positive-going waveform between 500-700ms with a parietal
distribution, which is larger in amplitude for weak and strong items compared to new
items, and is also maximal at P7; this component is similar in polarity, timing and
distribution to the LPC old/new effect (e.g., Johnson, 1995; Allan et al., 1998). At
frontal electrodes there is also a negative-going waveform between 500-700ms which
is larger in amplitude for strong items than new or weak items.
234
Figure 6-4: Grand average (N = 18) ERPs for correctly identified new, weak (presented once) and
strong (presented three times) items at midline frontal, central and parietal electrodes. Mean
numbers of individual ERP trials per strength condition per participant were: new: 63.61; weak: 46.72;
strong: 57.83. The scale bar indicates amplitude (in µV) and time course of activity (in ms). Positive
plotted upwards.
Figure 6-5: Topographical distribution of old/new differences in mean amplitude (µV) for weak items (first row) and strong items (second row) between 300-700ms.
235
236
The topography of the old/new effects is illustrated in Figure 6-5, which highlights the
posterior distribution of the old/new difference waves elicited by weak and strong
items in both the 300-500ms and 500-700ms time windows, and a left hemisphere
distribution of the reversed old/new effect for strong items in the 300-500ms time
window.
Consistent with previous studies into the neural correlates of familiarity and
recollection, and to facilitate comparison with the old/new effects found by other labs
(e.g., Allan et al., 1998, Curran, 2000; Friedman, & Johnson, 2000; MacKenzie, &
Donaldson, 2007; Mecklinger, 2000; Rugg et al., 1998a; Wilding, & Sharpe, 2003),
ERPs were quantified for analysis by computing the mean amplitude relative to the
mean of the 200ms pre-stimulus baseline period for 300-500ms and 500-700ms
post-stimulus. Separate within-subjects ANOVAs were conducted for the two time
windows using electrodes equivalent to F7, F3, Fz, F4, F8, T7, C3, Cz, C4, T8, P7,
P3, Pz, P4, P8 (see Chapter 2, section 2.2.1) in analyses of location (including
factors of caudal position, hemisphere, and site). The old/new effects were
characterised separately for the weak and strong conditions, and finally comparisons
were made between conditions. As ERP effects are only of interest when they
reflect differences between conditions, only significant effects involving the factor of
condition are reported. The literature reports qualitatively separate old/new effects
involving different ERP components in the two time windows; therefore analysis over
time was not carried out.
6.2.2.4.
Old/New Effects for Weak and Strong Items
The old/new effects were characterised separately for the weak and strong
conditions. Mean amplitude was analysed in a 2 x 2 x 2 x 3 repeated measures
237
ANOVA with within-subject factors of item-type (old, new), hemisphere (left, right),
site (superior, inferior) and caudal position (frontal, central, parietal).
6.2.2.5.
Weak Old vs. New Items
300-500ms Time Window
Analysis of the 300-500ms time window revealed no main effects. There was a
significant item-type by caudal position interaction (F(2,34) = 6.03, MSE = 2.02, p
<.01, ηp2 =.262), reflecting differences between new and weak ERPs at parietal
electrodes that were absent at frontal and central sites. A priori t-tests showed that
these differences were significant at parietal electrodes (t(17) = 3.74, p <.01, r
=.451), demonstrating the presence of an early old/new effect at parietal electrodes
(see Figure 6-6).
New
Weak
Mean Amplitude (uV)
3.0
p < .01
2.5
2.0
1.5
1.0
0.5
0.0
Frontal
Central
Parietal
Caudal Position
Figure 6-6: Mean amplitude for new and weak items at frontal, central and parietal electrodes at 300500ms. Error bars show standard error of mean.
The 3-way item-type by caudal position by site interaction just failed to reach
significance (F(2,34) = 2.76, MSE = 0.5, p =.077, ηp2 =.140) – in general the
old/new effect exhibited a superior distribution at parietal sites. Examination of
238
the data revealed that the old/new effect was maximal at P3 (t(17) = 2.30, p <.05,
r =.238).
500-700ms Time Window
Analysis of the 500-700ms time window revealed that the main effect of item-type
just failed to reach significance (F(1,17) = 3.42, MSE = 10.2, p =.08, ηp2 =.167) – in
general the mean amplitude was more positive-going for weak items compared to
new items. There was a significant item-type by caudal position interaction
(F(1.35,23.0) = 3.80, MSE = 4.23, p <.05, ηp2 =.183), reflecting differences between
new and weak ERPs at parietal electrodes that were absent at frontal and central
sites. A priori t-tests showed that these differences were significant at parietal
electrodes (t(17) = 4.07, p <.001, r =.494), demonstrating the presence of a late
old/new effect at parietal electrodes (see Figure 6-7).
New
Weak
4.0
p < .001
Mean Amplitude (uV)
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Frontal
Central
Parietal
Caudal Position
Figure 6-7: Mean amplitude for new and weak items at frontal, central and parietal electrodes at 500700ms. Error bars show standard error of mean.
The 3-way item-type by caudal position by site interaction was also significant
(F(2,34) = 5.43, MSE = 0.917, p <.01, ηp2 =.242) – the old/new effect exhibited a
superior distribution at parietal sites. Examination of the data revealed that the
old/new effect was maximal at P3 (t(17) = 2.30, p <.05, r =.238).
239
6.2.2.6.
Strong Old vs. New Items
300-500ms Time Window
Analysis of the 300-500ms time window revealed a significant item-type by caudal
position interaction (F(1.40,23.7) = 3.54, MSE = 3.75, p <.05, ηp2 =.172; Mauchly’s
test indicated that the assumption of sphericity had been violated (χ2(2) = 9.08, p
<.05), therefore degrees of freedom were corrected using Greenhouse-Geisser
estimates of sphericity (ε = 0.698)). The interaction reflected larger differences
between new and strong ERPs at frontal electrodes than at central and parietal
sites. Although on average new items were more negative-going than strong items
at parietal electrodes, a priori t-tests showed that these differences were not
significant (t(17) = 0.96, p >.05, ns), and new and strong items did not differ
significantly at frontal (t(17) = 1.26, p >.05, ns) or central electrodes (t(17) = 0.85, p
>.05, ns; see Figure 6-8).
New
Strong
Mean Amplitude (uV)
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Frontal
Central
Parietal
Caudal Position
Figure 6-8: Mean amplitude for new and strong items at frontal, central and parietal electrodes at
300-500ms. Error bars show standard error of mean.
The item-type by hemisphere interaction just failed to reach significance (F(1,17) =
3.39, MSE = 2.39, p =.08, ηp2 =.166) – in general the old/new effect was larger over
the left hemisphere. The item-type by caudal position by site 3-way interaction was
240
significant (F(2,34) = 3.85, MSE = 0.631, p <.05, ηp2 =.185) – the old/new effect
exhibited a superior distribution at frontal sites. Examination of the data revealed
that the old/new effect was maximal (albeit reversed in polarity) at F3 (t(17) = 2.10, p
<.05, r =.206).
500-700ms Time Window
Analysis of the 500-700ms time window revealed a significant item-type by caudal
position interaction (F(1.47,25.1) = 5.63, MSE = 4.12, p <.01, ηp2 =.249; Mauchly’s
test indicated that the assumption of sphericity had been violated (χ2(2) = 70.7, p
<.05), therefore degrees of freedom were corrected using Greenhouse-Geisser
estimates of sphericity (ε = 0.737)). The interaction reflected larger differences
between new and strong ERPs at frontal and central electrodes than at parietal
sites. Although on average strong items were more positive-going than new items at
parietal electrodes, a priori t-tests showed that this difference was not significant
(t(17) = 1.14, p >.05, ns), and new and strong items did not differ significantly at
central (t(17) = 1.53, p >.05, ns) or frontal electrodes (t(17) = 1.81, p =.09, r = .162;
see Figure 6-9).
New
Strong
3.0
Mean Amplitude (uV)
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
Frontal
Central
Parietal
Caudal Position
Figure 6-9: Mean amplitude for new and strong items at frontal, central and parietal electrodes at
500-700ms. Error bars show standard error of mean.
241
The interactions between item-type and hemisphere (F(1,17) = 3.11, MSE = 1.97, p
=.096, ηp2 =.155), and item-type and site (F(1,17) = 3.08, MSE = 0.827, p =.098, ηp2
=.153) just failed to reach significance – in general the old/new effects were larger
over the left hemisphere and at superior sites. The item-type by caudal position by
site 3-way interaction was significant (F(2,34) = 6.32, MSE = 0.854, p <.01, ηp2
=.271) – the old/new effect exhibited a superior distribution at frontal sites.
Examination of the data revealed that the old/new effect was maximal (albeit
reversed in polarity) at Fz (t(17) = 2.65, p <.05, r =.293).
6.2.2.7.
Effect of Presentation Frequency on Old/New Effect
Difference waves were calculated for weak minus new items, and strong minus new
items, to allow direct comparison of the magnitude and distribution of the old/new
effects in each condition. Mean amplitude difference was analysed in a 2 x 2 x 2 x 3
repeated measures ANOVA with within-subject factors of condition (weak, strong),
hemisphere (left, right), site (superior, inferior) and caudal position (frontal, central,
parietal).
300-500ms Time Window
Analysis of the 300-500ms time window revealed that the main effect of condition
just failed to reach significance (F(1,17) = 3.22, MSE = 12.0, p =.09, ηp2 =.159) – in
general the old/new effect was larger in the weak condition than in the strong
condition. The interactions between condition and hemisphere (F(1,17) = 3.91, MSE
= 2.93, p =.065, ηp2 =.187), and condition and site (F(1,17) = 3.27, MSE = 0.61, p
=.089, ηp2 =.161) just failed to reach significance – in general differences between
old/new effects in the two conditions were larger over the left hemisphere and at
242
superior sites. Examination of the data revealed that the difference between old/new
effects in the two conditions was maximal at P3 (t(17) = 3.05, p <.01, r =.353).
500-700ms Time Window
Analysis of the 500-700ms time window revealed a significant main effect of
condition (F(1,17) = 10.5, MSE = 9.94, p <.01, ηp2 =.383) – the old/new effect was
larger in the weak condition than in the strong condition. The 3-way condition by
caudal position by site interaction was also significant (F(2,34) = 4.06, MSE = 0.65, p
<.05, ηp2 =.193), reflecting differences between old/new effects in the two conditions
were largest at inferior sites and frontal electrodes. Examination of the data
revealed that the difference between old/new effects in the two conditions was
maximal at F7 (t(17) = 2.12, p <.05, r =.208).
6.2.2.8.
Early Effects (80-150ms)
From visual inspection of the waveforms, a very early negative-going effect was
observed for strong items relative to new items at frontal electrodes, between 80150ms, which resembles the N1 component associated with attention orientation.
Therefore an additional analysis was performed within the 80-150ms time window in
a 3 x 2 x 2 x 5 repeated measures ANOVA with within-subject factors of item-type
(new, weak, strong), hemisphere (left, right), site (superior, inferior) and caudal
position (frontal, central, parietal).
Analysis revealed a significant item-type by hemisphere interaction (F(2,34) = 6.51,
MSE = 1.796, p <.01, ηp2 =.277), reflecting an old/new effect for strong items that
was reversed in polarity over the left hemisphere. The 3-way item-type by
hemisphere by site interaction was also significant (F(2,34) = 3.55, MSE = 0.235, p
<.05, ηp2 =.173) – this old/new effect exhibited an inferior distribution for strong items
243
but a superior distribution for weak items over the right hemisphere. The 3-way
item-type by caudal position by site interaction just failed to reach significance
(F(4,68) = 2.17, MSE = 0.254, p =.08, ηp2 =.113) – in general the old/new difference
for strong items exhibited an inferior distribution at parietal sites.
6.2.3.
Discussion
Experiment 8 sought to replicate an ERP study which used a memory strength
manipulation to produce graded early and late ERP old/new effects, and to extend
the study to include concurrent pupil-size recording, with the aim of also observing a
graded pupil-size response.
The pupillometry findings showed a main effect of item-type – in general participants’
pupils dilated more in response to correctly identified weak old items than to
correctly identified new items. Somewhat surprisingly, and counter to predictions,
participants’ pupils also dilated more to correctly identified weak old items than to
correctly identified strong old items, and the difference between new and strong
items only approached significance.
ERP findings in the 300-500ms window showed an old/new effect at parietal
electrodes for weak items, and although the literature generally reports mid-frontal
old/new effects, some studies have reported parietal old/new effects in the 300500ms time window (MacKenzie, & Donaldson, 2007; Yovel, & Paller, 2004),
including the study that was being replicated here (Finnigan et al., 2002). Contrary
to predictions, strong items did not show enhanced positivity compared to new
items; instead, the old/new effect for strong items, which was exhibited at frontal
electrodes, was reversed in polarity – new items were more positive than strong old
items.
244
In the 500-700ms window there was also an old/new effect at parietal electrodes for
weak items, consistent with the literature (e.g., Curran 2004; Düzel et al., 1997;
Friedman, 2004; Rugg et al., 1998b; Smith, 1993; Trott et al., 1999). This was not,
however, the case for strong items which, as for the earlier time window, did not
show enhanced positivity compared to new items; instead, the old/new effect for
strong items, which was exhibited at frontal electrodes, was reversed in polarity –
new items were more positive than strong old items.
To try to understand whether the reversal of the old/new effect for strong items was
due to an absence of enhanced positivity of the FN400 and LPC components, or
whether it could be due to modification of an earlier component, a follow-up analysis
was performed on the 80-150ms window, to see whether the memory-strength
manipulation may have influenced early attention processes such as those reflected
by the N1. Analysis revealed reversed polarity old/new differences for strong items
over the left hemisphere than the right hemisphere. In the later time windows,
interactions with hemisphere were not found to be significant, however there was a
trend for hemisphere differences in both time windows in the analysis of the triple
presentation condition (strong items in relation to new items). In general the strong
old/new effect was larger over the left hemisphere than the right hemisphere in both
the 300-500ms and 500-700ms time windows.
Comparing the old/new effects between the conditions in the 300-500ms time
window, an interaction approaching significance suggested the old/new effect in the
weak condition was slightly larger over the left hemisphere than the right
hemisphere, whereas in the strong condition the old new effect was much larger
over the right hemisphere than the left hemisphere. There appear to be laterality
differences between weak and strong words, and single and triple presentation
245
old/new effects, which begin as early as 80-150ms after stimulus presentation and
persist in the later components. The recognition positivity reported in the literature is
usually maximal over the left hemisphere, suggesting that in this experiment it may
be absent or reduced for strong items.
6.3. Experiment 9 – Pupil and Behavioural Data Only
Experiment 8 yielded the surprising finding that rather than a graded strength of
memory effect, both ERPs and pupillometry results demonstrated only a very weak
old/new effect for strongly encoded items (those that were repeated three times
during study), but a much more robust old/new effect for the weakly encoded items
(which were only encountered once during study). Therefore, Experiment 9 was
carried out as a near-identical behavioural/pupillometry only replication of
Experiment 8, in the laboratory environment used for Experiments 1-7, to determine
whether this pattern of results was attributable to the task itself (which differed
considerably from those used in previous experiments), or whether it was somehow
associated with the ERP testing environment and procedure. For example, the
illumination level in the EEG room is low compared to the room in which the other
experiments were performed, and it is possible that the procedure itself increased
overall levels of arousal in the participants.
6.3.1.
Method
6.3.1.1.
Participants
Thirty-one participants (7 male; age range: 18.2-43.5, M = 21.7, SD = 6.79) with
normal or corrected-to-normal vision were recruited from the psychology coursecredit and subject pools at the University of Sussex, and through personal contact.
246
6.3.1.2.
Materials/Apparatus/Design/Procedure
Stimuli, design and procedure were the same as in Experiment 8 (see sections
6.2.1.2 and 6.2.1.3) with the following changes: Study and recognition lists were
presented in black 20pt Monospaced font on a light grey background under fixed
illumination. Words were presented using Experiment Builder software (SRResearch, Ontario) on a 21” CRT monitor. Participants viewed the monitor from a
distance of 70cm and the visual angle subtended by the words was approximately
3o. Eye movements were recorded with an EyeLink II (SR-Research, Ontario), with
a sampling rate of 500Hz. All items are presented in Appendix G.
6.3.1.3.
Pupil Recording
Maximum pupil-size was recorded from the right eye during each recognition period.
A Pupil Dilation Ratio (PDR; see Chapter 2, section 2.1.2.1) was calculated
expressing the maximum pupil-size for each 2000ms recognition trial as a proportion
of the maximum pupil-size during that trial’s 200ms baseline.
6.3.2.
Results
6.3.2.1.
Behavioural Data
The proportions of correct responses to new, weak (presented once at learning) and
strong (presented three times at learning) items were calculated and averaged
80.4% (SD = 8.3%) correct new items, 54.1% (SD = 13.3%) correct weak items and
73.9% (SD = 13.5%) correct strong items. A one-way repeated-measures ANOVA
revealed a significant main effect of item-type (F(1.14,34.3) = 53.1, MSE = 0.019, p
<.001, ηp2 =.639). Mauchly’s test indicated that the assumption of sphericity had
been violated (χ2(2) = 40.3, p <.001), therefore degrees of freedom were corrected
247
using Greenhouse-Geisser estimates of sphericity (ε = 0.571). Bonferroni-corrected
subsidiary t-tests revealed that in general participants correctly identified significantly
fewer weak items than strong (t(30) = 20.36, p <.001, r =.932) or new items (t(30) =
Proportion Correct Responses
8.29, p <.001, r =.696; see Figure 6-10).
p < .001
0.90
0.85
0.80
0.75
0.70
0.65
0.60
0.55
0.50
0.45
0.40
p < .001
New
Weak
Strong
Item type
Figure 6-10: Proportion of correct responses to new, weak and strong items. Error bars show
standard error of mean.
6.3.2.2.
Pupil-Size Data
Average PDR for old and new items was calculated for correctly identified new,
weak and strong items. As PDR is a function of baseline pupil-size, baseline pupilsizes for old and new items in each condition were compared to ensure that any
differences in PDR were not due to baseline differences. The difference was not
significant (F(2,60) = 1.69, p >.05, ns, ηp2 =.053).
Average PDR for correctly identified new, weak and strong items were compared in
a one-way repeated measures ANOVA, which showed a main effect of item-type
(F(2,60) = 7.07, MSE < 0.001, p <.01, ηp2 =.191). Planned t-tests revealed that in
general participants’ pupils dilated more to correctly identified strong old items (M =
1.076, SD = 0.0278) than to correctly identified new items (M = 1.064, SD = 0.0306,
248
t(30) = 4.85, p <.001, r =.440). There was a trend for participants’ pupils to dilate
more to correctly identified strong old items than to correctly identified weak items (M
= 1.069, SD = 0.0292; t(30) = 1.90, p =.06, r =.108; see Figure 6-11).
Average Pupil Dilation Ratio
Maximum
p = .06
1.085
p < .001
1.080
1.075
1.070
1.065
1.060
New
Weak
Strong
Item type
Figure 6-11: Pupil dilation ratio for correctly identified new, weak and strong items. Error bars show
standard error of mean.
6.3.3.
Discussion
Experiment 9 was designed to establish whether the strength of memory
manipulation adopted in Experiment 8 (Finnigan et al., 2002) would produce a
graded pupil response in addition to the old/new effect between weak and new
items. As in Experiment 8, in general participants correctly identified more strong
and weak items than new items, and the proportion of correctly identified items in
each category was roughly the same (new: 81.2% vs. 80.4%; weak: 57.2% vs.
54.1%; strong: 72.9% vs. 73.9%).
Participants’ pupils dilated more to correctly identified strong old items than to
correctly identified new items. There was also a trend for participants’ pupils to
dilate more to correctly identified strong old items than to correctly identified weak
items, with no significant difference between weak and new items. This result is in
contrast to Experiment 8 where participants’ pupils dilated more to correctly
249
identified weak old items than to correctly identified strong old items and correctly
identified new items. As such, it suggests that some feature of the ERP testing
environment led to the unexpected finding of a stronger pupil (and ERP) response to
the weak compared to strong items in experiment 8. This possibility is explored
further in the general discussion.
6.3.4.
General Discussion
Experiment 8 sought to replicate an ERP study which used a memory strength
manipulation to produce graded early and late ERP old/new effects (Finnigan et al.,
2002). Using a novel approach, it also sought to extend the study to include
concurrent pupil-size recording, with the aim of also observing a graded pupil-size
response, which might suggest that the two psychophysiological measures are
linked. Concurrent measurement was designed to allow the memory processes
from the same group of participants, during precisely the same task, to be quantified
in two different psychophysiological indices of recognition memory.
Results showed that for weak items (studied once during learning) there was a
PONE, and early and late ERP old/new effects at parietal electrodes. However,
contrary to predictions, for strong items (studied three times during learning) early
and late ERP old/new effects occurred only at frontal electrodes and were reversed
in polarity; in addition the PONE for strong items was absent. Therefore the
predicted graded strength of memory effect was not obtained in either ERP or
pupillometry data. In order to test the effects of the strength of memory manipulation
on pupil-size alone, Experiment 9 was carried out as a near-identical
behavioural/pupillometry only replication of Experiment 8, in the laboratory
environment used for Experiments 1-7. Experiment 9 showed that in contrast to
Experiment 8 there was a PONE for strong old items – participants’ pupils dilated
250
more to correctly identified strong old items than to correctly identified new items.
There was also a trend for participants’ pupils to dilate more to correctly identified
strong old items than to correctly identified weak old items, with no significant
difference between weak and new items.
The strength of memory manipulation was expected to be effective because stimulus
repetition has been shown to enhance memory performance on behavioural
measures (e.g., Leding, & Lampinen, 2009; Yonelinas, 2002) and previous studies
(Finnigan et al., 2002; MacKenzie, & Donaldson, 2007; Van Petten et al., 1991)
found graded ERP components using memory strength manipulations. However it
did not appear to work in Experiment 8. Analysis of the ERP mean amplitudes for
weak and new items revealed a parietal old/new effect at 300-500ms, and although
the literature generally reports frontal/central old/new effects, some studies have
reported parietal old/new effects in the 300-500ms time window (MacKenzie, &
Donaldson, 2007; Paller, Gonsalves, Grabowecky, Bozic, & Yamada, 2000; Yovel, &
Paller, 2004), including the study that was being replicated here (Finnigan et al.,
2002). The 500-700ms window demonstrated a parietal LPC old/new effect for
weak items, consistent with the literature (e.g., Curran, 2004; Düzel et al., 1997;
Friedman, 2004; Rugg et al., 1998b; Smith, 1993; Trott et al., 1999).
Analysis of the ERP mean amplitudes for strong and new items showed that strong
items did not show the expected enhanced positivity compared to new items in
either the 300-500ms or 500-700ms time window. Instead new items showed
enhanced positivity compared to strong items at frontal electrodes, reversing the
polarity of the old/new effect. This meant that the predicted graded memory effect
did not occur for ERPs at frontal, central or parietal electrodes. This was echoed in
the pupil-size data which showed a significant old/new difference for weak but not
251
strong items, similar to ERPs at parietal electrodes. Norman, Tepe, Nyhus and
Curran (2008) strengthened memory for some stimuli by presenting faces three
times (weak) or six times (strong) during learning. They demonstrated that the
strength manipulation had no impact on the FN400 old/new effect (300-500ms), and
that the LPC old/new ERP effect (400-800ms) was only present for weak items – the
positive-going waveform evoked for strong items was equivalent in magnitude to the
waveform for new items. Because participant-reported “remember” judgements
were reduced by the manipulation, whilst “know” judgements were not, Norman et al.
(2008) concluded that the interference caused by increasing the “strength” of items
affected recollection but not familiarity, and reduced the LPC old/new effect. This
explanation is consistent with the absence of a parietal LPC old/new for strong items
in Experiment 8, however it does not explain why the expected FN400 old/new effect
was also missing for strong items, or why in both time windows strong old/new
effects were reversed at frontal electrodes.
Although quantitatively different, weak and strong items demonstrated qualitatively
the same topographical changes, whereas the distribution for new items was
qualitatively different to both strong and weak items. This could indicate that an
earlier latent ERP component was modified for strong items as a result of the triple
presentation (for example an early negative component being larger, or an early
positive component being smaller than for weak items). This was explored in a
follow-up analysis in the 80-150ms time window (selected by inspection of the
grand-averages), to see whether the memory-strength manipulation may have
influenced early attentional processes such as those reflected by N1. Analysis of
the 80-150ms window revealed early old/new differences over the left hemisphere
for strong but not weak items. In the FN400 and LPC time windows (300-500ms and
500-700ms) hemisphere by condition interactions also showed a trend towards
252
larger old/new differences over the left hemisphere for strong items. There appear
to be laterality differences between weak and strong words, and single and triple
presentation old/new effects, which begin as early as 80-150ms after stimulus
presentation and persist in the later components FN400 and LPC.
If stimulus repetition meant that fewer processing resources needed to be allocated
to strong items, then this might account for differences between weak and strong
items in the 80-150ms time window. Target stimuli can be sorted and selected for
additional processing at a very early stage, so early components such as the N1 are
affected by attention (Hillyard, Hink, Schwent, & Picton, 1973; Näätänen, 1990),
therefore if a repeated strong stimulus becomes less interesting, and therefore less
attended, the N1 might be attenuated.
Repetition may lead to priming, whereby subsequent presentations of an item are
processed more quickly and efficiently (cf., implicit memory, Chapter 3; Tulving, &
Schacter, 1990), however, negative priming can lead to poorer episodic encoding for
highly primed items (Tipper, 1985; Wagner, Maril, & Schacter, 2000). A related
phenomenon is repetition suppression, where repeated stimuli produce less neural
activation than new stimuli (for review, see Schacter, & Buckner, 1998; Buckner et
al., 1998; Grill-Spector, & Malach, 2001; van Turennout et al., 2000) or stimuli
presented once at learning (e.g., weak; Jiang, Haxby, Martin, Ungerleider, &
Parasuraman, 2000).
Monkey single-cell recordings (Desimone 1996; Miller, & Desimone, 1994) and
human fMRI studies have shown that repetition leads to decreased activation in
brain regions involved in stimulus processing, such as the left inferior prefrontal
cortex (LIPC; Wagner, Desmond, Demb, Glover, & Gabrieli, 1997) and hippocampal
253
and parahippocampal regions (Brozinsky, Yonelinas, Kroll, & Ranganath, 2005;
Suzuki, Johnson, & Rugg, 2011), and is suggested to occur automatically during
learning (Wiggs, & Martin, 1998). Guo, Lawson and Jiang (2007) found that the late
posterior ERP repetition effect (>550ms) showed more positive-going ERP
amplitudes to items at initial presentation compared to repetition.
Whilst the pupil response was not graded as expected, it did echo the ERPs at
parietal electrodes in that there was an old/new effect in the expected direction for
weak items, and weak items elicited a larger response than strong items. The
similar pattern of results in the ERPs and PDR old/new effects suggests that both
measures index the same underlying cognitive processes, and, importantly, lends
support to Otero et al.’s (2011) argument that the PONE represents neurocognitive
activity underlying recognition memory. Similarly, like the potential ERP repetition
suppression effect which may explain findings in Experiment 8, Van Rijn, Dalenberg,
Borst, and Sprenger (submitted) found that the phasic pupil response to repeated
stimuli decreased by 2% per repetition, which could account for the equivalent effect
seen in the pupil response.
Behavioural performance was consistent between the two experiments, suggesting
that the implementation of Experiment 9 in a different laboratory was similar enough
to produce comparable task performance, and that the ERP procedure and
equipment did not massively distract participants from the stimuli. However, the
manipulation did not produce the expected graded pupil-size in either experiment,
although the results of Experiment 9 were more in line with the expected effect with
a PONE for strong items and weak items on average being intermediate in size
between strong and new (albeit non-significant).
254
A possible influencing factor in Experiment 8 was the use of a remote eye-tracker.
Unlike the head-mounted eye-tracker used in Experiment 9, remote cameras are
usually located 50-100cm from the eye, and therefore measure the pupil with lower
precision than head-mounted eye-trackers (see Chapter 2, section 2.1.1.1; Klingner,
2010; Marshall, 2002). It is possible that the use of a remote eye-tracker in these
experiments (necessitated by the ERP acquisition) reduced the precision of the pupil
measurement, and perhaps underestimating pupil-size when it was at its largest
(e.g., for strong items). Eye-tracking during Experiment 8 was more vulnerable to
loss of signal due to wires across the participant’s face and change of focus, and
was less amenable to correction except for between recording blocks for fear of loss
of ERP data. There was a 2% loss of correctly eye-tracked new trials only (81.2%
correct trials vs. 78.7% eye-tracked correct trials; t(17) = 10.7, p <.001, r =.871), and
the total number of tracked strong trials was 98.9%, compared to 99.8% weak (t(17)
= 4.50, p <.001, r =.544) and 100% of new trials (t(17) = 3.43, p =.003, r =.409).
Additionally, the ambient illumination in the room used for Experiment 8 was lower
than that used for Experiment 9, which would have influenced pupil-size.
Although the number of artefact-free correct trials as a percentage of correct trials
per item-type (and therefore the number that were included in each grand average)
was not significantly different (all ts < 2, ns), there were significantly fewer absolute
numbers of trials contributing to the ERP grand averages for weak items (M =
51.9%, SD = 15.1%) than for new (M = 70.7%, SD = 15.4%; t(17) = 3.91, p =.001, r
=.474), or strong (M = 64.3%, SD = 14.3%; t(17) = 8.53, p <.001, r =.810) items.
Although issues may arise with waveforms formed from differing numbers of trials,
because of the resultant differing signal-to-noise ratios (Luck, 2010; see Chapter 2,
section 2.2.2), this is more of a concern when measuring peak amplitude due to the
greater influence a spurious peak has over the peak measurement in averages
255
containing fewer contributing trials (Luck, 2010). Experiment 8 measured mean
amplitude, an unbiased measure even when trial numbers differ, and means that this
difference in numbers is unlikely to have biased the results (Luck, 2010).
In considering why the results of Experiment 8 are different to those of the replicated
study by Finnigan et al. (2002) several factors should be acknowledged. Although
as far as possible the study design and procedure were replicated, ultimately there
were methodological differences, for example Finnigan et al. (2002) used a 30
electrode cap with lower maximum impedance than the 128 electrode net used here.
Presentation duration was increased from 400ms to 1000ms at both learning and
recognition to bring the procedure in line with Experiments 1-7, and a mask of
“HHHHH” preceded and followed stimuli at recognition, rather than the blank screen
used by Finnigan et al. (2002), in order to minimise the influence of the light reflex.
Although they were the same length (5 letters), different stimuli were used in
Experiment 8 to those used by Finnigan et al. (2002), and font size and distance
from screen may also have differed as these were not provided. Therefore it is
possible that basic visual stimuli features influenced the ERP waveform, particularly
with respect to early visual components (Luck, 2005; Schloerscheidt, & Rugg, 2004).
Experiment 8 used an online vertex reference electrode, whereas Finnigan et al.
(2002) used physically linked earlobe electrodes as an online reference, which
although not biased towards either hemisphere, creates “a zero-resistance electrical
bridge between the hemispheres, distorting the voltage distribution and reducing
hemispheric asymmetries” (Luck, 2005, p. 107). Electrodes in Experiment 8 were
re-referenced offline to a virtual average mastoid after the left and right mastoid
recordings were checked for artifacts, whereas Finnigan et al. (2002) do not appear
to have re-referenced offline.
256
In the separate analyses of their two experiments, Finnigan et al. (2002) collapsed
data across responses, whereas the data for Experiment 8 was analysed for correct
items only to limit analysis to items most likely to elicit a genuine memory. In their
analysis of correct and incorrect trials, Finnigan et al. (2002) collapsed data over
their two experiments because they had insufficient incorrect trials to form a grand
average in either experiment separately. Although their experiments were very
similar (procedurally only the length of the recognition list varied), the study-test
repetition lag varied between experiments, and lag can itself influence repetition
effects such as suppression (e.g., Brozinsky et al., 2005).
Experiment 8 and the experiments by Finnigan et al. (2002) took place in different
laboratory environments with different researchers, equipment and sources of noise,
and ultimately a different group of participants, therefore it would not necessarily be
expected that the two produce the same results for any or all of the above reasons.
In addition, contrary to Finnigan et al. (2002), Opitz (2010) found no difference in late
parietal old/new ERP effects between items presented once and items presented
three times.
Despite focusing on two components, FN400 and LPC, repetition of learning items in
this manipulation may have had wider influence than the single ‘memory strength’
effect intended. If the manipulation affected more than these two components, then
this could explain the apparent lack of difference between new and strong items, for
example modulation of overlapping positive- or negative-going ERPs. The issue of
latent components makes it more difficult to interpret the waveform, which is a local
sum of voltage differences. A reduction of the amplitude of the FN400 or LPC
components, as manifest in the grand-average, may not reflect a reduction in the
underlying neural activity of interest (see Luck, 2005, for further discussion).
257
Although the literature reports separate ERP components in the 300-500ms and
500-700ms time windows analysed, very similar patterns of results were reported for
the two time periods in Experiment 8. An interesting development of Experiment 8
might be to perform an analysis over time, to test whether or not the early and late
old/new effects were statistically different, therefore reflecting separate components.
258
7. General Discussion
Conclusions, Limitations and Priorities
7.1.1.
Summary
The central aim of this thesis was to explore the cognitive processes associated with
the recently reported Pupil Old/New Effect (PONE), whereby the pupil dilates to a
larger maximum size in relation to a baseline when participants view old items
compared to when they view new items during a recognition memory test. This
concluding chapter will summarise the key results of Experiments 1 to 9, noting
some of the limitations, and relating the findings back to the main issues in
pupillometry and recognition memory research outlined in the introductory chapter.
Finally, it will offer some suggestions for the future direction of this research.
Experiments 1 and 2 set out to replicate the PONE observed in explicit tests of
recognition memory and determine whether it would also be present in an “implicit”
test of memory using perceptual fluency. Results showed that the PONE was
replicated in a standard test of recognition memory, but not in an “implicit” test of
perceptual fluency, and it did not occur when participants were asked to read word
stimuli rather than make a recognition decision. Experiments 3 and 4 extended this
finding by examining whether the PONE would be present when recognition memory
was tested using artificial grammar learning, a form of implicit learning that relies on
a sense of familiarity to facilitate recognition. The PONE was again replicated in a
standard explicit recognition task, but was not present when participants were
judging grammatical vs. ungrammatical letter strings. Experiments 5 and 6
examined the effects of asking participants to deliberately perform poorly during
recognition, and crucially demonstrated that the PONE is still present when
259
participants are asked to give false behavioural answers in a malingering task, and
even when they are asked not to respond at all, but is absent if participants
randomly preload an answer without making a recognition decision. A further
important finding from Experiment 7 was that despite a slight impairment of
performance levels, the PONE is still present when attention is divided both at
learning and/or during the recognition phase. Experiments 8 and 9 set out to
explore whether the PONE and concurrent ERPs responded in a graded manner to
an ERP memory strength manipulation, and showed that the PONE is accompanied
by parietal ERP old/new effects at 300-500ms and 500-700ms, showing enhanced
positivity for old items presented once at learning, compared to new items, and that
neither the PONE nor the ERP old/new effects are enhanced by repetition of items
during learning.
Across all the experiments reported in this thesis that employed a standard
recognition memory procedure, maximum pupil-size was larger when participants
looked at old items compared to when they looked at novel items. Taken as a
whole, these findings support the theory that the PONE reflects mnemonic
processes recruited when participants make a recognition decision. It is important to
note that even when an item is new, mnemonic processes are activated – in part
due to prior exposure to the common English words used in the experiments, but
also because participants are actively seeking to reject novel items, for example
searching their memory to ensure that the item was not presented. This may
account for the fact that pupil-size to new items, in conditions where participants
make a recognition decision, was still often larger than pupil size to old or new items
in conditions where no recognition decision was required (such as reading, although
in the short duration reading condition of Experiment 2 pupil size was larger than for
260
the short duration recognition condition). In our results we have evidence for both
effort effects and mnemonic effects.
There is an extensive literature documenting the effects of “cognitive effort” on pupil
size (see Chapter 1, section 1.2.2.1) with some authors proposing that the PONE is
nothing more than the result of the greater cognitive effort required to correctly
identify old compared to new stimuli. Võ et al. (2008) suggest that recollection
requires the retrieval of qualitative contextual information, including the experience
of an old item during the study phase, which is more cognitively demanding than the
correct rejection of a new item, which does not. Their theory would predict that the
PONE should be smaller for deeply encoded items than shallowly encoded items
because less effort is required for recollection, however this was not what was found
when tested in Experiment 1. The central argument of this thesis, therefore, is that
the PONE is the result of conscious recollective processes that accompany the
recognition decision, and that items that are better remembered, or have a “stronger”
memory, are associated with a larger pupil-size (in line with Otero et al., 2011;
Papesh et al., 2011).
Although the experiments reported here did not directly measure participants’
introspective remember-know judgements, Experiments 3 and 4 used artificial
grammar, which Scott and Dienes (2008) propose elicits decisions based on
familiarity in the absence of recollection. In these experiments, no PONE was found
in response to familiar versus unfamiliar grammatical strings. This finding suggests
that within a dual-process model of recognition memory, the PONE reflects primarily
recollective processes. Whilst others (e.g., Otero et al., 2011) have found a larger
pupil size in response to old items rated as “known”, compared to new items, at
trend levels, it is difficult to exclude the possibility of recollective experience
261
contaminating familiarity judgements, with weaker memory strength leading to both
the intermediate pupil size and the failure to say that the item is remembered.
Wixted and Mickes (2010) propose that the R-K paradigm merely distinguishes
strong and weak memories. It is therefore possible to interpret the findings of thesis
in line with continuous strength models of recognition memory, such as the signal
detection unequal variance model (Wixted, 2007a), STREAK (Rotello et al., 2004),
and single-trace dual-process models (e.g., Greve, Donaldson, & van Rossum,
2010). These models assume that like familiarity, recollection also lies on a
continuum, and that rather than recognition decisions being based on either
recollection OR familiarity, both sources of memory information are summed into a
unitary combined memory strength that is then compared with a criterion value to
make a recognition decision (Wixted, & Stretch, 2004; Wixted, 2007a).
Experiment 8 attempted to provide further evidence for a memory strength signal in
the pupil by replicating a graded memory strength ERP study, which demonstrated
greater positivity for strongly encoded items relative to items with weaker encoding
(Finnigan et al., 2002), and concurrently measuring a graded pupil response. For
weak items an enhanced positivity relative to new items was present at parietal
electrodes and maximal at P7 in both time windows. Whilst this is consistent with
the left parietal old/new effect seen in the literature around 500-700ms, the 300500ms old/new effect for word recognition is typically seen maximally at frontocentral sites rather than at parietal sites (e.g., Curran, 2000). Other recent studies
have also shown an early old/new effect with a posterior scalp distribution, however
these studies have been concerned with face recognition rather than word
recognition (MacKenzie, & Donaldson, 2007; Paller et al., 2000; Yovel, & Paller,
2004). Finnigan et al. (2002) found a posterior old/new effect between 300-500ms
for word stimuli, but did not discuss possible origins, they merely referred to it as a
262
“posterior N400 strength effect” (p. 2300) and interpreted their results within a singleprocess model of recognition memory. Although a posterior distribution for the early
old/new effect might suggest that the same cognitive processes underlie familiarity
and recollection, as MacKenzie and Donaldson (2007) point out, even when the
topography of ERPs overlap in this way, it is not possible to determine whether or
not common neural generators are implicated.
Although the expected old/new effects were not present for strong items, early and
late frontal old/new effects of reversed polarity were demonstrated. In addition, very
early hemisphere differences (80-150ms) between weak and strong items persisted
throughout the trial, reflecting a difference in magnitude and/or location of neural
activity. The different pattern of old/new effects for strong items may be due to
interference effects, as proposed by Norman et al. (2008), leading to a reduction in
recollection. An alternative explanation is that one or both old/new effects has been
attenuated by repetition suppression, where less neural activity occurs for repeated
stimuli (e.g., strong) compared to new stimuli (see Schacter, & Buckner, 1998) or
stimuli presented once at learning (e.g., weak) (Jiang et al., 2000). Repetition
suppression is well documented in the ERP and other psychophysiological
literatures (e.g., Guo et al., 2007; Suzuki et al., 2011), and has recently been
reported in the pupil literature (Van Rijn et al., submitted).
Although the graded memory manipulation didn’t produce the predicted pattern in
either the ERP data or pupil-size data, the two psychophysiological measures did
respond in qualitatively the same manner, producing old/new differences for weak
items but not strong items at parietal electrodes. The parallel occurrence of an
old/new effect for weak items, and a possible repetition suppression effect for strong
items, in both the ERP and pupil-size data, raises the possibility that the two
263
measures index the same underlying neurocognitive processes occurring during
recognition memory.
Experiments 1 and 5 included measures of participant confidence, which is a
potential covariant to pupil size, through the positive emotion associated with a
correct response (Kahneman, 1973; Muldner et al., 2009). As discussed in Chapter
1, section 1.2.3.1 and Chapter 3, section 3.2.3, another link between confidence and
pupil-size is memory strength. If, as argued earlier, pupil-size during explicit
recognition decisions reflects the strength of the underlying memory trace, it is to be
expected that trials that have a “strong” memory lead both to a high level of
confidence and a larger PDR. In other words, if confidence ratings are taken as a
reflection of participants’ subjective experience of the strength of this aggregate
signal, and the pupil-size increase reflects the cognitive processes that drive this
signal, then pupil-size increases should be greater for high compared to low
confidence judgments, as was indeed the case. Interestingly, when analyses were
restricted to highly confident answers only, pupil size was still significantly larger for
old items than new items, suggesting that confidence and pupil-size may both be
downstream effects of memory strength but remain, to some extent at least,
independent.
Beatty and Wagoner (1975; 1976) measured confidence and pupil-size in a target
detection task and found that largest pupil-sizes were evoked by highly confident hits
and the smallest pupil-sizes occurred for highly confident rejections, with low
confidence hits and misses in between, suggesting that confidence and pupil-size
are not always tightly coupled. In addition, the Quiet condition of Experiment 6
demonstrated a PONE without a behavioural response, suggesting that the PONE
cannot simply be the result of confidence in a correct response. An experiment that
264
could potentially explore this idea further is the Deese-Roediger-McDermott (DRM)
paradigm (Roediger, & McDermott, 1995), which elevates the rate of false alarms
(new items identified as old), and therefore confidence in items which are new but
are confidently identified as old. Using this paradigm, Otero et al. (2011) found that
average PDR for false alarms was significantly smaller than for correctly identified
old items, therefore a replication which also measured confidence might show that
false-alarms have a significantly smaller average pupil-size but an equivalently high
confidence rating to old items.
There are a number of methodological issues within these that, with the benefit of
hindsight, could have been improved. As acknowledged in Chapter 3 section 3.2.3,
the larger pupil size to new items than old items in the Implicit condition of
Experiment 1 may have been a novelty or orienting response, visible in the absence
of the PONE (Laeng et al., 2007; Lynn, 1966; Pavlov, 1927; Sokolov, 1963).
However, as this finding was not replicated in the short duration reading condition of
Experiment 2, it is difficult to draw conclusions about the underlying cause without
further replication. Given the effect sizes seen in the Implicit condition of Experiment
1, Experiment 2 had sufficient participants (n = 28) to give a statistical power to
detect similar sized effects of 80%.
The pupil results of Experiments 8 and 9 are slightly contradictory in that for
Experiment 8 the PONE was only present for weak items, but in Experiment 9 the
PONE was only present for strong items. In addition, in Experiment 8, pupil size for
weak items was also significantly larger than for strong items, however this was
echoed by the ERP data which also showed an old/new effect for weak items only
and more positive-going ERPs for weak than strong items. PDRs in Experiment 9
were smaller than expected, and given the effect sizes seen in Experiment 8,
265
Experiment 9 needed a larger number of participants (n = 56) to increase the
statistical power to 80% to detect the same size effect.
It should be acknowledged that different outcomes can arise from different analyses
of the same pupil-size data. One limitation of the analyses included in this thesis is
that they focussed on the main effect of item-type, comparing PDR to old items with
PDR to new items. Alternative interpretations and conclusions may have been
drawn had data been analysed and presented in terms of interaction effects, i.e.
differences in PDR between old and new items in each condition. This approach is
equivalent to the construction and comparison of difference waves to analyse ERP
data, such as in section 6.2.2.7 of Chapter 6, and would be carried out by
subtracting PDR to new items from PDR to old items and subjecting the PONE
subtraction data to the ANOVA. For example, the results of Experiment 1,
summarised in Figure 3-4, highlight the difference in overall PDR to correctly
identified new items, which in the Implicit condition is larger than in the Explicit
condition. This has the effect of detracting from the old/new effect, and provides
support for the effort account of the PONE (Võ et al., 2008). Had the data been
presented as old/new differences, then it would more clearly show support for the
strength account of the PONE (Otero et al., 2011) because the PONE in the Explicit
condition is larger than in the Implicit condition.
In addition, had the analysis of the Levels Of Processing (LOP) manipulation,
illustrated by Figure 3-5, been presented as old/new differences, it would have been
clear that the PONE was larger for deeply encoded items than shallowly encoded
items in the Explicit condition – better supporting the argument that deeply encoded
items are associated with a stronger memory signal. In Experiment 2, the results
analysed and presented in Figure 3-10 appear to emphasise an effort-related main
266
effect of short vs. long presentation duration, with short duration conditions requiring
more cognitive effort. Had data been presented as old/new differences in each
condition, then it would have been clear that the difference was largest for the two
recognition conditions, with the strongest memory signal for the long duration
recognition condition. Similarly stronger support for the memory strength account of
the PONE can be made by alternative presentations of the data for the remaining
seven experiments.
In light of this issue, it is important to reflect on the balance of evidence, and how the
data presented here speaks to the effort vs. memory strength debate outlined in the
introductory chapter. As discussed above, evidence for the effort account of the
PONE can be found in the results of Experiment 2, which show a larger overall pupilsize to items presented in the more difficult short duration conditions than in the
easier long duration conditions. In addition, the pattern of results in Experiment 8
could reflect effort-related changes in pupil-size. Weak items were the hardest to
correctly identify, as demonstrated by the behavioural data (57.2% vs. 81.2% for
new and 72.9% for strong items), and mirroring the changes in pupil-size which were
largest for weak items and not significantly different for new and strong items.
However, behavioural performance in Experiment 9 was very similar to that of
Experiment 8 (54.1% weak, 80.4% new and 73.9% strong items), yet the pupil-size
data showed a different result – largest for strong items and not significantly different
for new and weak items. In addition, in Experiment 8 pupil responses exhibited the
same pattern as the late ERP old/new effect at parietal electrodes, associated with
recognition memory processes in an extensive literature (e.g., Curran, 2004; Düzel
et al., 1997; Friedman, 2004; Rugg et al., 1998b; Smith, 1993; Trott et al., 1999). If
behavioural performance is to be taken to indicate task difficulty, on some level at
least, then additional evidence in support of a memory strength account of the
267
PONE comes from Experiments 3 and 4. Here, performance in the Implicit condition
was much lower than that in the standard explicit recognition task (only 61.9%
compared to 74.6%), yet pupil-size increase was larger in the Explicit condition. A
similar line of argument can be used in interpreting the findings of Experiment 7, in
which the divided attention manipulation reduced behavioural performance but did
not affect pupil-size between conditions.
In Experiment 5 pupil-size is largest in the Standard condition, despite the fact that in
the Malingering condition participants are carrying out a more complex task,
involving suppressing a number of correct responses and implementing a covert
malingering strategy. Deception itself has been associated with increased cognitive
effort (e.g., Dionisio et al., 2011) and/or increased anxiety (e.g., Berrien, &
Huntington, 1943), therefore it seems unlikely that the effort explanation of the
PONE can account for the largest pupil-sizes occurring in the Standard condition.
Further evidence comes from the Incomplete Effort condition of Experiment 6, where
participants were required to not pay attention to stimuli but only during the learning
phase – at recognition they were to try their best to correctly identify the items.
Therefore, the smaller PDR that occurred in the Incomplete Effort condition
compared to the Standard condition, is unlikely to be the result of reduced effort and
instead fits with the memory strength account of the PONE. The experimental
manipulation that was introduced specifically to provide evidence for one account
over the other was varying the LOP in Experiment 1. The effort account of the
PONE predicts that pupil-size for deeply encoded items should be smaller than for
shallowly encoded items because they are easier to remember. In contrast, the
results showed that pupil-size for deeply encoded items was larger than for shallowly
encoded items, supporting the memory strength account of the PONE.
268
PDR increased in response to all tasks, in all experiments across the thesis. As
discussed in Chapter 1, a vast number of factors influence pupil-size, many of which
cannot be controlled or easily isolated in an experimental context – including
increases associated with: making an overt response, participant confidence or
anxiety, variations in attention, and cognitive effort related to either task difficulty or
how hard the participant is trying on a particular trial. It is not the intention to argue
that all task-evoked increase in pupil-size is due to memory strength; rather that the
balance of evidence suggests that in these experiments the increase in participants’
pupil-size when they correctly identify old items compared to new items is
predominantly due to a memory-strength signal (e.g., Otero et al., 2011). With hindsight the subtractive difference analysis would have been a stronger way of
presenting this central argument.
7.1.2.
Future Directions
Future research could extend the present findings in several ways. Research into
the cognitive correlates of pupil-size is currently enjoying something of a
renaissance, but even now there have been very few studies that have explored the
role of mnemonic processes. The topic is still in its infancy and several important
issues remain unresolved, with some key methodological issues yet to be refined.
The eye-tracker outputs other data, so other aspects which might be interesting to
analyse include latency to peak pupil-size, in order to understand more about the
timecourse of the underlying neurophysiological processes and further characterise
the PONE. The data collected during the course of this thesis could be analysed as
waveforms in a manner similar to the analysis of ERPs. This would allow inspection
of grand-average waveforms and selection of smaller time windows within the
2000ms trials for analysis if there appears to be a consistent pattern of response.
269
Then a mean ‘amplitude’ measure from a specific time window could be extracted.
For example, Kuipers and Thierry (2011) selected an epoch within each trial from
-100 to 850ms, where 0ms represents stimulus presentation, baseline-corrected to
the 100ms prestimulus section, and applied a 10Hz low-pass filter offline prior to
measurement of mean pupil amplitude. Other researchers report filtering pupil-size
data using a 10Hz low-pass filter (e.g., Hupé et al., 2009) as cognition-induced
changes in pupil size are of a lower frequency, but some sources of noise (e.g., from
estimation of pupil-size) produce artifactual changes at a higher frequency. The
application of artifact-detection might also give cleaner data with which to work (e.g.,
Hupé et al., 2009).
One way in which the findings of this thesis could be taken forward would be to
explore the effects of other types of memory strength manipulations on the PONE.
If, as argued above, the PONE essentially reflects a strength of memory signal, then
other manipulations of memory strength should also result in a “graded” pupil-size.
For example the use of established mnemonic strategies, such as
visualisation/imagery for half the stimuli at learning, should enhance memory,
facilitating a much stronger recollection at recognition, and therefore a bigger pupil.
Additionally it would be interesting to investigate the effects on the PONE of
manipulations which have been shown to dissociate behavioural measures of
familiarity and recollection in the ERP literature. For example, Yovel and Paller’s
(2004) unknown faces with spoken occupations design could be adapted for use
with words by presenting related information with each item at learning and asking
participants to recall the additional information at recognition. This would allow
separation of items remembered with different degrees of clarity for analysis.
270
Given the promising preliminary finding that a similar pattern emerges in the PONE
as in ERP correlates of recognition memory, further concurrent recordings could be
made to investigate whether the PONE responses to ‘strength of memory’
manipulations were matched by variations in the magnitude of the ERP old/new
effects at 300-500ms and 500-700ms. Once the time-course of corresponding
effects had been clarified using ERPs, a technique with better spatial resolution
could be applied, such as EEG source localisation or fMRI, making links with what is
already known about the neural substrate of recognition memory and of the pupillary
control system.
In recent years, researchers have shown that it is possible to record from single
neurons in conscious human brains (see for example Quian Quiroga’s work on the
Jennifer Anniston neuron, 2008; 2010; Rutishauser et al., 2008). If the opportunity
arose to test recognition memory in patients undergoing awake brain surgery whilst
also measuring pupil-size and directly recording the activity of Locus Coeruleus (LC)
neurons, this would provide valuable information confirming the hypothesised link
between LC activity and pupil-size in humans during cognitive tasks. It would also
provide information on whether the LC increases phasic firing when participants are
correctly recognising old items compared to new items, or whether this increased
dilation arises from a different part of the brain.
It is noted that no studies so far deal directly with pupil-size during visual word
recognition in amnesic patients, reflecting a gap in the literature. In terms of
pupillometry research, this suggests that future work needs to pay more attention to
what happens to the apparently conscious recognition-related PONE in a patient
who can’t make an explicit recognition decision. These ideas were introduced in
Chapter 5, and it was suggested that the PONE might be suitable as a means for
271
detecting malingered memory-impairment. However, the only way to truly determine
the PONE’s utility would be by establishing its parameters within a genuine patient
population in a standard recognition memory task.
As well as exploring the PONE in amnesic patients, memory could be manipulated in
healthy participants using Transcranial Magnetic Stimulation (TMS), which has been
shown to impair some types of memory (e.g., Prime, Vesia, & Crawford, 2008), and
enhance others (Kirschen, Davis-Ratner, Jerde, Schraedley-Desmond, & Desmond,
2006). Machizawa, Kalla, Walsh and Otten (2010) found that TMS applied to the left
or right inferior frontal gyrus of the prefrontal cortex during the learning phase
affected performance on a recognition memory test fifteen minutes later. Turriziani
et al. (2008) found that familiarity was impaired in a test of recognition after TMS
was used to stimulate the right and left dorsolateral prefrontal cortex (DLPFC) prior
to encoding, and that recollection was impaired after stimulation of the right DLPFC.
Clearly it is possible to influence behavioural performance measures on a
recognition test and it would be fascinating to see whether this temporary
impairment also extends to the PONE.
272
8. Bibliography
Aboyoun, D.C., & Dabbs, J.N. (1998). The Hess pupil dilation findings: Sex or
novelty? Social Behavior and Personality, 26(4), 415-419.
doi:10.2224/sbp.1998.26.4.415
Achilles, E.M. (1920). Experimental studies in recall and recognition. New York: G.E.
Stechert & Co. Retrieved 14 October 2011 from
http://www.archive.org/details/experimentalstud00achi
Aggleton, J.P., & Brown, M.W. (1999). Episodic memory, amnesia, and the
hippocampal-anterior thalamic axis. Behavioral and Brain Sciences, 22(3),
425–489.
Ahern, S.K., & Beatty, J. (1979). Pupillary responses during information processing
vary with scholastic aptitude test scores. Science, 205(4412), 1289-1292.
doi:10.1126/science.472746
Ahern, S.K., & Beatty, J. (1981). Physiological evidence that demand for processing
capacity varies with intelligence. In M.P. Friedman, J.P. Das, & N. O'Connor
(Eds.), Intelligence and learning (pp. 121-128). New York: Plenum.
Allan, K., & Rugg, M.D. (1997). An event-related potential study of explicit memory
on tests of cued recall and recognition. Neuropsychologia, 35(4), 387-397.
doi:10.1016/S0028-3932(96)00094-2
Allan, K.,Wilding, E.L., & Rugg, M.D. (1998). Electrophysiological evidence for
dissociable processes contributing to recollection. Acta Psychologica, 98(23), 231–252. doi:10.1016/S0001-6918(97)00044-9
Allen, J., & Mertens, R. (2009). Limitations to the detection of deception: True and
false recollections are poorly distinguished using an event-related potential
procedure. Social Neuroscience, 4(6), 473-490.
doi:10.1080/17470910802109939
Allen, R., & Reber, A.S. (1980). Very long-term memory for tacit knowledge.
Cognition, 8(2), 175-185. doi:10.1016/0010-0277(80)90011-6
Ally, B.A., Waring, J.D., Beth, E.H., McKeever, J.D., Milberg, W.P., & Budson, A.E.
(2008). Aging memory for pictures: Using high-density event-related
potentials to understand the effect of aging on the picture superiority effect.
Neuropsychologia, 46(2), 679-689.
doi:10.1016/j.neuropsychologia.2007.09.011
273
American Clinical Neurophysiology Society (2006). Guideline Five: Guidelines for
standard electrode position nomenclature. Journal of Clinical
Neurophysiology, 23(2), 107-110.
American Electroencephalographic Society (1991). Guidelines for standard
electrode position nomenclature. Journal of Clinical Neurophysiology, 8, 200202.
American Electroencephalographic Society (1994). Guideline thirteen: Guidelines for
standard electrode position nomenclature. Journal of Clinical
Neurophysiology, 11, 111-113.
Andreassi, J.L. (2000). Psychophysiology: Human behavior and physiological
response. Mahwah, NJ: Lawrence Erlbaum Associates.
Ardolf, B.R., Denney, R.L., & Houston, C.M. (2007). Base rates of negative response
bias and malingered neurocognitive dysfunction among criminal defendants
referred for neuropsychological evaluation. The Clinical Neuropsychologist,
21(6), 899-916. doi:10.1080/13825580600966391
Aston-Jones, G., Chiang, C., & Alexinsky, T. (1991). Discharge of noradrenergic
locus coeruleus neurons in behaving rats and monkeys suggests a role in
vigilance. Progress in Brain Research, 88, 501-20.
Aston-Jones, G., & Cohen, J.D. (2005). An integrative theory of locus coeruleusnorepinephrine function: Adaptive gain and optimal performance. Annual
Review of Neuroscience, 28, 403-450. doi:10.1146/annurev.neuro.
28.061604.135709
Aston-Jones, G., Foote, S.L., & Bloom, F.E. (1984). Anatomy and physiology of
locus coeruleus neurons: Functional implications. In M. Ziegler, & C.R. Lake
(Eds.), Norepinephrine: Frontiers of clinical neuroscience (pp. 92-116).
Baltimore: Williams and Wilkins.
Aston-Jones, G., Rajkowski, J., & Kubiak, P. (1997). Conditioned responses of
monkey locus coeruleus neurons anticipate acquisition of discriminative
behavior in a vigilance task. Neuroscience, 80(3), 697-715.
doi:10.1016/S0306-4522(97)00060-2
Aston-Jones, G., Rajkowski, J., Kubiak, P., & Alexinsky, T. (1994). Locus coeruleus
neurons in monkey are selectively activated by attended cues in a vigilance
task. The Journal of Neuroscience, 14(7), 4467-4460.
274
Aston-Jones, G., Rajkowski, J., Lu, W., Zhu, Y., Cohen, J.D., & Morecraft, R.J.
(2002). Prominent projections from the orbital prefrontal cortex to the locus
coeruleus in monkey. Society for Neuroscience Abstracts, 28, 86-89.
Atkinson, R.C., & Juola, J.F. (1973). Factors influencing speed and accuracy of word
recognition. In S. Kornblum (Ed.) Attention and performance IV (pp. 583-612).
New York: Academic Press.
Atkinson, R.C., & Juola, J.F. (1974). Search and decision processes in recognition
memory. In D.H. Krantz, R.C. Atkinson, R.D. Luce, & P. Suppes (Eds.),
Contemporary developments in mathematical psychology. I:
Learning,memory and thinking (pp. 243-293). San Francisco, CA: Freeman.
Axelrod, B.N., Barlow, A., & Paradee, C. (2010). Evaluation of the WMS-III rarely
missed index in a naïve clinical sample. Clinical Neuropsychologist, 24(1), 95102. doi:10.1080/13854040902960083
Backs, R.W., & Walrath, L.C. (1992). Eye movement and pupillary response indices
of mental workload during visual search of symbolic displays. Applied
Ergonomics, 23(4), 243-254. doi:10.1016/0003-6870(92)90152-L
Baddeley, A. (2000). The episodic buffer: A new component of working memory?
Trends in Cognitive Sciences, 4(11), 417-423. doi:10.1016/S13646613(00)01538-2
Baddeley, A.D., & Hitch, G.J.L. (1974). Working memory. In G.A. Bower (Ed.), The
psychology of learning and motivation: Advances in research and theory (Vol.
8, pp. 47–89), New York: Academic Press.
Bailey, B.P., Busbey, C.W., & Iqbal, S.T. (2007). TAPRAV: An interactive analysis
tool for exploring workload aligned to models of task execution. Interacting
with Computers, 19(3), 314-329. doi:10.1016/j.intcom.2007.01.004
Bailey, B.P., & Iqbal, S.T. (2008). Understanding changes in mental workload during
execution of goal-directed tasks and its application for interruption
management. ACM Transactions on Computer-Human Interaction, 14(4), 128. doi:10.1145/1314683.1314689
Bailey, D.L., Townsend, D.W., Valk, P.E., & Maisey, M.N. (2005). Positron emission
tomography: Basic sciences. London: Springer.
Balota, D.A., & Neely, J.H. (1980). Text expectancy and word-frequency effects in
recall and recognition. Journal of Experimental Psychology: Human Learning
and Memory, 6(5), 576–587. doi:10.1037/0278-7393.6.5.576
275
Banks, W.P. (1970). Signal detection theory and human memory. Psychological
Bulletin, 74(2), 81-99. doi:10.1037/h0029531
Barbur, J.L., Wolf, J., & Lennie, P. (1998). Visual processing levels revealed by
response latencies to changes in different visual attributes. Proceedings of
the Royal Society of London, Series B, 265(1412), 2321–2325.
doi:10.1098/rspb.1998.0578
Barlow, J.D. (1969). Pupillary size as a index of preference in political candidates.
Perceptual & Motor Skills, 28, 927-932. doi:10.2466/ pms.1969.28.2.587
Barlow, J.D. (1970). Pupillary size as an index of preference. Perceptual & Motor
Skills, 31(1), 331-336. doi:10.2466/pms.1970.31.1.331
Barreto, A., Zhai, J., Rishe, N., & Gao, Y. (2007). Significance of pupil diameter
measurements for the assessment of affective state in computer users. In E.
Elleithy (Ed.), Advances and innovations in systems, computing sciences and
software engineering (pp. 59-64). Dordrecht, The Netherlands: Springer.
doi:10.1007/978-1-4020-6264-3_11
Bartholomew, A. (1483). De proprietatibus rerum, Liber V. Nuremberg, Germany:
Anton Koberger. (Original work published around 1240). Retrieved 25 August
2011 from http://catalog.hathitrust.org/Record/001906553
Bauer, L.M., Olheiser, E.L., Altarriba, J., & Landi, N. (2009). Word type effects in
false recall: Concrete, abstract, and emotion word critical lures. American
Journal of Psychology, 122(4), 469-481.
Baxendale, S. (2004). Memories aren’t made of this: Amnesia at the movies. British
Medical Journal, 329(7480), 18-25. doi:10.1136/bmj.329.7480.1480
Beatty, J. (1982a). Phasic not tonic pupillary responses vary with aduitory vigilance
performance. Psychophysiology, 19(2), 167-172. doi:10.1111/j.14698986.1982.tb02540.x
Beatty, J. (1982b). Task-evoked pupillary responses, processing load, and the
structure of processing resources. Psychological Bulletin, 91, 276–292.
doi:10.1037/0033-2909.91.2.276
Beatty, J. (1986). The pupillary system. In G.H. Coles, E. Donchin, & S.W. Porges
(Eds.), Psychophysiology: Systems, process, and applications (pp. 43-50).
New York: Guilford Press.
Beatty, J. (1988). Pupillometric signs of selective attention in man. In G.C. Galbraith,
M.L. Kietzman, & E. Donchin (Eds.), Neurophysiology and psychophysiology:
Experimental and clinical applications (pp. 138-143). Hillsdale, NJ: Erlbaum.
276
Beatty, J., & Kahneman, D. (1966). Pupillary changes in two memory tasks.
Psychonomic Science, 5, 371-372.
Beatty, J., & Lucero-Wagoner, B. (1975). Pupillary measurement of sensory and
decision processes in a signal-detection task. Paper presented at The
Meeting of the Psychonomic Society. Denver: Psychonomic Society.
Beatty, J., & Lucero-Wagoner, B. (2000). The pupillary system. In J.T. Cacioppo,
L.G. Tassinary, & G.G. Berntson (Eds.), Handbook of psychophysiology (2nd
ed., pp. 142-162). USA: Cambridge University Press.
Beatty, J., & Wagoner, B.L. (1976). Pupillometric assessment of signal analysis and
decision processes in acoustic signal detection. Unpublished manuscript,
University of California at Los Angeles, CA, USA.
Beatty, J., & Wagoner, B.L. (1978). Pupillometric signs of brain activation vary with
level of cognitive processing. Science, 199(4334), 1216-1218.
doi:10.1126/science.628837
Becker, W., Iwase, K., Jürgens, R., & Kornhuber, H.H. (1976). Brain potentials
preceding slow and rapid hand movements. In W.C. McCallum, & J.R. Knott
(Eds.), The responsive brain (pp. 99-102). Bristol: Wright.
Bender, W.R.G. (1933). The effect of pain and emotional stimuli and alcohol upon
pupillary reflex activity. Psychological Monographs, 44(2), 1-32.
Ben-Shakhar, G., & Elaad, E. (2003). The validity of psychophysiological detection
of information with the Guilty Knowledge Test: A meta-analytic review.
Journal of Applied Psychology, 88(1), 131-151. doi:10.1037/00219010.88.1.131
Bentin, S., & McCarthy, G. (1994). The effects of immediate stimulus repetition on
reaction time and event-related potentials in tasks of different complexity.
Journal of Experimental Psychology: Learning,Memory and Cognition, 20(1),
130–149.
Bentin, S., & Peled, B. (1990). The contribution of task-related factors to ERP
repetition effects at short and long lags. Memory & Cognition, 18(4), 359-366.
doi:10.3758/BF03197125
Bernard, C. (1852). Sur les effets de la section de la portion céphalique du grand
sympathique. Chronicles of the Royal Society of Biologists, Paris, 4, 168-170.
Bernbach, H.A. (1967). Decision processes in memory. Psychological Review,
74(6), 462-480. doi:10.1037/h0025132
277
Bernhardt, P.C., Dabbs, J.M., & Riad, J.K. (1996). Pupillometry system for use in
social psychology. Behavior Research Methods, 28(1), 61–66.
doi:10.3758/BF03203637
Bernick, N., Altman, F., & Mintz, D.L. (1972). Pupil responses of addicts in treatment
to drug culture argot: II. Responses during verbalization of visually presented
words. Psychonomic Science, 28(2), 81-82.
Bernick, N., Kling, A., & Borowitz, G. (1971). Physiologic differentiation of sexual
arousal and anxiety. Psychosomatic Medicine, 33(4), 341-352.
Berridge, C.W., & Waterhouse, B.D. (2003). The locus coeruleus- noradrenergic
system: Modulation of behavioral state and state-dependent cognitive
processes. Brain Research Review, 42, 33–84. doi:10.1016/S01650173(03)00143-7
Berrien, F.K., & Huntington, G.H. (1943). An exploratory study of pupillary responses
during deception. Journal of Experimental Psychology, 32(5), 443-449.
doi:10.1037/h0063488
Berry, D., & Dienes, Z.P. (1993). Implicit learning: Theoretical and empirical issues.
Hove: Lawrence Erlbaum Associates Ltd.
Birren, J.E., Casperson, R.C., & Botwinick, J. (1950). Age changes in pupil-size.
Journal of Gerentology, 5(3), 216-221. doi:10.1093/geronj/5.3.216
Bitsios, P., Philpott, A., Langley, R.W., Bradshaw, C.M., & Szabadi, E. (1999).
Comparison of the effects of diazepam on the fear-potentiated startle reflex
and the fear-inhibited light reflex in man. Journal of Psychopharmacology,
13(3), 226-234. doi:10.1177/026988119901300303
Bitsios, P., Prettyman, R., & Szabadi, E. (1996a). Changes in autonomic function
with age: A study of pupillary kinetics in healthy young and old people. Age &
Ageing, 25(6), 432-438. doi:10.1093/ageing/25.6.432
Bitsios, P., Szabadi, E., & Bradshaw, C.M. (1996b). The inhibition of the pupillary
light reflex by the threat of an electric shock: A potential laboratory model of
human anxiety. Journal of Psychopharmacology, 10(4), 279-287.
doi:10.1177/026988119601000404
Bitsios, P., Szabadi, E., & Bradshaw, C.M. (2002). Relationship of the 'fear-inhibited
light reflex' to the level of state/trait anxiety in healthy subjects. International
Journal of Psychophysiology, 43(2), 177-184. doi:10.1016/S01678760(01)00173-8
278
Botvinick, M., Braver, T., Barch, D., Carter, C., & Cohen, J.D. (2001). Conflict
monitoring and cognitive control. Psychological Review, 108(3), 624-652.
doi:10.1037//0033-295X.I08.3.624
Bouma, H. (1962). Size of the static pupil as a function of wavelength and luminosity
of the light incident on the human eye. Nature, 193, 690–691.
doi:10.1038/193690a0
Bouma, H., & Baghuis, L.C.J. (1971). Hippus of the pupil: Periods of slow
oscillations of unknown origin. Vision Research, 11(11), 1345-1351.
doi:10.1016/0042-6989(71)90016-2
Bourget, D., & Whitehurst, L. (2007). Amnesia and crime. Journal of the American
Academy of Psychiatry and the Law, 35(4), 469–480.
Bowles, B., Crupi, C., Mirsatarri, S.M., Pigott, S.E., Parrent, A.G., Pruessner, J.C., …
Kohler, S. (2007). Impaired familiarity with preserved recollection after
anterior temporal lobe resection that spares the hippocampus. Proceedings
for the National Academy of Sciences, 104(41), 16382-16387.
doi:10.1073/pnas.0705273104
Bowles, B., Crupi, C., Pigott, S., Parrent, A., Wiebe, S., Janzen, L., & Kohler, S.
(2010). Double dissociation of selective recollection and familiarity
impairments following two surgical treatments for temporal lobe epilepsy.
Neuropsychologia, 48(9): 2640–2647.
doi:10.1016/j.neuropsychologia.2010.05.010
Bradley, J.C., Bentley, K.C., Mughal, A.I., Bodhireddy, H., Young, R.S., & Brown,
S.M. (2010). The effect of gender and iris color on the dark-adapted pupil
diameter. Journal of Ocular Pharmacolology and Therapeutics, 26(4), 335-40.
doi:10.1089/jop.2010.0061
Bradley, M.T., & Janisse, M.P. (1979). Pupil-size and lie detection: The effect of
certainty on detection. Psychology, 16(4), 33-39.
Bradley, M.T., & Janisse, M.P. (1981). Accuracy demonstrations, threat, and the
detection of deception: Cardiovascular, electrodermal, and pupillary
measures. Psychophysiology, 18(3), 307-315. doi:10.1111/j.14698986.1981.tb03040.x
Bradshaw, J. (1967). Pupil-size as a measure of arousal during information
processing. Nature, 216(5114), 515-516. doi:10.1038/216515a0
Bradshaw, J.L. (1968a). Pupil-size and problem solving. Quarterly Journal of
Experimental Psychology, 20(2), 116-122. doi:10.1080/14640746808400139
279
Bradshaw, J.L. (1968b). Pupillary changes and reaction time with varied stimulus
uncertainty. Psychonomic Science, 13(2), 69-70.
Bradshaw, J.L. (1969). Background light intensity and the pupillary response in a
reaction time task. Psychonomic Science, 14(6), 271-272.
Bradshaw, J.L. (1970). Pupil-size and drug state in a reaction time task.
Psychonomic Science, 18(2), 112-113.
Brainerd, C.J., & Reyna, V.F. (2002). Recollection rejection: How children edit their
false memories. Developmental Psychology, 38(1), 156-172.
doi:10.1037/0012-1649.38.1.156
Brainerd, C.J., Wright, R., Reyna, V.F., & Mojardin, A.H. (2001). Conjoint recognition
and phantom recollection. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 27(2), 307-327. doi:10.1037/0278-7393.27.2.307
Brontë, C. (1946). Jane Eyre. M. Lamberton Becker, (Ed.). Cleveland, NY: The
World Publishing Co. (Original work published 1847).
Brown, A.A., & Bodner, G.E. (2011). Re-examining dissociations between
remembering and knowing: Binary judgments vs. independent ratings.
Journal of Memory and Language, 65(2), 98-108.
doi:10.1016/j.jml.2011.04.003
Brown, G.G., Kinderman, S.S., Siegle, G.J., Granholm, E., Wong, E.C., & Buxton,
R.B. (1999). Brain activation and pupil response during covert performance of
the Stroop colour word task. Journal of the International Neuropsychological
Society, 5(4), 308–319.
Browndyke, J.N., Paskavitz, J., Sweet, L.H., Cohen, R.A., Tucker, K.A., WelshBohmer, K.A., … Schmechel, D.E. (2008). Neuroanatomical correlates of
malingered memory-impairment: Event-related fMRI of deception on a
recognition memory task. Brain Injury, 22(6), 481-489.
doi:10.1080/02699050802084894
Brozinsky, C.J., Yonelinas, A.P., Kroll, N.E.A., & Ranganath, C. (2005). Lagsensitive repetition suppression effects in the anterior parahippocampal
gyrus. Hippocampus, 15(5), 557-561. doi:10.1002/hipo.20087
Brunia, C.H.M., & Vingerhoets, A.J.J.M. (1980). CNV and EMG preceding a plantar
flexion of the foot. Biological Psychology, 11(3-4), 181-191.
Buckner, R.L., Goodman, J., Burock, M., Rotte, M., Koutstaal, W., Schacter, D., …
Dale, A.M. (1998). Functional-anatomic correlates of object priming in
280
humans revealed by rapid presentation event-related fMRI. Neuron, 20(2),
285–296. doi:10.1016/S0896-6273(00)80456-0
Budge, J., & Waller, A. (1851). Action de la portion cervicale du nerf grand
sympathique et d'une portion de la moelle epinière sur la dilatation de la
pupille. Gazette Médicale de Paris: Journal de Médecine et des Sciences
Accessoires, 3(6), 648-649.
Bull, R., & Shead, G. (1979). Pupil-dilation, sex of stimulus, and age and sex of
observer. Perceptual & Motor Skills, 49(1), 27-30.
doi:10.2466/pms.1979.49.1.27
Cacioppo, J.T., & Tassinary, L.G. (1990). Inferring psychological significance from
physiological signals. American Psychologist, 45(1), 16-28.
Cacioppo, J.T., Tassinary, L.G., & Berntson, G.G. (2000). Psychophysiological
science. In J.T. Cacioppo, L.G. Tassinary, & G.G. Berntson (Eds.), Handbook
of psychophysiology (2nd ed., pp. 3-23). New York: Cambridge University
Press.
Cain, B. (2007). A review of the mental workload literature (NATO Report No. RTOTR-HFM-121-Part-II). Toronto, Canada: Defence Research and
Development.
Carver, R.P. (1971). Pupil dilation and its relationship to information processing
during reading and listening. Journal of Applied Psychology, 55(2), 126-134.
doi:10.1037/h0030664
Casey, B.J., & de Haan, M. (2002). Introduction: New methods in developmental
science. Developmental Science, 5(3), 265-267. doi:10.1111 /14677687.00365
Chan, C. (1992). Implicit cognitive processes: Theoretical issues and applications in
computer systems design. (Unpublished doctoral dissertation). University of
Oxford, Oxford.
Chapman, C.R., Oka, S., Bradshaw, D.H., Jacobson, R.C., & Donaldson, G.W.
(1999). Phasic pupil dilation response to noxious stimulation in normal
volunteers: Relationship to brain evoked potentials and pain report.
Psychophysiology, 36(1), 44-52. doi:10.1017/S0048577299970373
Chatham, C.H, Frank, M.J., & Munakata, Y. (2009). Pupillometric and behavioral
markers of a developmental shift in the temporal dynamics of cognitive
control. Proceedings of the National Academy of Sciences, 106(14), 55295533, 2009. doi:10.1073/pnas.0810002106
281
Chelazzi, L., Miller, E.K., Duncan, J., & Desimone, R. (1993). A neural basis for
visual search in inferior temporal cortex. Nature, 363, 345–347.
doi:10.1038/363345a0
Cherry, K.E., & Stadler, M.A. (1995). Implicit learning of a non-verbal sequence in
younger and older adults. Psychology & Aging, 10(3), 379-394.
doi:10.1037/0882-7974.10.3.379
Cima, M., Nijman, H., Merckelbach, H., Kremer, K., & Hollnack, S. (2004). Claims of
crime-related amnesia in forensic patients. International Journal of Law &
Psychiatry, 27(3), 215–221. doi:10.1016/j.ijlp.2004.03.007
Clark, J.W. (1885). Civilisation and eyesight. Nature, 31(802), 433–434.
doi:10.1038/031433d0
Clark, K.B., Krahl, S.E., Smith, D.S., & Jensen, R.A. (1995). Post-training unilateral
vagal stimulation enhances retnetion performance in the rat. Neurobiology of
Learning & Memory, 63(3), 213-216. doi:10.1006/nlme.1995.1024
Clark, K.B., Naritoku, D.K., Smith, D.C., Browning, R.A., & Jensen, R.A. (1999).
Enhanced recognition memory following vaus nerve stimulation in human
subjects. Nature Neuroscience, 2, 94-98. doi:10.1038/4600
Clark, S.E., & Gronlund, S.D. (1996). Global matching models of recognition
memory: How the models match the data. Psychonomic Bulletin & Review,
3(1), 37–60. doi:10.3758/BF03210740
Clark, W.R. (1975). A comparison of pupillary response, heart rate, and GSR during
deception. Paper presented at The Annual Meeting of the Midwestern
Psychological Association. Chicago: Midwestern Psychological Association.
Clark, W.R., & Ertas, M.A. (1975). A comparison of pupillary reactions to visual and
auditory stimuli in a test of preferences for presidential candidates. Journal
Supplement Abstract Service Catalog of Selected Documents in Psychology,
12, 20-25.
Clark, W.R., & Johnson, D.A. (1970). Effects of instructional set on pupillary
responses during a short-term memory task. Journal of Experimental
Psychology, 85(2), 315-317. doi:10.1037/h0029543
Clayton, E.C., Rajkowski, J., Cohen, J.D., & Aston-Jones, G. (2004). Phasic
activation of monkey locus ceruleus neurons in a forced-choice task. Journal
of Neuroscience, 24(44), 9914–9920. doi:10.1523/JNEUROSCI.244604.2004
282
Cleeremans, A., & Dienes, Z. (2008). Computational models of implicit learning. In
R. Sun (Ed.), The Cambridge handbook of computational modeling (pp. 396421). Cambridge, UK: Cambridge University Press.
Cohen, J.D., Aston-Jones, G., & Gilzenrat, M.S. (2004). A systems-level perspective
on attention and cognitive control: Guided activation, adaptive gating, conflict
monitoring, and exploitation versus exploration. In M.I. Posner (Ed.),
Cognitive neuroscience of attention (pp. 71–90). New York: Guilford Press.
Cohen, J.D., Botvinick, M., & Carter, C.S. (2000). Anterior cingulate and prefrontal
cortex: Who’s in control? Nature Neuroscience, 3, 421-423.
doi:10.1038/74783
Cohn, M., Moscovitch, M., & Davidson, P.S.R. (2010). Double dissociation between
familiarity and recollection in Parkinson's disease as a function of encoding
tasks. Neuropsychologia, 48(14), 4142-4147.
doi:10.1016/j.neuropsychologia.2010.10.013
Coles, M.G.H., & Strayer, D.L. (1985). The psychophysiology of the cardiac cycle
time effect. In J.F. Orlebeke, G. Mulder, & L.J.P. van Doornen (Eds.),
Psychophysiology of cardiovascular control: Models, methods and data (pp.
517-534). New York: Plenum Press.
Colman, F.D., & Paivio, A. (1969). Pupillary response and galvanic skin response
during an imagery task. Psychonomic Science, 16(6), 296-297.
Coltheart, M. (1981). The MRC Psycholinguistic Database. Quarterly Journal of
Experimental Psychology, 33A, 497-505. Accessed at
http://www.psy.uwa.edu.au/mrcdatabase/uwa_mrc.htm
Conati, C., & Merten, C. (2007). Eye-tracking for user modeling in exploratory
learning environments: An empirical evaluation. Knowledge-Based Systems,
20(6), 557-574. doi:10.1016/j.knosys.2007.04.010
Conway, C.A., Jones, B.C., DeBruine, L.M., Little, A.C., & Sahraie, A. (2008).
Transient pupil constrictions to faces are sensitive to orientation and species.
Journal of Vision, 8(3), 17, 1-11. doi:10.1167/8.3.17
Cottingham, M.E., & Boone, K.B. (2010). Non-credible language deficits following
mild traumatic brain injury. The Clinical Neuropsychologist, 24(6), 1006-1025.
doi:10.1080/13854046.2010.481636
Craik, F.I.M., Govoni, R., Naveh-Benjamin, M., & Anderson, N.D. (1996). The effects
of divided attention on encoding and retrieval processes in human memory.
Journal of Experimental Psychology: General, 125(2), 159-180.
283
Craik, F.I.M., & Lockhart, R.S. (1972). Levels of processing: A framework for
memory research. Journal of Verbal learning & Verbal Behaviour, 11(6), 671684. doi:10.1016/S0022-5371(72)80001-X
Craik, F.I.M., & McDowd, J.M. (1987). Age differences in recall and recognition.
Journal of Experimental Psychology: Learning, Memory, & Cognition, 13(3),
474–479.
Criss, A.H. (2006). The consequences of differentiation in episodic memory:
Similarity and the strength based mirror effect. Journal of Memory &
Language: Special Issue on Computational Models of Memory, 55(4), 461478. doi:10.1016/j.jml.2006.08.003
Curran, T. (1999). The electrophysiology of incidental and intentional retrieval: ERP
old/new effects in lexical decision and recognition memory.
Neuropsychologia, 37(7), 771-785. doi:10.1016/S0028-3932(98)00133-X
Curran, T. (2000). Brain potentials of recollection and familiarity. Memory &
Cognition, 28(6), 923–938.
Curran, T. (2004). Effects of attention and confidence on the hypothesized ERP
correlates of recollection and familiarity. Neuropsychologia, 42(8), 1088–
1106. doi:10.1016/j.neuropsychologia.2003.12.011
Curran, T., & Cleary, A.M. (2003). Using ERPs to dissociate recollection from
familiarity in picture recognition. Cognitive Brain Research, 15(2), 191–205.
doi:10.1016/S0926-6410(02)00192-1
Curran, T., DeBuse, C., Woroch, B., & Hirshman, E. (2006). Combined
pharmacological and electrophysiological dissociation of familiarity and
recollection. The Journal of Neuroscience, 26(7), 1979–1985.
doi:10.1523/JNEUROSCI.5370-05.2006
Curran, T., & Doyle, J. (2011). Picture superiority doubly dissociates the ERP
correlates of recollection and familiarity. Journal of Cognitive Neuroscience,
23(5), 1247-1262. doi:10.1162/jocn.2010.21464
Curran, T., Schacter, D.L., Johnson, M.K., & Spinks, R. (2001). Brain potentials
reflect behavioral differences in true and false recognition. Journal of
Cognitive Neuroscience, 13, 201–216. doi:10.1162/089892901564261
Curran, T., Tepe, K.L., & Piatt, C. (2006). ERP explorations of dual processes in
recognition memory. In H.D. Zimmer, A. Mecklinger, & U. Lindenberger
(Eds.), Binding in human memory: A neurocognitive approach (pp. 467-492).
Oxford: Oxford University Press.
284
Curran, T., Tucker, D.M., Kutas, M., & Posner, M.I. (1993). Topography of the N400:
Brain electrical activity reflecting semantic expectation.
Electroencephalography & Clinical Neurophysiology, 88(3), 188–209.
Dabbs, J.M., & Milun, R. (1999). Pupil dilation when viewing strangers: Can
testosterone moderate prejudice? Social Behavior & Personality, 27(3), 297301.
Dahlberg, J. (2010). Eye tracking with eye glasses. (Unpublished masters
dissertation). Umea University, Sweden.
Dain, S.J., Cassimaty, V.T., & Psarakis, D.T. (2004). Differences in FM100-Hue test
performance related to iris colour may be due to pupil-size as well as
presumed amounts of macular pigmentation. Clinical & Experimental
Optometry, 87(4-5), 322-325. doi:10.1111/j.1444-0938.2004.tb05061.x
Daly, R.M. (1966). Pupillary size and its relationship to the problem solving process.
(Unpublished doctoral dissertation). Loyola University, Chicago.
Darwin, C.R. (1872). The expression of the emotions in man and animals. London:
John Murray.
Daugman, J. (2003). The importance of being random: statistical principles of iris
recognition. Pattern Recognition, 36(2), 279–291. doi:10.1016/S00313203(02)00030-4
Davies, L.N., Wolffsohn, J.S., & Gilmartin, B. (2005). Cognition, ocular
accommodation, and cardiovascular function in emmetropes and late-onset
myopes. Investigative Ophthalmology & Visual Science, 46(5), 1791-1796.
doi:10.1167/iovs.04-0986
de Balzac, H. (2000). Le curé du village [The village rector]. (K. Prescott Wormeley,
Trans.). Cirencester: The Echo Library. (Original work published 1841).
de Greef, T., Lafeber, H., van Oostendorp, H., & Lindenberg, J. (2009). Eye
movement as indicators of mental workload to trigger adaptive automation.
Foundations of augmented cognition: Neuroergonomics and operational
neuroscience (pp. 219-228): Springer-Verlag.
Deese, J. (1960). Frequency of usage and number of words in free recall: The role
of association. Psychological Review , 7, 337-344. doi: 10.2466/PR0.7.6.337344
Dehais, F., Causse, M., & Pastor, J. (2008). Embedded eye tracker in a real aircraft:
new perspectives on pilot/aircraft interaction monitoring. In Proceedings from
285
The 3rd International Conference on Research in Air Transportation. Fairfax,
USA: Federal Aviation Administration.
DeLaunay, J. (1949). A note on the photo-pupil reflex. Journal of the Optical Society
of America, 39(5), 364-367. doi:10.1364/JOSA.39.000364
Denney, R.L. (2007). Assessment of malingering in criminal forensic
neuropsychological settings. In K.B. Boone (Ed.), Assessment of feigned
cognitive impairment: A neuropsychological perspective (pp. 428–452). New
York: Guilford.
Dennis, S., & Humphreys, M.S. (2001). A context noise model of episodic word
recognition. Psychological Review, 108(2), 452–478. doi:10.1037/0033295X.108.2.452
DePaulo, B.M., Lindsay, J.J., Malone, B.E., Muhlenbruck, L., Charlton, K., & Cooper,
H. (2003). Cues to deception. Psychological Bulletin, 129(1), 74-118.
doi:10.1037/0033-2909.129.1.74
Desimone, R. (1996). Neural mechanisms for visual memory and their role in
attention. Proceedings of the National Academy of Sciences of the United
States of America, 93(24), 13494–13499.
Diana, R., Reder, L.M., Arndt, J., & Park, H. (2006). Models of recognition: A review
of arguments in favor of a dual process account. Psychonomic Bulletin &
Review, 13(1), 1-21.
Dien, J. (1998). Issues in the application of the average reference: Review, critiques,
and recommendations. Behavior Research Methods, 30(1), 34–43.
doi:10.3758/BF03209414
Dienes, Z., & Altmann, G. (1997). Transfer of implicit knowledge across domains:
How implicit and how abstract? In D. Berry (Ed.), How implicit is implicit
learning? (pp 107-123). Oxford: Oxford University Press.
Dienes, Z., Altmann, G., Kwan, L., & Goode, A. (1995). Unconscious knowledge of
artificial grammars is applied strategically. Journal of Experimental
Psychology: Learning, Memory, & Cognition, 21(5), 1322-1338.
doi:10.1037/0278-7393.21.5.1322
Dienes, Z., & Berry, D. (1997). Implicit learning: Below the subjective threshold.
Psychonomic Bulletin & Review, 4(1), 3-23.
Dienes, Z., Broadbent, D.E., & Berry, D.C. (1991). Implicit and explicit knowledge
bases in artificial grammar learning. Journal of Experimental Psychology:
Learning, Memory, & Cognition, 17(5), 875-887.
286
Dienes, Z., & Fahey, R. (1998). The role of implicit memory in controlling a dynamic
system. Quarterly Journal of Experimental Psychology A, 51(3), 593-614.
Dienes, Z., & Scott, R. (2005). Measuring unconscious knowledge: Distinguishing
structural knowledge and judgment knowledge. Psychological Research,
69(5-6), 338-351. doi:10.1007/s00426-004-0208-3
Diller, D.E., Nobel, P.A., & Shiffrin, R.M. (2001). An ARC–REM model for accuracy
and response time in recognition and recall. Journal of Experimental
Psychology: Learning, Memory, & Cognition, 27(2), 414-435.
doi:10.1037/0278-7393.27.2.414
Dionisio, D.P., Granholm, E., Hillix, W.A., & Perrine, W.F. (2001). Differentiation of
deception using pupillary responses as an index of cognitive processing.
Psychophysiology, 38(2), 205-211.
Donaldson, D.I., Allan, K., & Wilding, E.L. (2002). Fractionating episodic memory
retrieval using event-related potentials. In A. Parker, E.L. Wilding, & T.
Bussey (Eds.), The cognitive neuroscience of memory: Encoding and
retrieval (pp.39-58). Hove: Psychology Press.
Donaldson, D.I., Petersen, S.E., Ollinger, J.M., & Buckner, R.L. (2001). Dissociating
state and item components of recognition memory using fMRI. NeuroImage,
13, 129-142. doi:10.1006/nimg.2000.0664
Donaldson, W. (1996). The role of decision processes in remembering and knowing.
Memory & Cognition, 24(4), 523-533.
Dooley, R.P., & Lehr, D.J. (1967). Critique of a pupillary response experiment.
Perceptual & Motor Skills, 25(2), 603-604. doi:10.2466/pms.1967.25.2.603
Dosher, B.A. (1984). Degree of learning and retrieval speed: Study time and multiple
exposures. Journal of Experimental Psychology: Learning, Memory &
Cognition, 10(4), 541–574. doi:10.1037/0278-7393.10.4.541
Doughty, M.J. (2001). Ocular pharmacology and therapeutics: A primary care guide.
Oxford: Butterworth-Heinemann.
Duarte, A., Ranganath, C., Winward, L., Hayward, D., & Knight, R.T. (2004).
Dissociable neural correlates for familiarity and recollection during the
encoding and retrieval of pictures. Cognitive Brain Research, 18(3), 255-272.
doi:10.1016/j.cogbrainres.2003.10.010
Duchowski, A.T. (2007). Eye tracking methodology: Theory and practice (2nd ed.).
London: Springer.
287
Dunn, J.A. (2004). Remember-Know: A matter of confidence. Psychological Review,
111(2), 524-542.
Düzel, E., Vargha-Khadem, F., Heinze, H.J., & Mishkin, M. (2001). Brain activity
evidence for recognition without recollection after early hippocampal damage.
Proceedings of the National Academy of Sciences of the United States of
America, 98(14), 8101-8106. doi:10.1073/pnas.131205798
Düzel, E., Yonelinas, A.P., Mangun, G.R., Heinze, H.J., & Tulving, E. (1997). Eventrelated brain potential correlates of two states of conscious awareness in
memory. Proceedings of the National Academy of Sciences of the United
States of America, 94(11), 5973-5978.
Einhauser, W., Stout, J., Koch, C., & Carter, O. (2008). Pupil dilation reflects
perceptual selection and predicts subsequent stability in perceptual rivalry.
Proceedings of the National Academy of Science of the United States of
America, 105(5), 1704-1709. doi:10.1073/pnas.0707727105
Ekman, I., Poikola, A., Mäkäräinen, M., Takala, T., & Hämäläinen, P. (2008a).
Voluntary pupil-size change as control in eyes only interaction. In
Proceedings from The 2008 Symposium on Eye Tracking Research &
Applications (pp. 115-188). New York, NY: ACM Press.
doi:10.1145/1344471.1344501
Ekman, I.M., Poikola, A.W., & Mäkäräinen, M.K. (2008b). Invisible eni: Using gaze
and pupil-size to control a game. In Proceedings from The CHI '08 Extended
Abstracts on Human Factors in Computing Systems (pp. 3135-3140). New
York, NY: ACM Press. doi:10.1145/1358628.1358820
Elaad, E. (2009). Lie-detection biases among male police interrogators, prisoners,
and laypersons. Psychological Reports, 105(3), 1047-1056.
doi:10.2466/pr0.105.F.1047-1056
Eldridge, L.L., Knowlton, B.J., Furmanski, C.S., Bookheimer, S.Y., & Engel, S.A.
(2000). Remembering episodes: A selective role for the hippocampus during
retrieval. Nature Neuroscience, 3(11), 1149–1152. doi:10.1038/80671
Elshtain, E.L., & Schaefer, T.Jr. (1968). Effects of storage load and word frequency
on pupillary responses during short-term memory. Psychonomic Science,
12(4), 143-144.
Engineer, N.D., Riley, J.R., Seale, J.D., Vrana, W.A., Shetake, J.A., Sudanagunta,
S.P., . . . Kilgard, M.P. (2011). Reversing pathological neural activity using
targeted plasticity. Nature, 470, 101-106. doi:10.1038/nature09656
288
Engelhard, I.M., Merckelbach, H., & van den Hout, M.A. (2003). The Guilty
Knowledge Test and the modified Stroop task in detection of deception: An
exploratory study. Psychological Reports, 92(2), 683-691.
Erichsen, J.T., Hodos, W., & Evinger, C. (2000). The pupillary light reflex,
accommodation and convergence: Comparative considerations. In O.
Franzen, H. Richter, & L. Stark (Eds.), Accommodation and vergence
mechanisms in the visual system (pp. 31-42). Basel: Birkhauser.
Evans, R.W., & Jacobson, D.M. (2003). Transient anisocoria in a migraineur.
Headache, 43(4), 416-418. doi:10.1046/j.1526-4610.2003.03081.x
Fabiani, M., & Donchin, E. (1995). Encoding processes and memory organization: A
model of the Von Restorff effect. Journal of Experimental Psychology:
Learning, Memory, & Cognition, 21(1), 224- 240. doi:10.1037/02787393.21.1.224
Fabiani, M., Gratton, G., & Coles, M.G.H. (2000). Event-related brain potentials. In
J.T. Cacioppo, L.G. Tassinary, & G.G. Berntson (Eds.), Handbook of
psychophysiology (pp. 53-84). New York: Cambridge University Press.
Fanciullacci, M., Galli, P., Pietrini, U., & Sicuteri, F. (1977). Adrenergic
supersensitivity of the pupil in idiopathic headache. Experientia, 33(8), 1082–
1083. doi:10.1007/BF01945983
Finnigan, S., Humphreys, M.S., Dennis, S., & Geffen, G. (2002). ERP ‘old/new’
effects: Memory strength and decisional factor(s). Neuropsychologia, 40(13),
2288-2304. doi:10.1016/S0028-3932(02)00113-6
Fogarty, C., & Stern, J.A. (1989). Eye movements and blinks: Their relationship to
higher cognitive processes. International Journal of Psychophysiology, 8(1),
35-42. doi:10.1016/0167-8760(89)90017-2
Forbes, T.R. (1977). Why is it called 'beautiful lady'? A note on belladonna. Bulletin
of the New York Academy of Medicine, 53(4), 403-406.
Ford, E.B. (2006). Lie detection: Historical, neuropsychiatric and legal dimensions.
International Journal of Law & Psychiatry, 29(3), 159-177.
doi:10.1016/j.ijlp.2005.07.001
Foster, M. (1891). A textbook of physiology volume IV (5th ed.). Cambridge: C.J.
Clay, M.A. & Sons.
Foster, J.K., & Jelicic, M. (Eds.) (1999). Memory: Systems, process or function.
Oxford: Oxford University Press.
289
Francis, R.D., & Kelly, M.R. (1969). An investigation of the relationship between
word stimuli and optical pupil-size. Australian Journal of Psychology, 21(2),
117-125. doi:10.1080/00049536908257774
Frank, M.G., & Ekman, P. (1997). The ability to detect deceit generalizes across
different types of high-stake lies. Journal of Personality & Social Psychology,
72(6), 1429-1439. doi:10.1037/0022-3514.72.6.1429
Franzen, M.D., & Iverson, G.L. (2000). Detecting negative response bias and
diagnosing malingering: The dissimulation exam. In J. Snyder, & P.J.
Nussbaum (Eds.), Clinical neuropsychology: A pocket handbook for
assessment (2nd ed.) (pp. 102-119). Washington, DC: American
Psychological Association. doi:10.1037/11299-006
Frensch, P.A. (1998). One concept, multiple meanings: On how to define the
concept of implicit learning. In M.A. Stadler, & P.A. Frensch (Eds.), Handbook
of implicit learning (pp. 47–104). Thousand Oaks, CA: Sage.
Frensch, P.A., & Miner, C.S. (1994). Effects of presentation rate and individual
differences in short-term memory capacity on an indirect measure of serial
learning. Memory & Cognition, 22(1), 95-110. doi:10.3758/BF03202765
Friedman, D. (2004). ERP studies of recognition memory: Differential effects of
familiarity, recollection, episodic priming. Cognitive Sciences, 1(1), 81-121.
Friedman, D., Hakerem, G., Sutton, S., & Fleiss, J.L. (1973). Effect of stimulus
uncertainty on the pupillary dilation response and the vertex evoked potential.
Electroencephalography & Clinical Neurophysiology, 34(5), 475-484.
doi:10.1016/0013-4694(73)90065-5
Friedman, D., Hamberger, M., Stern, Y., & Marder, K. (1992). Event-related
potentials (ERPs) during repetition priming in Alzheimer's patients and young
and older controls. Journal of Clinical & Experimental Neuropsychology,
14(4), 448-462. doi:10.1080/01688639208402837
Friedman, D., & Johnson Jr., R. (2000). Event-related potential (ERP) studies of
memory encoding and retrieval: A selective review. Microscopy Research &
Technique, 51(1), 6-28. doi:10.1002/1097-0029(20001001)51:1<6::AIDJEMT2>3.0.CO;2-R
Fukuda, K., Stern, J.A., Brown, T.B., & Russo, M.B. (2005). Cognition, blinks, eyemovements, and pupillary movements during performance of a running
memory task. Aviation, Space, & Environmental Medicine, 76(1S), C75-C85.
290
Gabay, S., Pertzov, Y., & Henik, A. (2011). Orienting of attention, pupil-size, and the
norepinephrine system. Attention, Perception & Psychophysics, 73(1), 123129. doi:10.3758/s13414-010-0015-4
Galappathie, N., & Vakili, K. (2009). Detecting memory malingering: A case study.
British Journal of Forensic Practice, 11(1), 20-24.
doi:10.1108/14636646200900005
Gambill, H.D., Ogle, K.N., & Kearns, T.P. (1967). Mydriatic effect of four drugs
determined with pupillograph. Archives of Ophthalmolology, 77(6), 740-746.
Gang, K. (1945). Psychosomatic factors in the control of pupillary movements.
Journal of Clinical Psychopathology Psychotherapy, 6, 461-472.
Gao, Y., Barreto, A., & Adjouadi, M. (2010). Affective assessment of a computer
user through the processing of the pupil diameter signal. In T. Sobh, & K.
Elleithy (Eds.), Innovations in computing sciences and software engineering
(pp. 189-194). New York: Springer. doi:10.1007/978-90-481-9112-3_32
Gardiner, J.M. (1988). Functional aspects of recollective experience. Memory &
Cognition, 16(4), 309–313. doi:10.3758/BF03197041
Gardiner, J.M., & Java, R.I. (1993). Recognizing and remembering. In S.E.
Gathercole, M.A. Conway, E.E. Morris, & A.E Collins (Eds.), Theories of
memory (pp. 163-188). Hove, England: Erlbaum.
Gardiner, J.M., & Parkin, A.J. (1990). Attention and recollective experience in
recognition memory. Memory & Cognition, 18(6), 579-583.
doi:10.3758/BF03197100
Gardiner, J.M., & Richardson-Klavehn, A. (2000). Remembering and knowing. In E.
Tulving, & F.I.M. Craik (Eds.), Oxford handbook of memory (pp. 229-244).
New York: Oxford University Press.
Gardner, R.M., Beltramo, J.S., & Krinsky, R. (1975). Pupillary changes during
encoding, storage, and retrieval of information. Perceptual & Motor Skills,
41(3), 951-955. doi:10.2466/pms.1975.41.3.951
Gardner, R.M., Mo, S.S., & Borrego, R. (1974a). Inhibition of pupillary orienting
reflex by novelty in conjunction with recognition memory. Bulletin of the
Psychonomic Society, 3(3-B), 237-238.
Gardner, R.M., Mo, S.S., & Krinsky, R. (1974b). Inhibition of pupillary orienting reflex
by heteromodal novelty. Bulletin of the Psychonomic Society, 4(5B), 510-512.
291
Gardner, R.M., Philp, P., & Radacy, S. (1978). Pupillary changes during recall in
children. Journal of Experimental Child Psychology, 25(1), 168-172.
doi:10.1016/0022-0965(78)90046-2
Garrett, J.C., Harrison, D.W., & Kelly, P.L. (1989). Pupillometric assessment of
arousal to sexual stimuli: Novelty effects or preference. Archives of Sexual
Behavior, 18(3), 191-201. doi:10.1007/BF01543194
Geacintov, T., & Peavler, W.S. (1974). Pupillography in industrial fatigue
assessment. Journal of Applied Psychology, 59(2), 213-216.
doi:10.1037/h0036529
Ghanem, T., & Early, S.V. (2006). Vagal nerve stimulator implantation: An
otolaryngologist's perspective. Otolaryngology Head Neck Surgery, 135(1),
46-51. doi:10.1016/j.otohns.2006.02.037
Ghodse, A.H., Greaves, J.L., & Lynch, D. (1999). Evaluation of the opioid addiction
test in an out-patient drug dependency unit. British Journal of Psychiatry, 175,
158-162. doi: 10.1192/bjp.175.2.158
Gillund, G., & Shiffrin, R.M. (1984). A retrieval model for both recognition and recall.
Psychological Review, 91(1), 1-67. doi:10.1037/0033-295X.91.1.1
Gilzentrat, M.S. (2006). The role of the locus coeruleus in cognitive control:
Pupillometric measures and computational mechanisms. (Unpublished
doctoral dissertation). Princeton University, New Jersey, USA.
Gilzenrat, M.S., Cohen, J.D., Rajkowski, J., & Aston-Jones, G. (2003). Pupil
dynamics predict changes in task engagement mediated by locus coeruleus.
Society for Neuroscience Abstracts, 515.19.
Gilzenrat, M.S., Nieuwenhuis, S., Jepma, M., & Cohen, J.D. (2010). Pupil diameter
tracks changes in control state predicted by the adaptive gain theory of locus
coeruleus function. Cognitive, Affective, & Behavioral Neuroscience, 10(2),
252-269. doi:10.3758/CABN.10.2.252
Glanzer, M., & Adams, J.K. (1985). The mirror effect in recognition memory. Memory
& Cognition, 13(1), 8-20.
Goldmann, R.E., Sullivan, A.L., Droller, D.B., Rugg, M.D., Curran, T., Holcomb, P.J.,
… Budson, A.E. (2003). Late frontal brain potentials distinguish true and false
recognition. Neuroreport, 14(13), 1717-1720.
doi:10.1097/01.wnr.0000087908.78892.23
Goldwater, B. (1972). Psychological significance of pupillary movements.
Psychological Bulletin, 77(5), 340-355. doi:10.1037/h0032456
292
Goodrich, G.L. (1974). Actual and apparent pupillary diameters by sex, age,
refraction, iris coloration, and fundus coloration. (Unpublished doctoral
dissertation). Washington State University, USA.
Gorman, A.M. (1961). Recognition memory for nouns as a function of abstractness
and frequency. Journal of Experimental Psychology, 61(1), 23-29.
doi:10.1037/h0040561
Gouvier, W.D., Prestholdt, P.H., & Warner, M.S. (1988). A survey of common
misconceptions about head injury and recovery. Archives of Clinical
Neuropsychology, 3(4), 331-343. doi:10.1093/arclin/3.4.331
Graf, P., Mandler, G., & Haden, P.E. (1982). Simulating amnesic symptoms in
normal subjects. Science, 218(4578), 1243-1244.
doi:10.1126/science.7146909
Granholm, E., Asarnow, R.F., Sarkin, A.J., & Dykes, K.L. (1996). Pupillary
responses index cognitive resource limitations. Psychophysiology, 33(4), 457461. doi:10.1111/j.1469-8986.1996.tb01071.x
Granholm, E., Morris, S., Galasko, D., Shults, C., Rogers, E., & Vukov, B. (2003).
Tropicamide effects on pupil-size and pupillary light reflexes in Alzheimer's
and Parkinson's disease. International Journal of Psychophysiology, 47(2),
95-115. doi:10.1016/S0167-8760(02)00122-8
Granholm, E., Morris, S.K., Sarkin, A.J., Asarnow, R.F., & Jeste, D.V. (1997).
Pupillary responses index overload of working memory resources in
schizophrenia. Journal of Abnormal Psychology, 106(3), 458-467.
doi:10.1037/0021-843X.106.3.458
Granholm, E., & Verney, S.P. (2004). Pupillary responses and attentional allocation
on the visual backward masking task in schizophrenia. International Journal
of Psychophysiology, 52, 37-52. doi:10.1016/j.ijpsycho.2003.12.004
Gratiolet, P., & Grandeau, L. (1865). De la physionomie et des mouvements
d’expression: Suivi d’une notice sur so vie et ses travaux et de la
nomenclature de ses ouvrages. Paris: J. Hetzel.
Green, P., & Astner, K. (1995). Manual for the Oral Word Memory Test. Durham,
NC: CogniSyst, Inc.
Greiffenstein, M.F., Baker, J.W., Gola, T., Donders, J., & Miller, L. (2002). The fake
bad scale in atypical and sever closed head injury litigants. Journal of Clinical
Psychology, 58(12), 1591–1600. doi:10.1002/jclp.10077
293
Gregg, V.H. (1976). Word frequency, recognition, and recall. In J. Brown (Ed.),
Recall and recognition (pp. 183-216). London: John Wiley & Sons.
Greve, A., Donaldson, D.I., & van Rossum, M.C.W. (2010). A single-trace dualprocess model of episodic memory: A novel computational account of
familiarity and recollection. Hippocampus, 20(2), 235-51.
doi:10.1002/hipo.20606
Grill-Spector, K., & Malach, R. (2001). fMRI-adaptation: A tool for studying the
functional properties of human cortical neurons. Acta Psychologica, 107(1-3),
293-321.
Gronlund, S.D., & Ratcliff, R. (1989). Time course of item and associative
information: Implications for global memory models. Journal of Experimental
Psychology: Learning, Memory & Cognition, 15(5), 846-858.
doi:10.1037/0278-7393.15.5.846
Groppe, D.M., Urbach, T.P., & Kutas, M. (2011). Mass univariate analysis of eventrelated brain potentials/fields I: A critical tutorial review. Psychophysiology,
48(12), 1711-1725. doi:10.1111/j.1469-8986.2011.01273.x
Groves, D.A., & Brown, V.J. (2005). Vagus nerve stimulation: A review of its
applications and potential mechanisms that mediate its clinical effects.
Neuroscience & Biobehavioral Reviews, 29(3), 493-500.
doi:10.1016/j.neubiorev.2005.01.004
Gruber, T., Tsivilis, D., Giabbiconi, C.M., & Muller, M.M. (2008). Induced
electroencephalogram oscillations during source memory: Familiarity is
reflected in the gamma band, recollection in the theta band. Journal of
Cognitive Neuroscience, 20(6), 1043-1053. doi:10.1162/jocn.2008.20068
Grunberger, J., Linzmayer, L., Majda, E.M., Reitner, A., & Walter, H. (1996).
Pupillary dilatation test and Fourier analysis of pupillary oscillations in patients
with multiple sclerosis. European Archives of Psychiatry & Clinical
Neuroscience, 246(4), 209-212. doi:10.1007/BF02188955
Grunberger, J., Linzmayer, L., Walter, H., Stohr, H., Saletu-Zyhlarz, G.M.,
Grunberger, M., & Lesch, O.M. (1998). Psychophysiological diagnostics in
alcohol dependency: Fourier analysis of pupillary oscillations and the receptor
test for determination of cholinergic deficiency. Alcohol & Alcoholism, 33(5),
541-548.
294
Guinan, J.F. (1967). An investigation of the relationship between pupil-size and
emotional words. (Doctoral dissertation). Dissertation Abstracts, 27(9-BO),
3286-3287.
Guo, C., Lawson, A.L., & Jiang, Y. (2007). Distinct neural mechanisms for repetition
effects of visual objects. Neuroscience, 149(4), 747-759.
doi:10.1016/j.neuroscience.2007.07.060
Habib, R., & LePage, M. (1999). Novelty assessment in the brain. In E. Tulving
(Ed.), Memory, consciousness, and the brain (pp. 265-277). Philadelphia, PA:
Psychology Press..
Hack Tuke, D. (1884). Illustrations of the Influence of the mind upon the body in
health and disease: Designed to elucidate the action of the imagination (2nd
ed.). Philadelphia, PA: Henry C Lea's Son & Co.
Hakerem, G. (1973). The effect of cognitive manipulation on pupillary diameter and
evoked vertex potentials. In E. Dodt, & K.E. Schrader (Eds.), The Normal and
The Perturbed Motion Pupil Symposium of the The German Society of
Ophthalmology. Munich: J.F. Bergmann Verlag.
Hakerem, G. (1974). Conceptual stimuli, pupillary dilation and evoked cortical
potentials. In M.P. Janisse (Ed.), Pupillary dynamics and behavior (pp. 135158). Plenum Press, New York.
Hakerem, G., & Sutton, S. (1966). Pupillary response at visual threshold. Nature,
212, 485-486. doi:10.1038/212485a0
Hannula, D.E., & Ranganath, C. (2009). The eyes have it: Hippocampal activity
predicts expression of memory in eye movements. Neuron, 63(5), 592-599.
doi:10.1016/j.neuron.2009.08.025
Hansen, D.W., & Pece, A.E. (2005). Eye tracking in the wild. Computer Vision and
Image Understanding, 98(1), 155-181. doi:10.1016/j.cviu.2004.07.013
Harle, D.E., Wolffsohn, J.S., & Evans, B.J.W. (2005). The pupillary light reflex in
migraine. Ophthalmic & Physiological Optics, 25(3), 240-245.
doi:10.1111/j.1475-1313.2005.00291.x
Harvey, W. (1649). Exercitatio duae anatomica de circulatione sanguinis ad
Joannem Riolanum filium. Cambridge: Ex officina Rogeri Danielis.
Hawkes, C.H., & Stow, B. (1981). Pupil-size and the pattern evoked visual response.
Journal of Neurology, Neurosurgery, and Psychiatry, 44(1), 90-91.
doi:10.1136/jnnp.44.1.90
295
Heaver, B., & Hutton, S.B. (2010). Keeping an eye on the truth: Pupil-size,
recognition memory and malingering. International Journal of
Psychophysiology, 77(3), 306-306. doi:10.1016/j.ijpsycho.2010.06.206
Heaver, B., & Hutton, S.B. (2011). Keeping an eye on the truth? Pupil-size changes
associated with recognition memory. Memory, 19(4), 398-405.
doi:10.1080/09658211.2011.575788
Heilbrun, K., Bennett, W.S., White, A.J., & Kelly, J. (1990). An MMPI-based
empirical model of malingering and deception. Behavioral Sciences & the
Law, 8(1), 45-53. doi:10.1002/bsl.2370080106
Heilveil, I. (1976). Deception and pupil-size. Journal of Clinical Psychology, 32(3),
675-676.
Heinrich, W. (1896). Die aufmerksamkeit und die funktion der sinnesorgane [The
attention and the function of the sensory organs]. Zeitschrift für Psychologie
und Physiologie der Sinnesorgane, 9, 342-388.
Helmholtz, H. (1853). Ueber einige gesetze der vertheilung elektrischer ströme in
körperlichen leitern mit anwendung auf die theirisch-elektrischen versuche
[Some laws about the distribution of electrical currents in volume conductors,
with application to animal electric experiments]. Annalen der Physik und
Chemie, 89, 211-233.
Henson, R.N., Hornberger, M., & Rugg, M.D. (2005). Further dissociating processes
in recognition memory using fMRI. Journal of Cognitive Neuroscience, 17(7),
1058-1073. doi:10.1162/0898929054475208
Henson, R.N.A., Rugg, M.D., Shallice, T., Josephs, O., & Dolan, R.J. (1999).
Recollection and familiarity in recognition memory: An event-related
functional magnetic resonance imaging study. Journal of Neuroscience,
19(10), 3962–3972.
Henson, R.N.A., Shallice, T., Gorno-Tempini, M.L., & Dolan, R.J. (2002). Face
repetition effects in implicit and explicit memory tests as measured by fMRI.
Cerebral Cortex, 12(2), 178-186. doi:10.1093/cercor/12.2.178
Hertel, P.T., & Parks, C. (2002). Emotional episodes facilitate word recall. Cognition
& Emotion, 16(5), 685–694. doi:10.1080/02699930143000491
Hess, E.H. (1965). Attitude and pupil-size. Scientific American, 212, 46–54.
Hess, E.H. (1972). Pupillometrics. In N.S. Greenfield, & R.A. Sternback (Eds.),
Handbook of psychophysiology (pp. 491-531). New York: Holt, Rinehart &
Winston.
296
Hess, E.H. (1975). The tell-tale eye. New York: Van Nostrand Reinhold.
Hess, E.H., Beaver, P.W., & Shrout, P.E. (1975). Brightness contrast effects in a
pupillometric experiment. Perception & Psychophysics, 18(2), 125-127.
doi:10.3758/BF03204099
Hess, E.H., & Polt, J.M. (1960). Pupil-size as related to interest value of visual
stimuli. Science, 132(3423), 349-350. doi:10.1126/science.132.3423.349.
Hess, E.H., & Polt, J.M. (1964). Pupil-size in relation to mental activity in simple
problem solving. Science, 143(3611), 1190-1192. doi:10.1126/science.
143.3611.1190
Hess, E.H., Seltzer, A.L., & Shlien, J.M. (1965). Pupil response of hetero- and
homosexual males to pictures of men and women: A pilot study. Journal of
Abnormal Psychology, 70(3), 165-168. doi:10.1037/h0021978
Hicks, R.A., Reaney, T., & Hill, L. (1967). Effects of pupil-size and facial angle on
preference for photographs of a young woman. Perceptual & Motor Skills,
24(2), 388-390. doi:10.2466/pms.1967.24.2.388
Hillyard, S.A., Hink, R.F., Schwent, V.L., & Picton, T.W. (1973). Electrical signs of
selective attention in the human brain. Science, 182(4108), 177-180.
doi:10.1126/science.182.4108.177
Hink, R.F., Van Voorhis, S.T., Hillyard, S.A., & Smith, T.S. (1977). The division of
attention and the human auditory evoked potential. Neuropsychologia, 15(45), 597-605. doi:10.1016/0028-3932(77)90065-3
Hintzman, D.L. (1988). Judgments of frequency and recognition memory in a
multiple-trace memory model. Psychological Review, 95(4), 528-551.
Hintzman, D.L., & Curran, T. (1994). Retrieval dynamics of recognition and
frequency judgments: Evidence for separate processes of familiarity and
recall. Journal of Memory & Language, 33, 1-18.
Hintzman, D.L., & Curran, T. (1995). When encoding fails: Instructions, feedback,
and registration without learning. Memory & Cognition, 23(2), 213–226.
doi:10.3758/BF03197223
Hintzman, D.L., & Curran, T. (1997). Comparing retrieval dynamics in recognition
memory and lexical decision. Journal of Experimental Psychology: General,
126(3), 228-247.
Hintzman, D.L., Curran, T., & Oppy, B. (1992). Effects of similarity and repetition on
memory: Registration without learning? Journal of Experimental Psychology:
297
Learning, Memory & Cognition, 18(4), 667-680. doi:10.1037/02787393.18.4.667
Holdstock, J.S., Mayes, A.R., Roberts, N., Cezayirli, E., Isaac, C.L., O’Reilly, R.C., &
Norman, K.A. (2002). Under what conditions is recognition spared relative to
recall after selective hippocampal damage in humans? Hippocampus, 12,
341-351. doi:10.1002/hipo.10011
Hou, R.H., Freeman, C., Langley, R.W., Szabadi, E., & Bradshaw, C.M. (2005).
Does modafinil activate the locus coeruleus in man? Comparison of modafinil
and clonidine on arousal and autonomic functions in human volunteers.
Psychopharmacology, 181(3), 537-549. doi:10.1007/s00213-005-0013-8
Howard, D.V., & Howard, J.H. (1989). Age differences in learning serial patterns:
Direct versus indirect measures. Psychology & Ageing, 4(3), 357-364. doi:
10.1037/0882-7974.4.3.357
Howard, D.V., & Howard, J.H. (1992). Adult age differences in the rate of learning
serial patterns: Evidence from direct and indirect tests. Psychology & Aging,
7(2), 232-241. doi:10.1037/0882-7974.7.2.232
Huber, D.E., Clark, T.F., Curran, T., & Winkielman, P. (2008). Effects of repetition
priming on recognition memory: Testing a perceptual fluency-disfluency
model. Journal of Experimental Psychology: Learning, Memory & Cognition,
34(6), 1305-1324. doi:10.1037/a0013370
Huettel, S.A., Song, A.W., & McCarthy, G. (2004). Functional magnetic resonance
imaging. Massachussets, USA: Sinauer Associates Inc
Humphreys, M.S., Bain, J.D., & Pike, R. (1989). Different ways to cue a coherent
memory system: A theory for episodic, semantic, and procedural tasks.
Psychological Review, 96(2), 208–233.
Hupé, J.M., Lamirel, C., & Lorenceau, J. (2009). Pupil dynamics during bistable
motion perception. Journal of Vision, 9(7), 10. doi:10.1167/9.7.10
Hutchinson, G.L. (2001). Disorders of simulation: Malingering, factitious disorders,
and compensation neurosis. Madison, Connecticut: Psychosocial Press.
Hyönä, J., Tommola, J., & Alaja, A.M. (1995). Pupil dilation as a measure of
processing load in simultaneous interpretation and other language tasks. The
Quarterly Journal of Experimental Psychology: A, 48(3), 598-612.
doi:10.1080/14640749508401407
Iijima, A., Haida, M., Ishikawa, N., Ueno, A., Minamitani, H., & Shinohara, Y. (2003).
Re-evaluation of tropicamide in the pupillary response test for Alzheimer's
298
disease. Neurobiology of Aging, 24(6), 789-796. doi:10.1016/S01974580(02)00235-X
Ingram, K.M., Mickes, L., & Wixted, J.T. (2011). Recollection can be weak and
familiarity can be strong. Journal of Experimental Psychology. Learning,
Memory, & Cognition, 38(2), 325-339. doi:10.1037/a0025483
Iqbal, S.T., Adamczyk, P.D., Zheng, X.S., & Bailey, B.P. (2005). Towards an index of
opportunity: Understanding changes in mental workload during task
execution. In Proceedings from The SIGCHI Conference on Human Factors
in Computing Systems (pp. 311-320). Portland, Oregon: ACM Press.
Iqbal, S.T., Zheng, X.S., & Bailey, B.P. (2004). Task-evoked pupillary response to
mental workload in human-computer interaction. In Proceedings from The
SIGCHI Conference on Human Factors in Computing Systems (pp. 14771480). New York: ACM Press. doi:10.1145/985921.986094
Isidore of Seville (2006). The etymologies of Isidore of Seville (S.A. Barney, W.J.
Lewis, J.A. Beach, & O. Berghof, Trans.). Cambridge: Cambridge University
Press. (Original work published 636 AD).
Jacoby, L.L., & Dallas, M. (1981). On the relationship between autobiographical
memory and perceptual learning. Journal of Experimental PsychologyGeneral, 110(3), 306-340. doi:10.1037/0096-3445.110.3.306
Jacoby, L.L., & Whitehouse, K. (1989). An illusion of memory: False recognition
influenced by unconscious perception. Journal of Experimental Psychology:
General, 118(2), 126-135. doi:10.1037/0096-3445.118.2.126
Jain, J., Siegle, G.J.,Gu. C., Moore, C.G., Ivanco, L.S., Studenski, S., … Steinhauer,
S.R. (2011). Pupillary unrest correlates with arousal symptoms and motor
signs in Parkinson disease. Movement Disorders, 26(7), 1344-1337.
doi:10.1002/mds.23628
Jainta, S., & Baccino, T. (2010). Analyzing the pupil response due to increased
cognitive demand: An independent component analysis study. International
Journal of Psychophysiology, 77, 1-7. doi:10.1016/j.ijpsycho.2010.03.008
Jainta, S., Hoormann, J., & Jaschinski, W. (2008). Ocular accommodation and
cognitive demand: An additional indicator besides pupil-size and
cardiovascular measures? Journal of Negative Results in BioMedicine, 7(6).
doi:10.1186/1477-5751-7-6
James, W. (1890). The principles of psychology, vol. 1. New York: Henry Holt & Co.
299
Janisse, M.P. (1973). Pupil-size and affect: A critical review of the literature since
1960. Canadian Psychologist, 14(4), 311-329. doi:10.1037/h0082230
Janisse, M.P. (Ed.) (1974). Pupillary dynamics and behavior. New York, NY: Plenum
Press.
Janisse, M.P. (1976). The relationship between pupil-size and anxiety: A review. In
I.G. Sarason, & C.D. Spielberger (Eds.), Stress and anxiety (Vol 3). Oxford,
England: Hemisphere.
Janisse, M.P. (1977). Pupillometry: The psychology of the pupillary response.
Washington, DC: Hemisphere Publishing Corporation.
Janisse, M.P., & Bradley, M.T. (1980). Deception, information and the pupillary
response. Perceptual & Motor Skills, 50(3), 748-750. doi:
10.2466/pms.1980.50.3.748
Janisse, M.P., & Peavler, W.S. (1974). Pupillary response today: Emotion in the eye.
Psychology Today, 7, 60-63.
Jasper, H.A. (1958). The ten–twenty electrode system of the International
Federation. Electroencepholography & Clinical Neurophysiology, 10(2), 371–
375. doi:10.1016/0013-4694(58)90053-1
Jepma, M., & Nieuwenhuis, S. (2011). Pupil diameter predicts changes in the
exploration-exploitation tradeoff: Evidence for the adaptive gain theory.
Journal of Cognitive Neuroscience, 23(7), 1587-1596.
doi:10.1162/jocn.2010.21548
Jiang, Y., Haxby, J.V., Martin, A., Ungerleider, L.G., & Parasuraman, R. (2000).
Complementary neural mechanisms for tracking items in human working
memory. Science, 287(5453), 643-646. doi:10.1126/science.287.5453.643
Jin, P. (1992). Towards a reconceptualization of the law of initial value.
Psychological Bulletin, 111(1), 176-184. doi:10.1037/0033-2909.111.1.176
Johnson, D.A. (1971). Pupillary responses during a short-term memory task:
Cognitive processing, arousal, or both? Journal of Experimental Psychology,
90(2), 311-318. doi:10.1037/h0031562
Johnson, R.J. (1986). A triarchic model of P300 amplitude. Psychophysiology, 23(4),
367-384. doi:10.1111/j.1469-8986.1986.tb00649.x
Johnson, R.J. (1995). Event-related potential insights into the neurobiology of
memory systems. In F. Boller, & J. Grafman (Eds.), Handbook of
neuropsychology (Vol. 10, pp. 135-163). Amsterdam: Elsevier.
300
Jones, T.C., & Jacoby, L.L. (2005). Conjunction errors in recognition memory:
Modality-free errors for older adults but not young adults. Acta Psychologica,
120(1), 55-73. doi:10.1016/j.actpsy.2005.03.003
Jouvet, M. (1969). Biogenic amines and the states of sleep. Science, 163(3862), 3241. doi:10.1126/science.163.3862.32
Just, M.A., & Carpenter, P.A. (1993). The intensity dimension of thought:
Pupillometric indices of sentence processing. Canadian Journal of
Experimental Psychology, 47(2), 310-339.
Just, M.A., & Carpenter, P.A. (1995). Spatial working memory: Cognitive capacity in
mental rotation. Unpublished manuscript, Carnegie Mellon University,
Pittsburgh, PA.
Just, M.A., Carpenter, P.A., Keller, T.A., Eddy, W.F., & Thulborn, K.R. (1996). Brain
activation modulated by sentence comprehension. Science, 274(5284), 114116. doi:10.1126/science.274.5284.114
Just, M.A., Carpenter, P.A., & Miyake, A. (2003). Neuroindices of cognitive
workload: Neuroimaging, pupillometric and event-related potential studies of
brain work. Theoretical Issues in Ergonomics Science, 4(1-2), 56-88.
doi:10.1080/14639220210159735
Kafkas, A., & Montaldi, D. (2011). Recognition memory strength is predicted by
pupillary responses at encoding while fixation patterns distinguish
recollection. The Quarterly Journal of Experimental Psychology, 64(10),
1971-1989. doi:10.1080/17470218.2011.588335
Kahneman, D. (1971). Changes in pupil-size and visual discrimination during mental
effort. In J.R. Pierce, & J.R. Levene (Eds.), Visual science: Proceedings of
the 1968 International Symposium. Bloomington, IN: Indiana University
Press.
Kahneman, D. (1973). Attention and effort. New Jersey: Prentice Hall.
Kahneman, D., & Beatty, J. (1966). Pupil diameter and load on memory. Science,
154(3756), 1583-1585. doi:10.1126/science.154.3756.1583
Kahneman, D., Beatty, J., & Pollack, I. (1967). Perceptual deficit during a mental
task. Science, 157(3785), 218-219. doi:10.1126/science.157.3785.218
Kahneman, D., Onuska, L., & Wolman, R.E. (1968a). Effects of grouping on the
pupillary response in a short-term memory task. Quarterly Journal of
Experimental Psychology, 20(3), 309-311. doi:10.1080/14640746808400168
301
Kahneman, D., & Peavler, W.S. (1969). Incentive effects and pupillary changes in
association learning. Journal of Experimental Psychology, 79(2), 312-318.
doi:10.1037/h0026912
Kahneman, D., Peavler, W.S., & Onuska, L. (1968b). Effects of verbalization and
incentive on the pupil response to mental activity. Canadian Journal of
Psychology, 22(3), 186-196. doi:10.1037/h0082759
Kahneman, D., Tursky, B., Shapiro, D., & Crider, A. (1969). Pupillary, heart rate, and
skin resistance changes during a mental task. Journal of Experimental
Psychology, 79(1), 164-167. doi:10.1037/h0026952
Kahneman, D., & Wright, P. (1971). Changes of pupil-size and rehearsal strategies
in a short-term memory task. Quarterly Journal of Experimental Psychology,
23(2), 187-196. doi:10.1080/14640747108400239
Kang, M.J., Hsu, M., Krajbich, I.M., Loewenstein, G., McClure, S.M., Wang, J.T., &
Camerer, C.F. (2009). The wick in the candle of learning: Epistemic curiosity
activates reward circuitry and enhances memory. Psychological Science,
20(8), 963-973. doi:10.1111/j.1467-9280.2009.02402.x
Kapucu, A., Macmillan, N.A., & Rotello, C.M. (2010). Positive and negative
remember judgments and ROCs in the plurals paradigm: Evidence for
alternative decision strategies. Memory & Cognition, 38(5), 541-554.
doi:10.3758/MC.38.5.541
Kardon, R. (1995). Pupillary light reflex. Current Opinion in Ophthalmology, 6(6), 2026.
Karis, D., Fabiani, M., & Donchin, E. (1984). "P300" and memory: Individual
differences in the von Restorff effect. Cognitive Psychology, 16(2), 177-216.
doi:10.1016/0010-0285(84)90007-0
Kasikci, T., & Bek, S. (2010). Malingering in neurology clinics. Turkiye Klinikleri
Journal of Medical Sciences, 30(4S), S35-S42.
Kelley, R., & Wixted, J.T. (2001). On the nature of associative information in
recognition memory. Journal of Experimental Psychology: Learning, Memory,
& Cognition, 27(3), 701-722. doi:10.1037/0278-7393.27.3.701
Kirschen, M.P., Davis-Ratner, M.S., Jerde, T.E., Schraedley-Desmond, P., &
Desmond, J.E. (2006). Enhancement of phonological memory following
transcranial magnetic stimulation. Behavioural Neurology, 17, 187-194.
302
Kiebel, S.J., & Friston, K.J. (2004). Statistical parametric mapping for event-related
potentials: I. Generic considerations. NeuroImage, 22(2), 492-502.
doi:10.1016/j.neuroimage.2004.02.012
Kim, M., Beversdorf, D.Q., & Heilman, K.M. (2000). Arousal response with aging:
Pupillographic study. Journal of the International Neuropsychological Society,
6(3), 348-350. doi:10.1017/S135561770000309X
Kintsch, W. (1967). Memory and decision aspects of recognition learning.
Psychological Review, 74(6), 496-504. doi:10.1037/h0025127
Klimesch, W. (1995). Memory processes described as brain oscillations in the EEGalpha and theta bands. Psycoloquy, 95(6), 23-37.
Knaggs, R.D., Crighton, I.M., Cobby, T.F., Fletcher, A.J.P., & Hobbs, G.J. (2004).
The pupillary effects of intravenous morphine, codeine, and tramadol in
volunteers. Anesthesia & Analgesia, 99(1), 108-112. doi:10.1213/
01.ANE.0000116924.16535.BA
Knopman, D., & Nissen, M.l. (1987). Implicit learning in patients with probable
Alzheimer's disease. Neurology, 37(5), 784-788.
Knowlton, B.J., Ramus, S., & Squire, L.R. (1992). Intact artificial grammar learning in
amnesia: Dissociation of classification learning and explicit memory for
specific instances. Psychological Science, 3(3), 172-179. doi:10.1111/j.14679280.1992.tb00021.x
Kohn, M., & Clynes, M. (1969). Color dynamics of the pupil. Annals of the New York
Academy of Sciences, 156, 931–950. doi:10.1111/j.17496632.1969.tb14024.x
Koikegami, H., & Yoshida, K. (1953). Pupillary dilatation induced by stimulation of
amygdaloid nuclei. Psychiatry & Clinical Neurosciences, 7(2), 109–126.
doi:10.1111/j.1440-1819.1953.tb00600.x
Kok, A. (1997). Event-related potential (ERP) reflections of mental resources: A
review and synthesis, Biological Psychology, 45(1-3), 19-56.
doi:10.1016/S0301-0511(96)05221-0
Konishi, S.,Wheeler, M.E., Donaldson, D.I., & Buckner, R.L. (2000). Neural
correlates of retrieval success. Neuroimage, 12(3), 276-286.
doi:10.1006/nimg.2000.0614
Kramer, A.F. (1991). Physiological metrics of mental workload: A review of recent
progress. In D.I. Damos (Ed.), Multi-task performance (pp. 279-323).
London: Taylor & Francis.
303
Krantz, D.H. (1969). Threshold theories of signal detection. Psychological Review,
76(3), 308-324. doi:10.1037/h0027238
Krinsky, R., & Gardner, R.M. (1977). Effects of retention interval on pupillary
responses during a simple discrimination. Perceptual & Motor Skills, 45(2),
457-458. doi:10.2466/pms.1977.45.2.457
Kruger, P.B. (1975). Luminance changes of the fundus reflex. American Journal of
Optometry & Physiological Optics, 52(12), 847-861.
Kruger, P.B. (1977a). Changes in fundus reflex luminance with increased cognitive
processing. American Journal of Optometry & Physiological Optics, 54(7),
445-51.
Kruger, P.B. (1977b). The role of accommodation in increasing the luminance of the
fundus reflex during cognitive processing. Journal of the American Optometric
Association, 48(12), 1493–1496.
Kucera, H., & Francis, W.N. (1967). Computational analysis of present-day
American English. Providence, RI: Brown University Press.
Kuipers, J.R., & Thierry, G. (2011). N400 amplitude reduction correlates with an
increase in pupil-size. Frontiers in Human Neuroscience, 5(61), 1-5.
doi:10.3389/fnhum.2011.00061
Kumnick, L.S. (1954). Pupillary psychosensory restitution and aging. Journal of the
Optical Society of America, 44(9), 735-741. doi:10.1364/JOSA.44.000735
Kumnick, L.S. (1956). Aging and pupillary response to light and sound stimuli.
Journal of Gerantology, 11(1), 38-45. doi:10.1093/geronj/11.1.38
Kuntz, A. (1929). The autonomic nervous system. Philadelphia: Lea & Febiger.
Kutas, M., & Hillyard, S.A. (1980). Reading senseless sentences: Brain potentials
reflect semantic incongruity. Science, 207(4427), 203-205.
doi:10.1126/science.7350657
Lacey, J.I. (1959). Psychophysiological approaches to the evaluation of
psychotherapeutic process and outcome. In E.A. Rubinstein, & M.B. Parloff
(Eds.), Research in psychotherapy (pp. 160-208). Washington, DC: American
Psychological Association.
Lacey, J.I. (1967). Somatic response patterning and stress: Some revisions of
activation theory. In M.H. Appley, & R. Trumbull (Eds.), Psychological stress:
Issues for research (pp. 14-42). New York: Appleton-Century-Crofts.
Laeng, B., Waterloo, K., Johnsen, S.H., Bakke, S.J., Låg, T., Simonsen, S.S., &
Høgsæt, J. (2007). The eyes remember it: Oculography and pupillometry
304
during recollection in three amnesic patients. Journal of Cognitive
Neuroscience, 19(11), 1888-1904. doi:10.1162/jocn.2007.19.11.1888
Lanatà, A., Armato, A., Valenza, G., & Scilingo, E.P. (2011). Eye tracking and pupilsize variation as response to affective stimuli: A preliminary study. In
Proceedings from The Pervasive Computing Technologies for Healthcare 5th
International Conference. Dublin, Ireland: Institute for Computer Sciences,
Social Informatics and Telecommunications Engineering.
doi:10.4108/icst.pervasivehealth.2011.246056
Lange, G., Janal, M.N., Maniker, A., FitzGibbons, J., Fobler, M., Cook, D., &
Natelson, B.H. (2011). Safety and efficacy of vagus nerve stimulation in
fibromyalgia: A phase I/II proof of concept trial. Pain Medicine, 12(9), 14061413. doi:10.1111/j.1526-4637.2011.01203.x
Lange, R.T., Senior, G., Douglas, L., & Dawes, S. (2003). Base rates and clinical
utility of the WMS-III Rarely Missed Index: Identifying cognitive exaggeration
in a head injury litigant sample. Archives of Clinical Neuropsychology, 18,
786.
Lange, R.T., Sullivan, K., & Anderson, D. (2005). Ecological validity of the WMS-III
Rarely Missed Index in personal litigation. Journal of Clinical & Experimental
Neuropsychology, 27(4), 412-424. doi:10.1080/13803390490520319
Langeluddecke, P.M. (2004). Validation of the Rarely Missed Index (RMI) in
detecting memory malingering in mild head injury litigants. Journal of
Forensic Neuropsychology, 4(1), 49-64.
Larrabee, G.J. (2003). Detection of malingering using atypical performance patterns
on standard neuropsychological tests. The Clinical Neuropsychologist, 17(3),
410-425. doi:10.1076/clin.17.3.410.18089
Larsen, R.J., Mercer, K.A., & Balota, D.A. (2006). Lexical characteristics of words
used in emotional Stroop experiments. Emotion, 6(1), 62-72.
doi:10.1037/1528-3542.6.1.62
Lawless, J.C., & Wake, F.R. (1969). Sex differences in pupillary response to visual
stimuli. Psychophysiology, 5(5), 568.
Leal, G., Neves, C., & Vieira, P. (2011). Wavelet analysis of the autonomic outflow of
the pupil's variation in normal subjects, given stimuli. In Proceedings from The
1st Portuguese Meeting in Biomedical Engineering (pp. 1-4). Lisbon,
Portugal: IEEE. doi:10.1109/ENBENG.2011.6026066
305
Leding, J.K., & Lampinen, J.M. (2009). Memory conjunction errors: The effects of
presentation duration and study repetition. Memory, 17(5), 597-607.
doi:10.1080/09658210902984518
Ledoux, K., Pickett, E., Van Droof, L.V., Buz, E., Billings, N.M., & Gordon, B. (2010).
Receptive vocabulary knowledge in individuals with autism as assessed by
eye movements, pupillary dilation, and event-related potentials. Presented at
The 2010 Neurobiology of Autism Conference. San Diego.
Lees-Haley, P.R., English, L.T., & Glenn, W.J. (1991). A fake bad scale for the
MMPI-2 for personal injury claimants. Psychological Reports, 68, 203-210.
Lehr, D.J., & Bergum, B. (1966). Note on pupillary adaptation. Perceptual & Motor
Skills, 23(3), 917-918. doi:10.2466/pms.1966.23.3.917
Liakos, A., & Crisp, A.H. (1971). Pupil-size in psychoneurotic patients.
Psychotherapy & Psychosomatics, 19(1-2), 104-110. doi:10.1159/000286311
Libby, W.L., Lacey, B.C, & Lacey, J.I. (1973). Pupillary and cardiac activity during
visual attention. Psychophysiology, 10(3), 270-294. doi:10.1111/j.14698986.1973.tb00526.x
Lin, T., Imamiya, A., & Mao, X. (2008). Using multiple data sources to get closer
insights into user cost and task performance. Interacting with Computers,
20(3), 364-374. doi:10.1016/j.intcom.2007.12.002
Lockhart, R.S., & Murdock, B.B. (1970). Memory and the theory of signal detection,
Psychological Review, 74(2), 100-109. doi:10.1037/h0029536
Loewenfeld, I. (1999). The pupil: Anatomy, physiology, and clinical applications (2nd
ed., Vol. 1). Oxford, UK: Butterworth-Heinemann.
Loewenfeld, I.E. (1958). Mechanisms of reflex dilatation of the pupil; historical review
and experimental analysis. Documenta Ophthalmologica Proceedings Series,
12, 185-448. doi:10.1007/BF00913471
Loewenfeld, I.E. (1966). Comment on Hess' findings. Survey of Ophthalmology, 11,
291-294.
Loewy, A.D. (1990). Autonomic control of the eye. In A.D. Loewy, & K.M. Spyer
(Eds.), Central regulation of autonomic function (pp. 268-285). New York:
Oxford University Press.
Lowenstein, O. (1955). Pupillary reflex shapes and topical clinical diagnosis.
Neurology, 5(9), 631-644.
306
Löwenstein, O., & Friedman, E.D. (1942). Pupillographic studies. 1. Present state of
pupillography: Its method and diagnostic significance. Archives of
Ophthalmology, 27(5), 969-993.
Löwenstein, O., & Loewenfeld, I.E. (1952). Disintegration of central autonomic
regulation during fatigue. Journal of Nervous and Mental Disease, 115(2),
121-145.
Löwenstein, O., & Loewenfeld, I.E. (1958). Electronic pupillography: A new
instrument and some clinical applications. American Medical Association
Archives of Ophthalmology, 59(3), 352-363.
Löwenstein, O., & Loewenfield, I.E. (1962). The pupil. In H. Davson (Ed.), The eye:
Muscular mechanisms (pp. 301-340). New York: Academic Press.
Löwenstein, O., & Loewenfield, I.E. (1964). The sleep-waking cycle and pupillary
activity. Annals of the New York Academy of Sciences, 117, 142-156.
doi:10.1111/j.1749-6632.1964.tb48169.x
Lubow, R.E., & Fein, O. (1996). Pupillary size in response to a visual guilty
knowledge test: New technique for the detection of deception. Journal of
Experimental Psychology-Applied, 2(2), 164-177. doi:10.1037//1076898X.2.2.164
Luck, S.J. (2005). An introduction to the event-related potential technique.
Cambridge, MA: The MIT Press.
Luck, S.J. (2010). Is it legitimate to compare conditions with different numbers of
trials? Retrieved 17 November 2011 from
http://erpinfo.org/Members/sjluck/Mean_Peak_Noise.pdf/view
Luu, P., & Ferree, T. (2000). Determination of the Geodesic Sensor Nets’ electrode
positions and their 10-10 international equivalents: Technical note. Eugene,
OR: Electrical Geodesics, Inc. Retrieved 26 June 2011 from
ftp://ftp.egi.com/pub/documentation/technotes/200_ElectrodePositions.pdf
Lykken, D.T. (1959). The GSR in the detection of guilt. Journal of Applied
Psychology, 43(6), 385-388. doi:10.1037/h0046060
Lykken, D.T. (1960). The validity of the guilty knowledge technique: The effects of
faking. Journal of Applied Psychology, 44(4), 258-262. doi:10.1037/h0044413
Lykken, D.T. (1972). Range correction applied to heart rate and GSR data.
Psychophysiology, 9(3), 373-379. doi:10.1111/j.1469-8986.1972.tb03222.x
Lykken, D.T. (1998). A tremor in the blood: Uses and abuses of the lie detector. New
York: Plenum Trade.
307
Lynn, R. (1966). Attention, arousal, and the orientation reaction. Oxford: Pergamon
Press.
Machizawa, M.G., Kalla, R., Walsh, V., & Otten, L.J. (2010). The time course of
ventrolateral prefrontal cortex involvement in memory formation. Journal of
Neurophysiology, 103(3), 1569-1579. doi:10.1152/jn.90937.2008
MacKenzie, G., & Donaldson, D.I. (2007). Dissociating recollection from familiarity:
Electrophysiological evidence that familiarity for faces is associated with a
posterior old/new effect. NeuroImage, 36, 454-463.
doi:10.1016/j.neuroimage.2006.12.005
MacLeod, C.M., & Kampe, K.E. (1996). Word frequency effects on recall,
recognition, and word fragment completion tests. Journal of Experimental
Psychology: Learning, Memory, & Cognition, 22(1), 132-142.
doi:10.1037/0278-7393.22.1.132
Macmillan, N.A., & Creelman, C.D. (1990). Response bias: Characteristics of
detection theory, threshold theory, and ‘‘nonparametric” indexes.
Psychological Bulletin, 107(3), 401-413. doi:10.1037//0033-2909.107.3.401
Macmillan, N.A., & Rotello, C.M. (2006). Deciding about decision models of
remember and know judgments: A reply to Murdock. Psychological Review,
113, 657-665. doi:10.1037/0033-295X.113.3.657
Malmberg, K.J. (2008). Recognition memory: A review of the critical findings and an
integrated theory for relating them. Cognitive Psychology, 57(4), 335-384.
doi:10.1016/j.cogpsych.2008.02.004
Malmberg, K.J. (2010). An annotated reading list for recognition memory. Retrieved
14 October 2011 from
http://uweb.cas.usf.edu/~malmberg/Recognition%20Readings.htm
Malmberg, K.J., Holden, J.E., & Shiffrin, R.M. (2004). Modeling the effects of
repetitions, similarity, and normative word frequency on judgments of
frequency and recognition memory. Journal of Experimental Psychology:
Learning, Memory, & Cognition, 30(2), 319-331. doi:10.1037/02787393.30.2.319
Mandler, G. (1980). Recognizing: The judgment of previous occurrence.
Psychological Review, 87(3), 252-271. doi:10.1037/0033-295X.87.3.252
Mandler, G. (2008). Familiarity breeds attempts: A critical review of dual process
theories of recognition. Perspectives in Psychological Science, 3(5), 392-401.
doi:10.1111/j.1745-6924.2008.00087.x
308
Mandler, G., & Boeck, W. (1974). Retrieval processes in recognition. Memory &
Cognition, 2(4), 613-615. doi:10.3758/BF03198129
Manns, J.R., Hopkins, R.O., Reed, J.M., Kitchener, E.G., & Squire, L.R. (2003).
Recognition memory and the human hippocampus. Neuron, 37(1), 171-180.
doi:10.1016/S0896-6273(02)01147-9
Maratos, E.J., Allan, K., & Rugg, M.D. (2000). Recognition memory for emotionally
negative and neutral words: An ERP study. Neuropsychologia, 38(11), 14521465. doi:10.1016/S0028-3932(00)00061-0
Marshall, S.P. (2002). The index of cognitive activity: measuring cognitive workload.
In Proceedings from The IEEE 7th Conference on Human Factors and Power
Plants, 7-5-7-9. doi:10.1109/HFPP.2002.1042860
Marshall, S.P. (2007). Identifying cognitive state from eye metrics. Aviation, Space,
& Environmental Medicine, 78(5,Suppl.), B165-B175.
Martineau, J., Hernandez, N., Hiebel, L., Roché, L., Metzger, A., & Bonnet-Brilhault,
F. (2011). Can pupil-size and pupil responses during visual scanning
contribute to the diagnosis of autism spectrum disorder in children? Journal of
Psychiatric Research, 45(8), 1077-1082.
doi:10.1016/j.jpsychires.2011.01.008
Maw, N.N., & Pomplun, M. (2004). Studying human face recognition with the gazecontingent window technique. In K. Forbus, D. Gentner, & T. Regier (Eds.),
Proceedings of the Twenty-Sixth Annual Meeting of the Cognitive Science
Society. Chicago, Illinois: Cognitive Science Society.
McCarthy, G., & Wood, C.C. (1985). Scalp distributions of event-related potentials:
An ambiguity associated with analysis of variance models. Clinical
Neurophysiology, 62(3), 203-208. doi:10.1016/0168-5597(85)90015-2
McClelland, J.L., & Chappell, M. (1998). Familiarity breeds differentiation: A
subjective-likelihood approach to the effects of experience in recognition
memory. Psychological Review, 105(4), 724-760. doi:10.1037/0033295X.105.4.734-760
McElvain, J.L. (1970). Pupillary dilations and galvanic skin response activity during
competitive and noncompetitive paired association learning. (Doctoral
dissertation). Dissertation Abstracts International, 30, 3410B-3411B.
McLaren, J.W., Erie, J.C., & Brubaker, R.F. (1992). Computerized analysis of
pupillograms in studies of alertness. Investigative Ophthalmology & Visual
Science, 33(3), 671-676.
309
Mecklinger, A. (2000). Interfacing mind and brain: A neurocognitive model of
recognition memory. Psychophysiology, 37(5), 565-582. doi:10.1111/14698986.3750565
Mecklinger, A., von Cramon, D.Y., & Matthes-von Cramon, G. (1998). Event-related
potential evidence for a specific recognition memory deficit in adult survivors
of cerebral hypoxia. Brain, 121(10), 1919-1935.
doi:10.1093/brain/121.10.1919
Medina, J.J. (2008). The biology of recognition memory. Psychiatric Times.
Retrieved March 14, 2009, from
http://www.brainrules.net/pdf/JohnMedina_PsychTimes_June08.pdf
Mentz, P. (1895). Die wirkung akustischer sinnesreize auf puls und athmung.
Philosophische Studien, 11, 61-124.
Merckelbach, H., & Christianson, S.A. (2007). Amnesia for homicides as a form of
malingering. In S.A. Christianson (Ed.), Offenders’ memories of violent crimes
(pp. 165–190). Chichester, England: Wiley
Merten, T., Thies, E., Schneider, K., & Stevens, A. (2009). Symptom validity testing
in claimants with alleged posttraumatic stress disorder: Comparing the morel
emotional numbing test, the structured inventory of malingered
symptomatology, and the word memory test. Psychological Injury & Law, 2(34), 284-293. doi:10.1007/s12207-009-9057-0
Michel, C.M., Murray, M.M., Lantz, G., Gonzalez, S., Spinelli, L., & Grave de Peralta,
R. (2004). EEG source imaging. Clinical Neurophysiology, 115(10), 21952222. doi:10.1016/j.clinph.2004.06.001
Miller, E.K., & Desimone, R. (1994). Parallel neuronal mechanisms for short-term
memory. Science, 263(5146), 520-523. doi:10.1126/science.8290960
Miller, G.A. (1956). The magic number seven, plus or minus two: Some limits on our
capacity for processing information. Psychological Review, 63(2), 81-97.
Miller, L.J., Ryan, J.J., Carruthers, C.A., & Cluff, R.B. (2004). Brief screening
indexes for malingering: A confirmation of Vocabulary minus Digit Span from
the WAIS-III and the Rarely Missed Index from the WMS-III. The Clinical
Neuropsychologist, 18, 327-333. doi:10.1080/13854040490501592
Miller, N.R., & Newman, N.J. (Eds.) (2005). Walsh & Hoyt's clinical neuroophthalmology: Volume one (6th ed.). USA: Lippincott Williams & Wilkins.
310
Miller, R.L. (1967). The clinical validation of the pupillary response: The effect of
chromatic and achromatic stimuli upon pupil responsivity. (Unpublished
doctoral dissertation). Michigan State University.
Miller, S.R., & Donders, J. (2001). Subjective symptomology after traumatic head
injury. Brain Injury, 15(4), 296-304. doi:10.1080/02699050118115
Minassian, A., Granholm, E., Verney, S., & Perry, W. (2004). Pupillary dilation to
simple versus complex tasks and its relationship to thought disturbance in
schizophrenia patients. International Journal of Psychophysiology, 52(1), 5362. doi:10.1016/j.ijpsycho.2003.12.008
Mittenberg, W., Patton, C., Canyock, E.M., & Condit, D.C. (2002). Base rates of
malingering and symptom exaggeration. Journal of Clinical & Experimental
Neuropsychology, 24(8), 1094-1102. doi:10.1076/jcen.24.8.1094.8379
Mohamed, F.B., Faro, S.H., Gordon, N.J., Platek, S.M., Ahmad, H., & Williams, J.M.
(2006). Brain mapping of deception and truth telling about an ecologically
valid situation: Functional MR imaging and polygraph investigation - Initial
experience. Neuroradiology, 238(2), 679-688. doi:10.1148/radiol.2382050237
Molnar, M. (1994). On the origin of the P3 event-related potential component.
International Journal of Psychophysiology, 17(2), 129-144. doi:10.1016/01678760(94)90028-0
Moloney, K.P., Jacko, J.A., Vidakovic, B., Sainfort, F., Leonard, V.K., & Shi, B.
(2006). Leveraging data complexity: Pupillary behavior of older adults with
visual impairment during HCI. ACM Transactions on Computer-Human
Interaction, 13(3), 376-402. doi:10.1145/1183456.1183460
Moray, N.E. (1979). Mental workload: Its theory and measurement. New York,
Plenum Press.
Mueller, D.J. (1970). Physiological techniques of attitude measurement. In G.F.
Summers (Ed.), Attitude measurement (pp. 534-552). Chicago, IL: Rand
McNally.
Muldner, K., Christopherson, R., Atkinson, R., & Burleson, W. (2009). Investigating
the utility of eye-tracking information on affect and reasoning for user
modeling user modeling, adaptation, and personalization. In G.-J. Houben, G.
McCalla, & F. Pianesi (Eds.), Proceedings from The 17th International
Conference on User Modeling, Adaptation, and Personalization (pp. 138149). Heidelberg, Germany: Springer-Verlag. doi:10.1007/978-3-642-022470_15
311
Mulhall, E.F. (1915). Experimental studies in recall and recognition. The American
Journal of Psychology, 26(2), 217-228.
Müller-Jensen, A., & Hagenah, R. (1978). Simultaneous recording of pupillary hippus
and EEG: Report of a case. Journal of Neurology, 217(3), 213-218.
doi:10.1007/BF00312964
Murdock, B.B. (1982). A theory of the storage and retrieval of item and associative
information. Psychological Review, 89(6), 609–626. doi:10.1037/0033295X.89.6.609
Murdock, B.B. (1985). An analysis of the strength-latency relationship. Memory &
Cognition, 13(6), 511-521. doi:10.3758/BF03198322
Murdock, B.B. (1997). Context and mediators in a theory of distributed associative
memory (TODAM2). Psychological Review, 104(4), 839-862.
Murdock, B.B. (2006). Decision-making models of remember–know judgments:
Comment on Rotello, Macmillan, and Reeder (2004). Psychological Review,
113(3), 648-656. doi:10.1037/0033-295x.113.3.648
Murphy, P.R., Robertson, I.H., Balsters, J.H., & O’Connell, R.G. (2011). Pupillometry
and P3 index the locus coeruleus-noradrenergic arousal function in humans.
Psychophysiology, 48(11), 1532-1543. doi:10.1111/j.14698986.2011.01226.x
Myers, C., & Connor, M. (1992). Age differences in skill acquisition and transfer in an
implicit learning paradigm. Applied Cognitive Psychology, 6(5), 429-442.
doi:10.1002/acp.2350060507
Mylius, V., Braune, H.J., & Schepelmann, K. (2003). Dysfunction of the pupillary light
reflex following migraine headache. Clininical Autonomic Research, 13(1), 1621. doi:10.1007/s10286-003-0065-y
Näätänen, R. (1990). The role of attention in auditory information processing as
revealed by event-related potentials and other brain measures of cognitive
function. Behavioral & Brain Sciences, 13, 201-288.
Naber, M., Rutishauser, U., & Einhäuser, W. (2011). Pupil size reflects familiarity
and predicts later retrieval success during encoding of declarative memories.
Unpublished manuscript. Philipps-Universität Marburg.
Nakayama, M. (2006). Influence of blink on pupillary indices. In Proceedings from
BioCAS: The Biomedical Circuits and Systems Conference (pp. 29-32).
London: BioCAS. doi:10.1109/BIOCAS.2006.4600300
312
Nakayama, M., & Shimizu, Y. (2004). Frequency analysis of task evoked pupillary
response and eye-movement. In Proceedings from The ACM Eye Tracking
Research & Applications Symposium (pp. 71-76). New York, NY: ACM Press.
doi:10.1145/968363.968381
Nakayama, M., Yasuike, I., & Shimizu, Y. (1990). Pupil-size changing by pattern
brightness and pattern contents. The Journal of the Institute of Television
Engineers of Japan, 44, 288-293.
National Research Council. (2003). The polygraph and lie detection. Washington,
DC: National Academies Press.
Nave, C.R. (2010). Accommodation. In Hyperphysics. Retrieved from
http://hyperphysics.phy-astr.gsu.edu/hbase/vision/accom.html
Naveh-Benjamin, M., Craik, F.I.M., Guez, J., & Dori, H. (1998). The effects of divided
attention on encoding and retrieval processes in human memory: Further
support for an asymmetry. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 24(5), 1091-1104.
Neiss, R. (1988). Reconceptualizing arousal: Psychobiological states in motor
performance. Psychological Bulletin, 103(3), 345-366. doi:10.1037/00332909.103.3.345
Nessler, D., Mecklinger, A., & Penney, T.B. (2005). Perceptual fluency, semantic
familiarity, and recognition-related familiarity: An electrophysiological
exploration. Cognitive Brain Research, 22(2), 265-288.
doi:10.1016/j.cogbrainres.2004.03.023
Nieuwenhuis, S., Aston-Jones, G., & Cohen, J.D. (2005a). Decision making, the P3,
and the locus coeruleus-norepinephrine system. Psychological Bulletin,
131(4), 510-532. doi:10.1037/0033-2909.131.4.510
Nieuwenhuis, S., De Geus, E.J., & Aston-Jones, G. (2011a). The anatomical and
functional relationship between the P3 and autonomic components of the
orienting response. Psychophysiology, 48(2), 162-175. doi:10.1111/j.14698986.2010.01057.x
Nieuwenhuis, S., Forstmann, B.U., & Wagenmakers, E.-J. (2011b). Erroneous
analyses of interactions in neuroscience: A problem of significance. Nature
Neuroscience, 14(9), 1105-1107. doi:10.1038/nn.2886
Nieuwenhuis, S., Gilzenrat, M.S., Holmes, B.D., & Cohen, J.D. (2005b). The role of
the locus coeruleus in mediating the attentional blink: A neurocomputational
313
theory. Journal of Experimental Psychology: General, 134(3), 291-307.
doi:10.1037/0096-3445.134.3.291
Nissen, M.J., & Bullemer, P. (1987). Attentional requirements of learning: Evidence
from performance measures. Cognitive Psychology, 19(1), 1-32.
doi:10.1016/0010-0285(87)90002-8
Nissen, M.J., Willingham, D., & Hartman, M. (1989). Explicit and implicit
remembering: When is learning preserved in amnesia? Neuropsychologia,
27(3), 341-352. doi:10.1016/0028-3932(89)90023-7
Norman, D.A., & Bobrow, D.J. (1975). On data limited and resource limited
processes. Cognitive Psychology, 7(1), 44-64. doi:10.1016/00100285(75)90004-3
Norman, E., Price, M., Duff, S., & Mentzoni, R. (2007). Gradations of awareness in a
modified sequence learning task. Consciousness and Cognition, 16(4):809837. doi:10.1016/j.concog.2007.02.004
Norman, K.A., & O’Reilly, R.C. (2003). Modeling hippocampal and neocortical
contributions to recognition memory: A complementary learning systems
approach. Psychological Review, 110(4), 611-646. doi:10.1037/0033295X.110.4.611
Norman, K.A., Tepe, K., Nyhus, E., & Curran, T. (2008). Event-related potential
correlates of interference effects on recognition memory. Psychonomic
Bulletin & Review, 15(1), 36-43. doi:10.3758/PBR.15.1.36
Nunez, P.L. (1981). Electric fields of the brain. New York: Oxford University Press.
Nunnally, C., Knott, P.D., Duchnowski, J., & Parker, K. (1967). Pupillary response as
a general measure of activation. Attention, Perception & Psychophysics, 2(4),
149-155. doi:10.3758/BF03210310
O’Bryant, S., & Lucas, J. (2006). Estimating the predictive value of the Test of
Memory Malingering: An illustrative example for clinicians. Clinical
Neuropsychology, 20(3), 533-540. doi:10.1080/13854040590967568
O'Connor, M.K., & Ally, B.A. (2010). Using stimulus form change to understand
memorial familiarity for pictures and words in patients with mild cognitive
impairment and Alzheimer's disease. Neuropsychologia, 48(7), 2068-2074.
doi:10.1016/j.neuropsychologia.2010.03.027
Ohira, H. (1996). Eyeblink activity in a word-naming task as a function of semantic
priming and cognitive load. Perceptual & Motor Skills, 82(3), 835-842.
doi:10.2466/pms.1996.82.3.835
314
Ohira, H., Winton, W.M., & Oyama, M. (1998). Effects of stimulus valence on
recognition memory and endogenous eyeblinks: Further evidence for
positive-negative asymmetry. Personality & Social Psychology Bulletin, 24(9),
986-993. doi:10.1177/0146167298249006
Ohno, T., & Mukawa, N. (2004). A free-head, simple calibration, gaze tracking
system that enables gaze-based interaction. In Proceedings from The 2004
Symposium on Eye Tracking Research & Applications (pp. 115-122). New
York, NY: ACM. doi:10.1145/968363.968387
Olichney, J.M., Van Petten, C., Paller, K.A., Salmon, D.P., Iragui, V.J., & Kutas, M.
(2000). Word repetition in amnesia: Electrophysiological measures of
impaired and spared memory. Brain, 123(9), 1948-1963.
doi:10.1093/brain/123.9.1948
Oliveira, F.T.P., Aula, A., & Russell, D.M. (2009). Discriminating the relevance of
web search results with measures of pupil-size. In Proceedings from The 27th
International Conference on Human Factors in Computing Systems (22092212). New York, NY: ACM Press. doi:10.1145/1518701.1519038
O'Neill, W.D., Oroujeh, A.M., & Merritt, S.L. (1998). Pupil noise is a discriminator
between narcoleptics and controls. IEEE Transactions on Biomedical
Engineering, 45(3), 314-322. doi:10.1109/10.661156
O'Neill, W.D., Oroujeh, A.M., Keegan, A.P., & Merritt, S.L. (1996). Neurological
pupillary noise in narcolepsy. Journal of Sleep Research, 5(4), 265-271.
doi:10.1111/j.1365-2869.1996.00265.x
O'Neill, W.D., & Trick, K.P. (2001). The narcoleptic cognitive pupillary response.
IEEE Transactions on Biomedical Engineering, 48(9), 963-968.
doi:10.1109/10.942585
Opitz, B. (2010). Context-dependent repetition effects on recognition memory. Brain
& Cognition, 73(2), 110-118. doi:10.1016/j.bandc.2010.04.003
Otero, S.C. (2010). False memory in normal ageing. (Unpublished doctoral
dissertation). University of Sussex, England.
Otero, S., Weekes, B., & Hutton, S. (2006). A novel association between pupil-size
and recollective experience during recognition memory. In Proceedings from
The Second Biennial Conference on Cognitive Science. Retreived 12 May
2011 from http://www.cogsci.ru/cogsci06/index_e.htm
315
Otero, S.C., Weekes, B.S., & Hutton, S.B. (2011). Pupil-size changes during
recognition memory. Psychophysiology, 48(10), 1346-1353.
doi:10.1111/j.1469-8986.2011.01217.x
Paivio, A. (1973). Psychophysiological correlates of imagery. In F.J. McGuigan, & R.
Schoonover (Eds.), The psychophysiology of thinking (pp. 263-295). New
York: Academic Press.
Paivio, A., & Simpson, H.M. (1966). The effect of word abstractness and
pleasantness on pupil-size during an imagery task. Psychonomic Science,
5(2), 55-56.
Paivio, A., & Simpson, H.M. (1968). Magnitude and latency of the pupillary response
during an imagery task as a function of stimuli abstractness and imagery
ability. Psychonomic Science, 12(2), 45-46.
Paller, K.A. (1990). Recall and stem-completion priming have different
electrophysiological correlates and are modified differentially by directed
forgetting. Journal of Experimental Psychology: Learning, Memory, &
Cognition, 16(6), 1021-1032. doi:10.1037/0278-7393.16.6.1021
Paller, K.A., Gonsalves, B., Grabowecky, M., Bozic, V.S., & Yamada, S. (2000).
Electrophysiological correlates of recollecting faces of known and unknown
individuals. NeuroImage, 11(2), 98-110. doi:10.1006/nimg.1999.0521
Paller, K.A., & Kutas, M. (1992). Brain potentials during memory retrieval provide
neurophysiological support of the distinction between conscious recollection
and priming. Journal of Cognitive Neuroscience, 4, 375-391.
doi:10.1162/jocn.1992.4.4.375
Paller, K.A., Kutas, M., & McIsaac, H.K. (1995). Monitoring conscious recollection
via the electrical activity of the brain. Psychological Science, 6(2), 107-111.
doi:10.1111/j.1467-9280.1995.tb00315.x
Paller, K.A., Voss, J.L., & Boehm, S.G. (2007). Validating neural correlates of
familiarity. Trends in Cognitive Sciences, 11(6), 243-250.
doi:10.1016/j.tics.2007.04.002
Papesh, M.H., & Goldinger, S.D. (2011). Your effort is showing! Pupil dilation reveals
memory heuristics. In P.A. Higham, & J.P. Leboe (Eds.), Constructions of
remembering and metacognition (pp. 215-224). Hampshire: Palgrave
Macmillan.
316
Papesh, M.H., Goldinger, S.D., & Hout, M.C. (2011). Memory strength and
specificity revealed by pupillometry. International Journal of
Psychophysiology, 83(1), 56-64. doi:10.1016/j.ijpsycho.2011.10.002
Park, H., Reder, L.M., & Dickison, D. (2005). The effects of word frequency and
similarity on recognition judgments: The role of recollection. Journal of
Experimental Psychology: Learning, Memory, & Cognition, 31(3), 568-578.
doi:10.1037/0278-7393.31.3.568
Parks, C.M., & Yonelinas, A.P. (2007). Moving beyond pure signal detection models:
Comment on Wixted (2007). Psychological Review, 114(1), 188-202.
doi:10.1037/0033-295X.114.1.188
Partala, T., & Surakka, V. (2003). Pupil-size variation as an indication of affective
processing. International Journal of Human-Computer Studies, 59, 185-198.
doi:10.1016/S1071-5819(03)00017-X
Pavlov, I.P. (1927). Conditioned reflexes. Oxford: Clarendon Press.
Payne, D.T., Parry, M.K., & Harasymiw, S.J. (1968). Percentage of pupillary dilation
as a measure of item difficulty. Perception & Psychophysics, 4(3), 139-143.
doi:10.3758/BF03210453
Peavler, W.S. (1974), Pupil-size, information overload, and performance differences.
Psychophysiology, 11(5), 559–566. doi:10.1111/j.1469-8986.1974.tb01114.x
Peavler, W.S., & McLaughlin, J.P. (1967). The question of stimulus content and
pupil-size. Psychonomic Science, 8(12), 505-506.
Peter, G., Merten, T., Merckelbach, H., & Oswald, M. (2010). Detection of feigned
crime-related amnesia: A multi-method approach. Journal of Forensic
Psychology Practice, 10(5), 440-463. doi:10.1080/15228932.2010.489875
Phillips, M.A., Bitsios, P., Szabadi, E., & Bradshaw, C.M. (2000a). Comparison of
the antidepressants reboxetine, fluvoxamine and amitriptyline upon
spontaneous pupillary fluctuations in healthy human volunteers.
Psychopharmacology, 149(1), 72-76. doi:10.1007/s002139900334
Phillips, M.A., Szabadi, E., & Bradshaw, C.M. (2000b). Comparison of the effects of
clonidine and yohimbine on spontaneous pupillary fluctuations in healthy
human volunteers. Psychopharmacology, 150(1), 85-89. doi: 10.1046/ j.13652125.2000.00225.x
Pivik, R.T., Broughton, R.J., Coppola, R., Davidson, R.J., Fox, N., & Nuwer, M.R.
(1993). Guidelines for the recording and quantitative analysis of
317
electroencephalographic activity in research contexts. Psychophysiology,
30(6), 547-588. doi:10.1111/j.1469-8986.1993.tb02081.x
Pliny the Elder (1938). Natural history with an English translation in ten volumes:
volume III, libri VIII-XI (H. Rackham, Trans.). London: William Heinemann.
(Original work published 79 AD).
Plonsey, R. (1963). Reciprocity applied to volume conductors and the EEG. IEEE
Transactions on Biomedical Engineering, 19, 9-12.
Polich, J., & Squire, L.E. (1993). P300 from amnesic patients with bilateral
hippocampal lesions. Electroencephalography & Clinical Neurophysiology,
86(6), 408-417. doi:10.1016/0013-4694(93)90136-J
Polt, J.M. (1970). Effect of threat of shock on pupillary response in a problem-solving
task. Perceptual & Motor Skills, 31(2), 587-593. doi:10.2466/pms.
1970.31.2.587
Pomplun, M., & Sunkara, S. (2003). Pupil dilation as an indicator of cognitive
workload in human-computer interaction. In D. Harris, V. Duffy, M. Smith, &
C. Stephanidis (Eds.), Proceedings from The 10th International Conference
on Human-Computer Interaction - Human-centred computing: Cognitive,
social, and ergonomic aspects, Vol. 3 (pp. 542-546). Crete, Greece: HCII.
Pomplun, M., Sunkara, S., Fairley, A.V., & Xiao, M. (submitted). Using pupil-size as
a measure of cognitive workload in video-based eye-tracking studies.
Poock, G.K. (1973). Information processing vs. pupil diameter. Perceptual & Motor
Skills, 37(3), 1000-1002.
Porter, G., Tales, A., Trościanko, T., Wilcock, G., Haworth, J., & Leonards, U.
(2010). New insights into feature and conjunction search: I. Evidence from
pupil-size, eye movements and ageing. Cortex, 46(5), 621-636.
doi:10.1016/j.cortex.2009.04.013
Porter, G., Trościanko, T., & Gilchrist, I.D. (2007). Effort during visual search and
counting: Insights from pupillometry. The Quarterly Journal of Experimental
Psychology, 60(2), 211-229. doi:10.1080/17470210600673818
Pothos, E.M. (2007). Theories of artificial grammar learning. Psychological Bulletin,
133(2), 227-244. doi:10.1037/0033-2909.133.2.227
Pratt, R.W. (1970). Cognitive processing of uncertainty: Its effects on pupillary
dilation and preference ratings. Perception & Psychophysics, 8(4), 193-198.
doi:10.3758/BF03210204
318
Pressman, M.R., Spielman, A.J., Korczyn, A.D., Rubenstein, A.E., Pollak, C.P., &
Weitzman, E.D. (1984). Patterns of daytime sleepiness in narcoleptics and
normals: A pupillometric study. Electroencephalography & Clinical
Neurophysiology, 57(2), 129-133. doi:10.1016/0013-4694(84)90171-8
Prime, S.L., Vesia, M., & Crawford, J.D. (2008). Transcranial magnetic stimulation
over posterior parietal cortex disrupts trans-saccadic memory of multiple
objects. Journal of Neuroscience, 28(27), 6938-6949.
doi:10.1523/JNEUROSCI.0542-08.2008
Pujol, M., & Kopelman, D. (2003). Psychogenic amnesia. Practical Neurology, 3(5),
292-299.
Qiyuan, J., Richer, F., Wagoner, B.L., & Beatty, J. (1985). The pupil and stimulus
probability. Psychophysiology, 22, 204-207. doi:10.1111/j.14698986.1985.tb01645.x
Quian Quiroga, R., Mukamel, R., Isham, E.A., Malach, R., & Fried, I. (2008). Human
single-neuron responses at the threshold of conscious recognition.
Proceedings of the National Academy of Sciences USA, 105(9), 3599-3604.
doi:10.1073/pnas.0707043105
Quian Quiroga, R., & Kreiman, G. (2010). Postscript: About Grandmother cells and
Jennifer Aniston neurons. Psychological Reviews, 117(1), 297-299.
doi:10.1037/0033-295X.117.1.297
Rabinowitz, J.C., & Graesser, A.C.I. (1976). Word recognition as a function of
retrieval processes. Bulletin of the Psychonomic Society, 7, 75-77.
Rajaram, S. (1993). Remembering and knowing: Two means of access to the
personal past. Memory & Cognition, 21(1), 89-102.
Rajaram, S., & Roediger, H.L., III. (1997). Remembering and knowing as states of
consciousness during retrieval. In J.W. Schooler, & J.D. Cohen (Eds.),
Scientific approaches to consciousness. Carnegie Mellon symposia on
cognition (pp. 213-240). Mahwah, NJ: Erlbaum.
Rajkowski, J., Kubiak, P., & Aston-Jones, G. (1993). Correlations between locus
coeruleus neural activity, pupil diameter and behavior in monkey support a
role of LC in attention. Society for Neuroscience Abstracts, 19, 974.
Rajkowski, J., Kubiak, P., & Aston-Jones, G. (1994). Locus coeruleus activity in
monkey: Phasic and tonic changes are associated with altered vigilance.
Brain Research Bulletin, 35, 607-616.doi:10.1016/0361-9230(94)90175-9
319
Rajkowski, J., Lu, W., Zhu, Y., Cohen, J.D., & Aston-Jones, G. (2000). Prominent
projections from the anterior cingulate cortex to the locus coeruleus (LC) in
rhesus monkey. Society for Neuroscience Abstracts, 26, 2230.
Rajkowski, J., Majczynski, H., Clayton, E., & Aston-Jones, G. (2004). Activation of
monkey locus coeruleus neurons varies with difficulty and performance in a
target detection task. Journal of Neurophysiology, 92, 361-371.
doi:10.1152/jn.00673.2003.
Ratcliff, R. (1978). A theory of memory retrieval. Psychological Review, 85(2), 59108. doi:10.1037/0033-295x.85.2.59
Ratcliff, R., Clark, S.E., & Shiffrin, R.M. (1990). List-strength effect: I. Data and
discussion. Journal of Experimental Psychology: Learning, Memory, &
Cognition, 16(2), 163-178. doi:10.1037/0278-7393.16.2.163
Ratcliff, R., & Murdock Jr., B.B. (1976). Retrieval processes in recognition memory.
Psychological Review, 83(3), 190-214. doi:10.1037/0033-295X.83.3.190
Ratcliff, R., Sheu, C.F., & Gronlund, S.D. (1992). Testing global memory models
using ROC curves. Psychological Review, 99(3), 518-535. doi:10.1037/0033295X.99.3.518
Raymond, J.E., Shapiro, K.L., & Arnell, K.M. (1992). Temporary suppression of
visual processing in an RSVP task: An attentional blink? Journal of
Experimental Psychology: Human Perception & Performance, 18(3), 849-860.
doi:10.1037/0096-1523.18.3.849
Reber, A.S. (1967). Implicit learning of artificial grammars. Journal of Verbal
Learning & Verbal Behaviour, 6(6), 855-863. doi:10.1016/S00225371(67)80149-X
Reber, A.S. (1969). Transfer of syntactic structure in synthetic languages. Journal of
Experimental Psychology, 81(1), 115-119. doi:10.1037/h0027454
Reber, A.S. (1989). Implicit learning and tacit knowledge. Journal of Experimental
Psychology: General, 118(3), 219-235.
doi:10.1093/acprof:oso/9780195106589.001.0001
Reber, A.S., Walkenfeld, E., & Hernstadt, R. ( 1991). Implicit and explicit learning:
Individual differences and IQ. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 17(5), 888-896. doi:10.1037/0278-7393.17.5.888
Recarte, M.A., & Nunes, L.M. (2003). Mental workload while driving: Effects on
visual search, discrimination, and decision making. Journal of Experimental
Psychology/Applied, 9(2), 119-137. doi:10.1037/1076-898X.9.2.119
320
Recarte, M.A., Perez, E., Conchillo, A., & Nunes, L.M. (2008). Mental workload and
visual impairment: Differences between pupil, blink, and subjective rating.
Spanish Journal of Psychology, 11(2), 374-385.
Reder, L.M., Nhouyvanisvong, A., Schunn, C.D., Ayers, M.S., Angstadt, P., & Hiraki,
K. (2000). A mechanistic account of the mirror effect for word frequency: A
computational model of remember-know judgments in a continuous
recognition paradigm. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 26(2), 294-320. doi:10.1037/0278-7393.26.2.294
Reed, A.V. (1973). Speed-accuracy tradeoff in recognition memory. Science,
181(4099), 574-576. doi:10.1126/science.181.4099.574
Reeves, A.G., & Swenson, R.S. (2004). Cranial nerves: Eye movements. In
Disorders of the nervous system: A primer (Chapter 4). Dartmouth, NH:
Dartmouth Medical School.
Reynolds, G.D., & Richards, J.E. (2009). Cortical source localisation of infant
cognition. Developmental Neuropsychology, 34(3), 312-329. doi:
10.1080/87565640902801890
Richer, F., & Beatty, J. (1985). Pupillary dilations in movement preparation and
execution. Psychophysiology, 22(2), 204-207. doi:10.1111/j.14698986.1985.tb01587.x
Richer, F., Silverman, C., & Beatty, J. (1983). Response selection and initiation in
speeded reactions: A pupillometric analysis. Journal of Experimental
Psychology: Human Perception & Performance, 9(3), 360-370, 1983.
doi:10.1037/0096-1523.9.3.360.
Roediger, H.L., & Byrne, J.H. (Eds.) (2008). Learning and memory: A
comprehensive reference (Vol. 2). Oxford: Elsevier.
Roediger, H.L.I., & McDermott, K.B. (1995). Creating false memories: Remembering
words not presented in lists. Journal of Experimental Psychology:
Learning,Memory & Cognition, 21, 803–814.
Roediger, H.L., Weldon, M.S., Stadler, M.L., & Riegler, G.L. (1992). Direct
comparison of two implicit memory tests: Word fragment and word stem
completion. Journal of Experimental Psychology: Learning, Memory, &
Cognition, 18(6), 1251-1269. doi:10.1037/0278-7393.18.6.1251
Rogers, R., & Bender, S.D. (2003) Evaluation of malingering and deception. In I.B.
Weiner, & A.M. Goldstein (Eds.), Handbook of psychology: Vol. 11. Forensic
psychology (pp. 109-129). New York: John Wiley & Sons, Inc.
321
Rohrbaugh, J.W., Syndulko, K., & Lindsley, D.B. (1976). Brain wave components of
the contingent negative variation in humans. Science, 191(4231), 1055-1057.
doi:10.1126/science.1251217
Rosenfeld, J.P., Ellwanger, J., & Sweet, J. (1995). Detecting simulated amnesia with
event-related brain potentials. International Journal of Psychophysiology,
19(1), 1-11. doi:10.1016/0167-8760(94)00057-L
Rosenfeld, J.P., Labkovsky, E., Winograd, M., Lui, M.A., Vandenboom, C., &
Chedid, E. (2008). The Complex Trial Protocol (CTP): A new,
countermeasure-resistant, accurate, P300-based method for detection of
concealed information. Psychophysiology, 45, 906-919. doi:10.1111/j.14698986.2008.00708.x
Rosenfeld, J.P., Soskins, M., Bosh, G., & Ryan, A., (2004). Simple, effective
countermeasures to P300-based tests of detection of concealed information.
Psychophysiology, 41(2), 205-219. doi:10.1111/j.1469-8986.2004.00158.x
Ross, S.R., Millis, S.R., Krukowski, R.A., Putnam, S.H., & Adams, K.M. (2004).
Detecting incomplete effort on the MMPI-2: An examination of the Fake Bad
Scale in mild head injury. Journal of Clinical & Experimental
Neuropsychology, 26(1), 115-124. doi:10.1076/jcen.26.1.115.23933
Rotello, C.M., & Heit, E. (1999). Two-process models of recognition memory:
Evidence for recall-to-reject? Journal of Memory & Language, 40(3), 432-453.
doi:10.1006/jmla.1998.2623
Rotello, C.M., & Heit, E. (2000). Associative recognition: A case of recall-to-reject
processing. Memory & Cognition, 28(6), 907-922. doi:10.3758/BF03209339
Rotello, C.M., & Macmillan, N.A. (2006). Remember-know models as decision
strategies in two experimental paradigms. Journal of Memory & Language,
55, 479-494. doi:10.1016/j.jml.2006.08.002
Rotello, C.M., Macmillan, N.A., & Reeder, J.A. (2004). Sum-difference theory of
remembering and knowing: A two-dimensional signal-detection model.
Psychological Review, 111(3), 588-616. doi:10.1037/0033-295X.111.3.588
Rotello, C.M., Macmillan, N.A., & Van Tassel, G. (2000). Recall-to-reject in
recognition: Evidence from ROC curves. Journal of Memory & Language,
43(1), 67-88. doi:10.1006/jmla.1999.2701
Roubinovitch, J. (1900). Des variations du diamétre pupillaire en rapport avec l'effort
intellectuel. In P. Janet (Ed.), Quantriéme congres internationale de
322
psychologie: Compte rendu des seances et texts des memoires, publies par
les soins du Doctuer Pierre Janet. Paris: F Alcan.
Rowe, D.W., Sibert, J., & Irwin, D. (1998). Heart rate variability: Indicator of user
state as an aid to human-computer interaction. In Proceedings from The
SIGCHI Conference on Human Factors in Computing Systems, 480-487.
New York, NY: ACM Press. doi:10.1145/274644.274709
Rowland, L.M., Thomas, M.L., Thorne, D.R., Sing, H.C., Krichmar, J.L., Davis, H.Q.,
… Belenky, G. (2005). Oculomotor responses during partial and total sleep
deprivation. Aviation, Space, & Environmental Medicine, 76(7S), C104-C113.
Rugg, M.D. (1995). ERP studies of memory. In M.D. Rugg, & M.G.H. Coles (Eds.),
Electrophysiology of mind (pp. 132-170). New York: Oxford University Press.
Rugg, M.D., & Allan, K. (1999) Memory retrieval: An electrophysiological
perspective. In M.S. Gazzaniga (Ed.) The cognitive neurosciences (pp.805816). USA: MIT Press.
Rugg, M.D., & Allan, K. (2000). Event-related potential studies of memory. In E.
Tulving, & F.I.M. Craik (Eds.), Oxford handbook of memory (pp. 521-537).
New York: Oxford University Press.
Rugg, M.D., Cox, C.J.C., Doyle, M.C., & Wells, T. (1995). Event-related potentials
and the recollection of low and high frequency words. Neuropsychologia,
33(4), 471-484. doi:10.1016/0028-3932(94)00132-9
Rugg, M.D., & Doyle, M.C. (1992). Event-related potentials and recognition memory
for low- and high-frequency words. Journal of Cognitive Neuroscience, 4(1),
69-79. doi:10.1162/jocn.1992.4.1.69
Rugg, M.D., Mark, R.E., Walla, P., Schloerscheidt, A.M., Birch, C.S., & Allan, K.
(1998a). Dissociation of the neural correlates of implicit and explicit memory.
Nature, 392(6676), 595-598. doi:10.1038/33396
Rugg, M.D., & Nagy, M.E. (1989). Event-related potentials and recognition memory
for words. Electroencephalography & Clinical Neurophysiology, 72(5), 395406. doi:10.1016/0013-4694(89)90045-X
Rugg, M.D., Pearl, S., Walker, P., Roberts, R.C., & Holdstock, J.S. (1994). Word
repetition effects on event-related potentials in healthy young and old
subjects, and in patients with Alzheimer-type dementia. Neuropsychologia,
32, 381-398. doi:10.1016/0028-3932(94)90085-X
Rugg, M.D., Roberts, R.C., Potter, D.D., Pickles, C.D., & Nagy, M.E. (1991). Eventrelated potentials related to recognition memory. Effects of unilateral temporal
323
lobectomy and temporal lobe epilepsy. Brain, 114(5), 2313-2332.
doi:10.1093/brain/114.5.2313
Rugg, M.D., Schloerscheidt, A.M., & Mark, R.E. (1998b). An electrophysiological
comparison of two indices of recollection. Journal of Memory & Language,
39(1), 47-69. doi:10.1006/jmla.1997.2555
Rugg, M.D., & Yonelinas, A.P. (2003). Human recognition memory: A cognitive
neuroscience perspective. Trends in Cognitive Sciences, 7(7), 313-319.
doi:10.1016/S1364-6613(03)00131-1
Rüsseler, J., Brett, A., Klaue, U., Sailer, M., & Münte, T.F. (2008). The effect of
coaching on the simulated malingering of memory-impairment. BioMed
Central Neurology, 8(37). doi:10.1186/1471-2377-8-37
Rutishauser, U. (2008). Learning and representation of declarative memories by
single neurons in the human brain. (Unpublished doctoral dissertation).
California Institute of Technology.
Rutishauser, U., Schuman, E.M., & Mamelak, A.N. (2008). Activity of human
hippocampal and amygdala neurons during retrieval of declarative memories.
Proceedings for the National Academy of Sciences, 105(1), 329-334.
doi:10.1073/pnas.0706015105
Sahraie, A., & Barbur, J.L. (1997). Pupil response triggered by the onset of coherent
motion. Graefe's Archive for Clinical and Experimental Ophthalmology,
235(8), 494-500. doi:10.1007/BF00947006.
Samuel, R.Z., & Mittenberg, M.D. (2005). Determination of malingering in disability
evaluations. Primary Psychiatry, 12(12), 60-68.
Samuels, E.R., & Szabadi, E. (2008a). Functional neuroanatomy of the
noradrenergic locus coeruleus: Its roles in the regulation of arousal and
autonomic function part I: Principles of functional organisation. Current
Neuropharmacology, 6(3), 235-253. doi:10.1006/jmla.1997.2555
Samuels, E.R., & Szabadi, E. (2008b). Functional neuroanatomy of the
noradrenergic locus coeruleus: Its roles in the regulation of arousal and
autonomic function part II: Physiological and pharmacological manipulations
and pathological alterations of locus coeruleus activity in humans. Current
Neuropharmacology, 6(3), 254-285. doi:10.2174/157015908785777193
Sanquist, T.F., Rohrbaugh, J.W., Syndulko, K., & Lindsley, D.B. (1980).
Electrocortical signs of levels of processing: Perceptual analysis and
324
recognition memory. Psychophysiology, 17(6), 568-576. doi:10.1111/j.14698986.1980.tb02299.x
Sara, S.J. (2009). The locus coeruleus and noradrenergic modulation of cognition.
Nature Reviews Neuroscience, 10(3), 211-223. doi:10.1038/nrn2573
Schacter, D.L. (1977). EEG theta waves and psychological phenomena: A review
and analysis. Biological Psychology, 5, 47-82. doi:10.1016/03010511(77)90028-X
Schacter, D.L., & Buckner, R.L. (1998). On the relations among priming, conscious
recollection, and intentional retrieval: evidence from neuroimaging research.
Neurobiology of Learning and Memory, 70(1-2), 284-303.
doi:10.1006/nlme.1998.3854
Schacter, D.L., McAndrews, M.P., & Moscovitch, M. (1988). Access to
consciousness: Dissociations between implicit and explicit knowledge in
neuropsychological syndromes. In L. Weiskrantz (Ed.), Thought without
language (pp. 242–278). Oxford: Oxford University Press.
Schaefer Jr., T., Ferguson, J.B., Klein, J.A., & Rawson, E.B. (1968). Pupillary
responses during mental activities. Psychonomic Science, 12(4), 137-138.
Schagen, S., Schmand, B., de Sterke, S., & Lindeboom, J. (1997). Amsterdam
Short-Term Memory Test: A new procedure for the detection of feigned
memory deficits. Journal of Clinical & Experimental Neuropsychology, 19(1),
43-51. doi:10.1080/01688639708403835
Schenk, K., & Sullivan, K.A. (2010). Do warnings deter rather than produce more
sophisticated malingering? Journal of Clinical & Experimental
Neuropsychology, 32(7), 752-762. doi:10.1080/13803390903512678
Schiff, M. (1875). La pupille consideré comme ethésiomètre. Paris: J.B. Bailliére.
Schiff, M., & Foa, P. (1874). La pupille considerée comme esthésiomètre (R.G. de
Choisity, Trans.). Marseille Médical, 2, 736-741.
Schlemmer, K.B., Kulke, F., Kuchinke, L., & Van Der Meer, E. (2005). Absolute pitch
and pupillary response: Effects of timbre and key color. Psychophysiology,
42(4), 465-472. doi:10.1111/j.1469-8986.2005.00306.x.
Schloerscheidt, A.M., & Rugg, M.D., (2004). The impact of change in stimulus format
on the electrophysiological indices of recognition. Neuropsychologia, 42, 451466. doi:10.1016/j.neuropsychologia.2003.08.010
Schluroff, M. (1982). Pupil responses to grammatical complexity of sentences. Brain
& Language, 17(1), 133-145. doi:10.1016/0093-934X(82)90010-4
325
Schluroff, M., Zimmermann, T.E., Freeman Jr., R.B., Hofmeister, K., Lorscheid, T., &
Weber, A. (1986). Pupillary responses to syntactic ambiguity of sentences.
Brain & Language, 27(2), 322-344. doi:10.1016/0093-934X(86)90023-4
Schulman, A.I. (1967). Word length and rarity in recognition memory. Psychonomic
Science, 9(4), 211-212.
Schultheis, H., & Jameson, A. (2004). Assessing cognitive load in adaptive
hypermedia systems: Physiological and behavioral methods. In P. De Bra, &
W. Nejdl (Eds.), Adaptive hypermedia and adaptive web-based systems (pp.
225-234). Berlin: Springer Verlag.
Schweitzer, N.M.J. (1956). Threshold measurements of the pupil the dark adapted
eye. Documenta Ophthalmologica, 10(1), 1-78. doi:10.1007/BF00172098
Scott, R.B., & Dienes, Z. (2008). The conscious, the unconscious, and familiarity.
Journal of Experimental Psychology: Learning, Memory, & Cognition, 34(5),
1264-1288. doi:10.1037/a0012943
Scott, R.B., & Dienes, Z. (2009). Knowledge applied to new domains: The
unconscious succeeds where the conscious fails. Consciousness &
Cognition, 19(1), 391-398. doi:10.1016/j.concog.2009.11.009
Scott, T.R., Wells, W.H., Wood, D.Z., & Morgan, D.I. (1967). Pupillary response and
sexual interest reexamined. Journal of Clinical Psychology, 23(4), 433-438.
doi:10.1002/1097-4679(196710)23:4<433::AID-JCLP2270230408>3.0.CO;22
Shanks, D.R., & St.John, M.F. (1994). Characteristics of dissociable human learning
systems. Behavioral & Brain Sciences, 17(3), 367-448.
doi:10.1017/S0140525X00035032
Shepard, R.N. (1967). Recognition memory for words, sentences, and pictures.
Journal of Verbal Learning & Verbal Behavior, 6(1), 156-163.
doi:10.1016/S0022-5371(67)80067-7
Shiffrin, R.M., & Steyvers, M. (1997). A model of recognition memory: REM –
Retrieving effectively from memory. Psychological Bulletin & Review, 4(2),
145-166.
Shiffrin, R.M., & Steyvers, M. (1998). The effectiveness of retrieval from memory. In
M. Oaksford & N. Chater (Eds.), Rational models of cognition (pp. 73-95).
London: Oxford University Press.
Shore, D.I., & Klein, R.M. (2000). On the manifestations of memory in visual search.
Spatial Vision, 14(1), 59-75. doi:10.1163/156856801741369
326
Siegle, G.J., Granholm, E., Ingram, R.E., & Matt, G.E. (2001). Pupillary and reaction
time measures of sustained processing of negative information in depression.
Biological Psychiatry, 49(7), 624-636.
Siegle, G.J., Ichikawa, N., & Steinhauer, S. (2008). Blink before and after you think:
blinks occur prior to and following cognitive load indexed by pupillary
responses. Psychophysiology, 45(5), 679-687. doi:10.1111/j.14698986.2008.00681.x
Siegle, G.J., Steinhauer, S.R., Carter, C.S., Ramel, W., & Thase, M.E. (2003a). Do
the seconds turn into hours? Relationships between sustained pupil dilation in
response to emotional information and self-reported rumination. Cognitive
Therapy & Research, 27(3), 365-382. doi:10.1023/A:102397 4602357
Siegle, G.J., Steinhauer, S.R., Carter, C.S., & Thase, M.E. (submitted). Is sustained
processing specific to emotional information in depression? Evidence from
pupil dilation.
Siegle, G.J., Steinhauer, S.R., Stenger, V.A., Konecky, R., & Carter, C.S. (2003b).
Use of concurrent pupil dilation assessment to inform interpretation and
analysis of fMRI data. Neuroimage, 20, 114-124. doi:10.1016/S10538119(03)00298-2
Siegle, G.J., Steinhauer, S.R., & Thase, M.E. (2004). Pupillary assessment and
computational modeling of the stroop task in depression. International Journal
Psychophysiology, 52, 63-76. doi:10.1016/j.ijpsycho.2003.12.010
Silk, J.S., Dahl, R.E., Ryan, N.D., Forbes, E.E., Axelson, D.A., Birmaher, B., &
Siegle, G.J. (2007). Pupillary reactivity to emotional information in child and
adolescent depression: Links to clinical and ecological measures. American
Journal of Psychiatry, 164(12), 1873–1880. doi:10.1176/appi.ajp.2007.
06111816
Silk, J.S., Siegle, G.J., Whalen, D.J., Ostapenko, L.J., Ladouceur, C.D., & Dahl, R.E.
(2009). Pubertal changes in emotional information processing: pupillary,
behavioral, and subjective evidence during emotional word identification.
Developmental Psychopathology, 21(1), 7-26.
doi:10.1017/S0954579409000029
Simpson, H.M. (1969). Effects of a task-relevant response on pupil-size.
Psychophysiology, 6(2), 115-121. doi:10.1111/j.1469-8986.1969.tb02890.x
327
Simpson, H.M., & Climan, M.H. (1971). Pupillary and electromyographic changes
during an imagery task. Psychophysiology, 8(4), 483-490. doi:10.1111/j.14698986.1971.tb00480.x
Simpson, H.M., & Hale, S.M. (1969). Pupillary changes during a decision-making
task. Perceptual & Motor Skills, 29(2), 495-498.
doi:10.2466/pms.1969.29.2.495
Simpson, H.M., & Molloy, F.M. (1971). Effects of audience anxiety on pupil-size.
Psychophysiology, 8, 491-496. doi:10.1111/j.1469-8986.1971.tb00481.x
Simpson, H.M., Molloy, F.M., Hale, S.M., & Climan, M.H. (1968). Latency and
magnitude of the pupillary response during an imagery task. Psychonomic
Science, 13, 293-294.
Simpson, H.M., & Paivio, A. (1966). Changes in pupil-size during an imagery task
without motor response involvement. Psychonomic Science, 5, 405-406.
Simpson, H.M., & Paivio, A. (1968). Effects on pupil-size of manual and verbal
indicators of cognitive task fulfillment. Perception & Psychophysics, 3(3A),
185-190. doi:10.3758/BF03212726
Skoglund, C.R. (1943). On the influence of alcohol on the pupillary light reflex in
man. Acta Physiologica Scandinavica, 6(1), 94-96. doi:10.1111/j.17481716.1943.tb02831.x
Slick, D., Hopp, G., Strauss, E., & Thompson, G.B. (1997). VSVT: Victoria Symptom
Validity Test (Version 1.0). Odessa, Florida: Psychological Assessment
Resources.
Slick, D.J., Sherman, E.M.S., & Iverson, G.L. (1999). Diagnostic criteria for
malingered neurocognitive dysfunction: Proposed standards for clinical
practice and research. The Clinical Neuropsychologist, 13(4), 545-561.
doi:10.1076/1385-4046(199911)13:04;1-Y;FT545
Slooter, J., & van Norren, D. (1980). Visual acuity measured with pupil responses to
checkerboard stimuli. Investigative Ophthalmology & Visual Science, 19(1),
105-108.
Smith, G.P., & Burger, G.K. (1997). Detection of malingering: Validation of the
Structured Inventory of Malingered Symptomatology. Journal of the American
Academy of Psychiatry & the Law, 25(2), 183-189.
Smith, M.E. (1993). Neurophysiological manifestations of recollective experience
during recognition memory judgments. Journal of Cognitive Neuroscience,
5(1), 1-13. doi:10.1162/jocn.1993.5.1.1
328
Smith, M.E., & Guster, K. (1993). Decomposition of recognition memory eventrelated potentials yields target, repetition, and retrieval effects.
Electroencephalography & Clinical Neurophysiology, 86(5), 335-343.
doi:10.1016/0013-4694(93)90046-X
Smith, M.E., & Halgren, E. (1989). Dissociation of recognition memory components
following temporal lobe lesions. Journal of Experimental Psychology:
Learning, Memory, & Cognition, 15(1), 50-60. doi:10.1037/0278-7393.15.1.50
Smith, M.E., Halgren, E., Sokolik, M., Baudena, P., Musolino, A., Liegois-Chauvel,
C., & Chauvel, P. (1990). The intracranial topography of the P3 event-related
potential elicited during auditory oddball. Electroencephalography & Clinical
Neurophysiology, 76(3), 235-248. doi:10.1016/0013-4694(90)90018-F
Smith, M.E., Stapleton, J.M., & Halgren, E. (1986). Human medial temporal lobe
potentials evoked in memory and language tasks. Electroencephalography &
Clinical Neurophysiology, 63(2), 145-159. doi:10.1016/0013-4694(86)90008-8
Soderlund, H., Parker, E.S., Schwartz, B.L., & Tulving, E. (2005). Memory encoding
and retrieval on the ascending and descending limbs of the blood alcohol
concentration curve. Psychopharmacology, 182(2), 305-317.
doi:10.1007/s00213-005-0096-2
Sokolov, E.N. (1963). Perception and the conditioned reflex. Oxford: Pergamon
Press.
Sokolski, K.N., Nguyen, B.D., & DeMet, E.M. (2000). Decreases in dilated pupil-size
in depressed patients with age may reflect adrenergic changes. Psychiatry
Research, 94(3), 267-272.
Spector, R.H. (1990). The pupils. In H.K. Walker, W.D. Hall, & J.W. Hurst (Eds.),
Clinical methods: The history, physical, and laboratory examinations (3rd ed.,
pp. 300-304). Boston: Butterworths.
Spence, S.A., Farrow, T.F.D., Herford, A.E., Wilkinson, I.D., Zheng, Y., & Woodruff,
P.W.R. (2001). Behavioural and functional anatomical correlates of deception
in humans. NeuroReport, 12(13), 2849-2853.
Spencer, K.M., Vila Abad, E., & Donchin, E. (2000). On the search for the
neurophysiological manifestation of recollective experience.
Psychophysiology, 37(4), 494-506. doi:10.1111/1469-8986.3740494
Squire, L.R., Wixted, J.T., & Clark, R.E. (2007). Recognition memory and the medial
temporal lobe: A new perspective. Nature, 8(11), 872-883.
doi:10.1038/nrn2154
329
Srinivasan, R., Tucker, D.M., & Murias, M. (1998). Estimating the spatial nyquist of
the human EEG. Behavioral Research Methods, 30(1), 8-19.
doi:10.3758/BF03209412
Staal, M.A. (2004). Stress, cognition, and human performance: A literature review
and conceptual framework (Tech. Memorandum 212824). Moffett Field, CA:
National Aeronautics and Space Administration Ames Research Center.
Stanners, R.F., Headley, D.B., & Clark, W.R. (1972). The pupillary response to
sentences: Influences of listening set and deep structure. Journal of Verbal
Learning & Verbal Behavior, 11(2), 257-263. doi:10.1016/S00225371(72)80086-0
Stark, C.E., & Squire, L.R. (2003). Hippocampal damage equally impairs memory for
single items and memory for conjunctions. Hippocampus, 13(2), 281-292.
doi:10.1002/hipo.10085
Stark, C.E.L., Bayley, P.J., & Squire, L.R. (2002). Recognition memory for single
items and for associations is similarly impaired following damage to the
hippocampal region. Learning & Memory, 9(5), 238-242.
doi:10.1101/lm.51802
Steinhauer, S.R. (1982). Emitted and evoked pupillary responses and event-related
potentials as a function of reward and task involvement. (Unpublished
doctoral dissertation). City University of New York, NY, USA.
Steinhauer, S.R. (2002). Pupillary responses, cognitive psychophysiology and
psychopathology. Retrieved 10 June 2007 from
http://www.wpic.pitt.edu/research/biometrics/Publications/Biometrics%20Archi
ves%20PDF/874-2002%20Steinhauer%20PUPILWEB.pdf
Steinhauer, S.R., Condray, R., & Kasparek, A. (2000). Cognitive modulation of
midbrain function: Task-induced reduction of the pupillary light reflex.
International Journal of Psychophysiology, 39(1), 21-30. doi:10.1016/S01678760(00)00119-7
Steinhauer, S.R., & Hakerem, G. (1992). The pupillary response in cognitive
psychophysiology and schizophrenia. In D. Friedman, & G. Bruder (Eds.),
Psychophysiology and experimental psychopathology: A tribute to Samuel
Sutton (pp. 182-204). New York, NY: Academy of Sciences.
Steinhauer, S.R., Hakerem, G., & Spring, B.J. (1979). The pupillary response as a
potential indicator of vulnerability to schizophrenia. Psychopharmacology
Bulletin, 15(1), 44-45.
330
Steinhauer, S.R., Siegle, G.J., Condray, R., & Pless, M. (2004). Sympathetic and
parasympathetic innervation of pupillary dilation during sustained processing.
International Journal of Psychophysiology, 52(1), 77-86.
doi:10.1016/j.ijpsycho.2003.12.005
Stelmack, R.M., & Leckett, W.J. (1974). Effect of artificial pupil-size on recognition
threshold. Perceptual & Motor Skills, 39(2), 739-742.
doi:10.2466/pms.1974.39.2.739
Stelmack, R.M., & Mandelzys, N. (1975). Extraversion and pupillary response to
affective and taboo words. Psychophysiology, 12(5), 536-540.
doi:10.1111/j.1469-8986.1975.tb00042.x
Steptoe, W., Steed, A., Rovira, A., & Rae, J. (2010). Lie tracking: Social presence,
truth and deception in avatar-mediated telecommunication. In Proceedings
from The Conference on Human Factors in Computing Systems, 2, 10391048. Atlanta, Georgia: ACM. doi:10.1145/1753326.1753481
Stern, J.A., Walrath, L.C., & Goldstein, R. (1984). The endogenous eyeblink.
Psychophysiology, 21(1), 22-33. doi:10.1111/j.1469-8986.1984.tb02312.x
Stern, P.C. (2003). The polygraph and lie detection. In Report of the National
Research Council Committee to Review the Scientific Evidence on the
Polygraph (pp. 11-28). Washington, DC: The National Academies Press.
Sternback, R.A. (1966). Principles of psychophysiology. New York: Academic Press.
Stevens, W.D., Wig, G.S., & Schacter, D.L. (2008). Implicit memory and priming. In
H.L. Roediger III, & J.H. Byrne (Eds.), Learning and memory, volume 2:
Cognitive psychology of memory (pp. 623-644). London: Elselvier.
Stone, G.M., & Rothenheber, E. (1992). Advanced scientific detection of deception ERP augmented polygraphy. In Proceedings from The International
Carnahan Conference on Security Technology: Crime Countermeasures (pp
72-73). Atlanta, GA: Institute of Electrical and Electronics Engineers.
doi:10.1109/CCST.1992.253758
Sullivan, K.N., Manfredi, F., & Behnke, R.H. (1968). Hippus in Cheyne-Stokes
respiration: Observations in three patients with rhythmic respiratory and
pupillary changes. Archives of Internal Medicine, 122(2), 116-121.
Suzuki, M., Johnson, J.D., & Rugg, M.D. (2011). Decrements in hippocampal activity
with item repetition during continuous recognition: An fMRI study. Journal of
Cognitive Neuroscience, 23(6), 1522-1532. doi:10.1162/jocn.2010.21535
331
Swihart, A.A, Harris, K.M., & Hatcher, L.L. (2008). Inability of the Rarely Missed
Index to identify simulated malingering under more realistic assessment
conditions. Journal of Clinical & Experimental Neuropsychology, 30(1), 120126. doi:10.1080/13803390701249044
Szabadi, E., & Bradshaw, C.M. (1996). Autonomic pharmacology of α2adrenoceptors. Journal of Psychopharmacology, 10(3), 6-18.
Tardif, H.P., Barry, R.J., Fox, A.M., & Johnstone, S.J. (2000). Detection of feigned
recognition memory-impairment using the old/new effect of the event-related
potential. International Journal of Psychophysiology, 36(1), 1-9.
doi:10.1016/S0167-8760(00)00083-0
Thackray, R.I. (1969) Patterns of physiological activity accompanying performance
in a perceptual motor task (Federal Aviation Administration Report. No. 69-8).
Washington DC, USA: Office of Aviation Medicine.
Tipper, S.P. (1985). The negative priming effect: Inhibitory priming by ignored
objects. Quarterly Journal of Experimental Psychology, 37(4), 571-590.
doi:10.1080/14640748508400920
Tombaugh, T.N. (1996). Test of Memory and Malingering. North Tonawanda, NY:
Multi-Health Systems.
Trott, C.T., Friedman, D., Ritter,W., & Fabiani, M. (1997). Item and source memory:
Differential age effects revealed by event-related potentials. NeuroReport,
8(15), 3373-3378.
Trott, C.T., Friedman, D., Ritter, W., Fabiani, M., & Snodgrass, J.G. (1999). Episodic
priming and memory for temporal source: Event-related potentials reveal agerelated differences in prefrontal functioning. Psychology & Aging, 14(3), 390413. doi:10.1037/0882-7974.14.3.390
Tryon, W.W. (1975). Pupillometry: A survey of sources of variation.
Psychophysiology, 12(1), 90-93. doi:10.1111/j.1469-8986.1975.tb03068.x
Tsivilis, D., Otten, L.J., & Rugg, M.D. (2001). Context effects on the neural correlates
of recognition memory: An electrophysiological study. Neuron, 31(3), 497505. doi:10.1016/S0896-6273(01)00376-2
Tucker, D.M. (1993). Spatial sampling of head electrical fields: The geodesic sensor
net. Electroencephalography & Clinical Neurophysiology, 87(3), 154-163.
doi:10.1016/0013-4694(93)90121-B
332
Tucker, D.M., Liotti, M., Potts, G.F., Russell, G.S., & Posner, M.I. (1994).
Spatiotemporal analysis of brain electrical fields. Human Brain Mapping, 1(2),
134-152. doi:10.1002/hbm.460010206
Tugcu, H. (2010). Forensic medicine and malingering. Turkiye Klinikleri Journal of
Medical Sciences, 30(4S), S86-S86.
Tulving, E. (1983). Elements of episodic memory. New York, USA: Oxford University
Press.
Tulving, E. (1985a). Memory and consciousness. Canadian Psychologist, 26, 1-12.
Tulving, E. (1985b). How many memory systems are there? American Psychologist,
40(4), 385-398. doi:10.1037/0003-066X.40.4.385
Tulving, E., & Kroll, N. (1995). Novelty assessment in the brain and long-term
memory encoding. Psychonomic Bulletin & Review, 2(3), 387-390.
doi:10.3758/BF03210977
Tulving, E., & Schacter, D.L. (1990). Priming and human memory systems. Science,
247(4940), 301-306. doi:10.1126/science.2296719
Turriziani, P., Oliveri, M., Salerno, S., Costanzo, F., Koch, G., Caltagirone, C., &
Carlesimo, G.A. (2008). Recognition memory and prefrontal cortex:
Dissociating recollection and familiarity processes using rTMS. Behavioural
Neurology, 19(1-2), 23-27.
Ucar, M., & Atac, A. (2010). The term of malingering in terms of medical ethics: The
ethical dimension of patient's attitudes and behaviours intended for
misleading physician. Turkiye Klinikleri Journal of Medical Sciences, 30(4S),
S1-S6.
Uhl, F., Franzen, P., Serles, W., Lang, W., Lindinger, G., & Deecke, L. (1990)
Anterior frontal cortex and the effect of proactive interference in paired
associate learning. A DC-potential study. Journal of Cognitive Neuroscience,
2(4), 373-382. doi:10.1162/jocn.1990.2.4.373
Ukai, K. (1985). Spatial pattern as a stimulus to the pupillary system. Journal of the
Optical Society of America, 2(7), 1094-1099. doi:10.1364/JOSAA.2.001094
Urbach, T.P., & Kutas, M. (2002). The intractability of scaling scalp distribution to
infer neuroelectric sources. Psychophysiology, 39(6), 791-808.
doi:10.1017/S0048577202010648
Urry, H.L., van Reekum, C.M., Johnstone, T., Kalin, N.H., Thurow, M.E., Schaefer,
H.S., . . . Davidson, R.J. (2006). Amygdala and ventromedial prefrontal cortex
are inversely coupled during regulation of negative affect and predict the
333
diurnal pattern of cortisol secretion among older adults. Journal of
Neuroscience, 26(16), 4415-4425. doi:10.1523/JNEUROSCI. 3215-05.2006
Usher, M., Cohen, J.D., Servan-Schreiber, D., Rajkowski, J., & Aston-Jones, G.
(1999). The role of locus coeruleus in the regulation of cognitive performance.
Science, 283(5401), 549-554. doi:10.1126/science.283. 5401.549
Vagnini, V.L., Berry, D.T.R., Clark, J.A., & Jiang, Y. (2008). New measures to detect
malingered neurocognitive deficit: Applying reaction time and event-related
potentials. Journal of Clinical & Experimental Neuropsychology, 30(7), 766776. doi:10.1080/13803390701754746
Van Droof, L.V., Ledoux, K., Pickett, E.J., Buz, E., Billings, N., & Gordon, B. (2010).
Advancing the measurement of receptive language in nonverbal individuals
with autism. Presented at The 2010 International Meeting for Autism
Research, Philadelphia, PA.
van Hooff, J.C., Brunia, C.H.M., & Allen, J.J.B. (1996). Event-related potentials as
indirect measures memory. International Journal of Psychophysiology, 21(1),
15-31. doi:10.1016/0167-8760(95)00043-7
van Hooff, J.C., Sargeant, E., Foster, J.K., & Schmand, B.A. (2009). Identifying
deliberate attempts to fake memory-impairment through the combined use of
reaction time and event-related potential measures. International Journal of
Psychophysiology, 73(3), 246-256. doi:10.1016/j.ijpsycho.2009.04.002
Van Orden, K.F., Limbert, W., Makeig, S., & Jung, T.P. (2001). Eye activity
correlates of workload during a visuospatial memory task. Human Factors:
The Journal of the Human Factors & Ergonomics Society, 43(1), 111-121.
doi:10.1518/001872001775992570
Van Petten, C., & Kutas, M. (1990). Interactions between sentence context and word
frequency in event-related brain potentials. Memory & Cognition, 18(4), 380393. doi:10.3758/BF03197127
Van Petten, C., & Kutas, M. (1991). Influences of semantic and syntactic context on
open- and closed-class words. Memory & Cognition, 19(1), 95-112.
doi:10.3758/BF03198500
Van Petten, C., Kutas, M., Kluender, R., Mitchiner, M., & McIsaac, H. (1991).
Fractionating the word repetition effect with event related potentials. Journal
of Cognitive Neuroscience, 3(2), 131-150. doi:10.1162/jocn.1991.3.2.131
Van Rijn, H., Dalenberg, J., Borst, J., & Sprenger, S.A. (submitted). Memory strength
correlates negatively with pupil dilation in a delayed paired-associate task.
334
Van Turennout, M.V., Ellmore, T., & Martin, A. (2000). Long-lasting cortical plasticity
in the object naming system. Nature Neuroscience, 3(12), 1329-1334.
doi:10.1038/81873
Verfaellie, M., Bauer, R.M., & Bowers, D. (1991). Autonomic and behavioral
evidence of “implicit” memory in amnesia. Brain & Cognition, 15(1), 10-25.
doi:10.1016/0278-2626(91)90012-W
Verleger, R. (1995). Memory-related EEG potentials: Slow negativities, priming
positivity, recognition positivity, and Dm: Commentary on Klimesch on
Memory-Brain. Psycoloquy, 6(27), 3.
Verney, S.P., Granholm, E., Marshall, S.P., Malcarne, V.L., & Saccuzzo, D.P.
(2005). Culture-fair cognitive ability assessment: Information processing and
psychophysiological approaches. Assessment, 12(3), 303-319.
doi:10.1177/1073191105276674
Vilberg, K.L., & Rugg, M.D. (2008). Memory retrieval and the parietal cortex: A
review of evidence from a dual-process perspective. Neuropsychologia,
46(7), 1787-1799. doi:10.1016/j.neuropsychologia.2008.01.004
Võ, M.L.H., Jacobs, A.M., Kuchinke, L., Hofmann, M., Conrad, M., Schacht, A., &
Hutzler, F. (2008). The coupling of emotion and cognition in the eye:
Introducing the pupil old/new effect. Psychophysiology, 45(1), 130-140.
doi:10.1111/j.1469-8986.2007.00606.x
Voss, J.L., & Paller, K.A. (2008). Brain substrates of implicit and explicit memory:
The importance of concurrently acquired neural signals of both memory
types. Neuropsychologia, 46, 3021-3029.
doi:10.1016/j.neuropsychologia.2008.07.010
Wagner, A.D., Desmond, J.E., Demb, J.B., Glover, G.H., & Gabrieli, J.D.E. (1997).
Semantic repetition priming for verbal and pictorial knowledge: A functional
MRI study of left inferior prefrontal cortex. Journal of Cognitive Neuroscience,
9(6), 714-726. doi:10.1162/jocn.1997.9.6.714
Wagner, A.D., Maril, A., & Schacter, D.L. (2000). Interactions between forms of
memory: When priming hinders new learning. Journal of Cognitive
Neuroscience, 12(S2), 52-60. doi:10.1162/089892900564064
Wais, P.E., Wixted, J.T., Hopkins, R.O., & Squire, L.R. (2006). The hippocampus
supports both the recollection and the familiarity components of recognition
memory. Neuron, 49(3), 459-466. doi:10.1016/j.neuron.2005.12.020
335
Walzl, M., Hagen, R., & Prummer, K. (2007). Pupillometric examinations for
sleepiness on professional drivers. Pupillometrische untersuchungen auf
schläfrigkeit bei berufskraftfahrern, 57(11), 349-364.
Wang, J.T. (2010). Pupil dilation and eye tracking. In M. Schulte-Mecklenbeck, A.
Kuehberger, & R. Ranyard (Eds.), Handbook of process-tracing methods (pp.
185-204). Hove: Psychology Press.
Wang, J.T., Spezio, M., & Camerer, C.F. (2010). Pinocchio's pupil: Using
eyetracking and pupil dilation to understand truth telling and deception in
sender-receiver games. American Economic Review, 100(3), 984-1007.
doi:10.1257/aer.100.3.984
Warren, L.R. (1980). Evoked-potential correlates of recognition memory. Biological
Psychology, 11(1), 21-35. doi:10.1016/0301-0511(80)90023-X
Wasserman, S., & Bockenholt, U. (1989). Bootstrapping: Applications to
psychophysiology. Psychophysiology, 26(2), 208-221. doi:10.1111/j.14698986.1989.tb03159.x
Webb, A.K., Honts, C.R., Kircher, J.C., Bernhardt, P., & Cook, A.E. (2009).
Effectiveness of pupil diameter in a probable-lie comparison question test for
deception. Legal & Criminological Psychology, 14, 279-292.
doi:10.1348/135532508X398602
Wechsler, D. (1997). Wechsler Memory Scale – third edition. San Antonio, TX: The
Psychological Corporation.
Weiermann, B., Stephan, M.A., Kaelin-Lang, A., & Meier, B. (2010). Is there a
recognition memory deficit in Parkinson’s Disease? Evidence from estimates
of recollection and familiarity. International Journal of Neuroscience, 120(3),
211-216. doi:10.3109/00207450903506510
Wenger, W., & Videbeck, R. (1969). Eye pupillary measurement of aesthetic
responses to forest scenes. Journal of Leisure Research, 1(2), 149-161.
Westerman, D.L. (2001). The role of familiarity in item recognition, associative
recognition, and plurality recognition on self-paced and speeded tests.
Journal of Experimental Psychology: Learning,Memory & Cognition, 27(3),
723-732. doi:10.1037/0278-7393.27.3.723
Wheeler, M.E., & Buckner, R.L. (2004). Functional-anatomic correlates of
remembering and knowing. NeuroImage, 21(4), 1337-1349.
doi:10.1016/j.neuroimage.2003.11.001
336
White, G.L., & Maltzman, I. (1978). Pupillary activity while listening to verbal
passages. Journal of Research in Personality, 12(3), 361-369.
doi:10.1016/0092-6566(78)90062-4
Wickelgren, W.A. (1979). Chunking and consolidation: A theoretical synthesis of
semantic networks, configuring in conditioning, S-R versus cognitive learning,
normal forgetting, the amnesic syndrome, and the hippocampal arousal
system. Psychological Review, 86(1), 44-60. doi:10.1037/0033-295X.86.1.44
Wiese, H., & Daum, I. (2006). Frontal positivity discriminates true from false
recognition. Brain Research, 1075, 183-192. doi:10.1016/j.brainres.2005.
12.117
Wiggs, C.L., & Martin, A. (1998). Properties and mechanisms of perceptual priming.
Current Opinion in Neurobiology, 8(2), 227-233. doi:10.1016/S09594388(98)80144-X
Wilder, J. (1958). Modern psychophysiology and the law of initial value. American
Journal of Psychotherapy, 12(2), 199-221.
Wilding, E.L., Doyle, M.C., & Rugg, M.D. (1995). Recognition memory with and
without retrieval of context: an event-related potential study.
Neuropsychologia, 33(6), 743-767. doi:10.1016/0028-3932(95)00017-W
Wilding, E.L., & Rugg, M.D. (1996). An event-related potential study of recognition
memory with and without retrieval of source. Brain, 119(3), 889-905.
doi:10.1093/brain/119.3.889
Wilding, E.L., & Rugg, M.D. (1997a). Event-related potential and the recognition
memory exclusion task. Neuropsychologia, 35(2), 119-128.
doi:10.1016/S0028-3932(96)00076-0
Wilding, E.L., & Rugg, M.D. (1997b). An event-related potential study of recognition
memory for words spoken aloud or heard. Neuropsychologia, 35(9), 11851195. doi:10.1016/S0028-3932(97)00048-1
Wilding, E.L., & Sharpe, H. (2003). Episodic memory encoding and retreival: Recent
insights from event-related potentials. In A. Zani, & A.M. Proverbio (Eds.),
The cognitive electrophysiology of mind and brain (pp. 169-196). San Diego,
CA: Academic Press.
Wilhelm, B.J., Heine, C., Widmann, A., Durst, W., Lüdtke, H., & Otto, G. (2010). How
sleepy are construction workers during extended shifts? An objective
investigation of spontaneous pupillary oscillations in tunnel workmen.
Somnologie, 14(3), 200-206. doi:10.1007/s11818-010-0480-y
337
Wilhelm, B.J., Widmann, A., Durst, W., Heine, C., & Otto, G. (2009). Objective and
quantitative analysis of daytime sleepiness in physicians after night duties.
International Journal of Psychophysiology, 72(3), 307-313.
doi:10.1016/j.ijpsycho.2009.01.008
Wilks, S. (1883). On the pupil in emotional states. Brain, 6(1), 1-6.
doi:10.1093/brain/6.1.1
Wilks, S. (1885). The pupil of the eyes during emotion. Nature, 31(803), 458.
doi:10.1038/031458c0
Winograd, M.R., & Rosenfeld, J.P. (2011). Mock crime application of the Complex
Trial Protocol (CTP) P300-based concealed information test.
Psychophysiology, 48, 155-161. doi:10.1111/j.1469-8986.2010.01054.x
Wixted, J.T. (2007a). Dual-process theory and signal-detection theory of recognition
memory. Psychological Review, 114(1), 152-176. doi:10.1037/0033295X.114.1.152
Wixted, J.T. (2007b). Spotlighting the probative findings: A reply to Parks and
Yonelinas (2007). Psychological Review, 114(1), 203-209. doi:10.1037/0033295X.114.1.203
Wixted, J.T., & Mickes, L. (2010). A continuous dual-process model of
remember/know judgments. Psychological Review, 117(4), 1025-1054.
doi:10.1037/a0020874
Wixted, J.T., & Stretch, V. (2004). In defense of the signal-detection interpretation of
Remember/Know judgments. Psychonomic Bulletin & Review, 11(4), 616641.
Wolk, D.A., Schacter, D.L., Lygizos, M., Sen, N.M., Holcomb, P.J., Daffner, K.R., &
Budson, A.E. (2006). ERP correlates of recognition memory: Effects of
retention interval and false alarms. Brain Research, 1096, 148-162.
doi:10.1016/j.brainres.2006.04.050
Wolk, D.A., Schacter, D.L., Lygizos, M., Sen, N.M., Chong, H., Holcomb, P.J., …
Budson, A.E. (2007). ERP correlates of Remember/ Know decisions:
Association with the late posterior negativity. Biological Psychology, 75(2),
131-135. doi:10.1016/j.biopsycho.2007.01.005
Woodmansee, J.J. (1966). Methodological problems in pupillographic experiments.
In Proceedings from The 74th Annual Convention of the American
Psychological Association, 1, 133-34.
338
Woodmansee, J.J. (1967). The pupil reaction as an index of positive and negative
affect. Paper presented at The meeting of the American Psychological
Association. Washington, DC: American Psychological Association.
Woodmansee, J.J. (1970). The pupil response as a measure of social attitudes. In
C. Sumers (Ed.), Attitude measurement. Chicago: Rand-McNally.
Wootton, A.C. (1910). Chronicles of Pharmacy. London: Macmillan & Co Ltd.
Wright, P., & Kahneman, D. (1971). Evidence for alternative strategies of sentence
retention. Quarterly Journal of Experimental Psychology, 23(2), 197-213.
doi:10.1080/14640747108400240
Wu, T.C.K., Allen, M.D., Goodrich-Hunsaker, N.J., Hopkins, R.O., & Bigler, E.D.
(2010). Functional neuroimaging of symptom validity testing in traumatic brain
injury. Psychological Injury & Law, 3(1), 50-62. doi:10.1007/s12207-0109067-y
Yamaji, K., Hirata, Y., & Usui, S. (2000). A method for monitoring autonomic nervous
activity by pupillary flash response. Systems & Computers in Japan, 31(4),
22-31. doi:10.1002/(SICI)1520-684X(200004)31:4<22::AID-SCJ3>3.0.CO;2W
Yochim, B.P., Kane, K.D., Horning, S., & Pepin, R. (2010). Malingering or expected
deficits? A case of herpes simplex encephalitis. Neurocase, 16(5), 451-460.
doi:10.1080/13554791003623334
Yonelinas, A.P. (1994). Receiver-operating characteristics in recognition memory:
Evidence for a dual-process model. Journal of Experimental Psychology:
Learning,Memory & Cognition, 20(6), 1341-1354. doi:10.1037/02787393.20.6.1341
Yonelinas, A.P. (1997). Recognition memory ROCs for item and associative
information: The contribution of recollection and familiarity.Memory &
Cognition, 25(6), 747-763. doi:10.3758/BF03211318
Yonelinas, A.P. (1999). The contribution of recollection and familiarity to recognition
and source-memory judgments: A formal dual-process model and an analysis
of receiver operating characteristics. Journal of Experimental Psychology:
Learning,Memory & Cognition, 25(6), 1415-1434. doi:10.1037/02787393.25.6.1415
Yonelinas, A.P. (2001a). Components of episodic memory: The contribution of
recollection and familiarity. Philosophical Transactions of the Royal Society B,
356, 1363-1374. doi:10.1098/rstb.2001.0939
339
Yonelinas, A.P. (2001b). Consciousness, control, and confidence: The 3 Cs of
recognition memory. Journal of Experimental Psychology: General, 130, 361379. doi:10.1037/0096-3445.130.3.361
Yonelinas, A.P. (2002). The nature of recollection and familiarity: A review of 30
years of research. Journal of Memory & Language, 46(3), 441-517.
doi:10.1006/jmla.2002.2864
Yonelinas, A.P., & Jacoby, L.L. (1996). Noncriterial recollection: Familiarity as
automatic, irrelevant recollection. Consciousness & Cognition, 5(1-2), 131141. doi:10.1006/ccog.1996.0008
Yonelinas, A.P., Kroll, N.E., Dobbins, I., Lazzara, M., & Knight, R.T. (1998).
Recollection and familiarity deficits in amnesia: Convergence of remember–
know, process dissociation, and receiver operating characteristic data.
Neuropsychology, 12, 323-339. doi:10.1037/0894-4105.12.3.323
Yoss, R.E. (1969). The sleepy driver: A test to measure ability to maintain alertness.
Mayo Clinic Proceedings, 44(11), 769-783.
Yoss, R.E. (1970). The inheritance of diurnal sleepiness as measured by
pupillography. Mayo Clinic Proceedings, 45(6), 426-437.
Yoss, R.E., Moyer, N.J., Carter, E.T., & Evans, W.E. (1970). Commercial airline pilot
and his ability to remain alert. Aerospace Medicine, 41, 1339-1346.
Yoss, R.E., Moyer, N.J., & Hollenhorst, R.W. (1970). Pupil-size and spontaneous
pupillary waves associated with alertness, drowsiness, and sleep. Neurology,
20(6), 545-554.
Yoss, R.E., Moyer, N.J., & Ogle, K.N. (1969). The pupillogram and narcolepsy. A
method to measure decreased levels of wakefulness. Neurology, 19(10), 921928.
Young, F.A., & Biersdorf, W. (1954). Pupillary contraction and dilation in light and
darkness. Journal of Comparative & Physiological Psychology, 47(3), 264268. doi:10.1037/h0057928
Yovel, G., & Paller, K.A. (2004). The neural basis of the butcher-on-the-bus
phenomenon: When a face seems familiar but is not remembered.
NeuroImage, 21, 789-800. doi:10.1016/j.neuroimage.2003.09.034
Zahn, T.P., Little, B.C., & Wender, P.H. (1978). Pupillary and heart rate reactivity in
children with minimal brain dysfunction. Journal of Abnormal Child
Psychology, 6(1), 135-147. doi:10.1007/BF00915789
340
Zekveld, A.A., Kramer, S.E., & Festen, J.M. (2011). Cognitive load during speech
perception in noise: The influence of age, hearing loss, and cognition on the
pupil response. Ear & Hearing, 32(4), 498-510.
doi:10.1097/AUD.0b013e31820512bb
Zhu, Y., Iba, M., Rajkowski, J., & Aston-Jones, G. (2004). Projection from the
orbitofrontal cortex to the locus coeruleus in monkeys revealed by
anterograde tracing. Society for Neuroscience Abstracts, 30, 211-213.
Zuckerman, M. (1971). Physiological measures of sexual arousal in the human.
Psychological Bulletin, 75(5), 297-329. doi:10.1037/h0030923
341
9. Appendices
A. Appendix: Experiments 1 and 1b
Study List
ANIMAL
ANSWER
ARTIST
BATTLE
BRIDGE
BUDGET
CHURCH
CIRCLE
COLUMN
CORNER
COTTON
DEGREE
DRIVER
ENGINE
FAMILY
FIGURE
FRIEND
GENIUS
HEAVEN
ISLAND
LETTER
MARBLE
NATION
NATURE
ORANGE
PENCIL
PERSON
PLANET
POCKET
POWDER
PRINCE
PRISON
REPORT
SADDLE
SQUARE
STRING
SYMBOL
THROAT
TONGUE
WINDOW
DOCTOR
FATHER
FLOWER
GROUND
HUNTER
JACKET
LAWYER
LESSON
LIQUID
MARKET
MIRROR
MOTHER
OFFICE
OXYGEN
PALACE
PARADE
PERMIT
RECORD
RESULT
SCHOOL
SHADOW
SIGNAL
SPIRIT
STABLE
STREET
SUMMER
TEMPLE
VACUUM
VALLEY
WEAPON
Recognition List
AVENUE
BARREL
BEAUTY
BOTTLE
BRANCH
BUTTER
CAREER
CATTLE
COFFEE
DANCER
B. Appendix: Experiment 2
Study List 1
BURIAL
CHISEL
TRIPOD
AVENUE
BUBBLE
STRIPE
SCHOOL
REWARD
RUMBLE
SQUIRE
LUXURY
BELIEF
COMEDY
PLEDGE
VOLUME
CAREER
DINNER
RECALL
KNIGHT
INFANT
RUDDER
TRANCE
SLEEVE
BLONDE
REMARK
TYRANT
COUSIN
AUTUMN
FUTURE
NICKEL
SUPPLY
COLUMN
REPORT
PARCEL
RIBBON
BUTTON
SYMBOL
HORROR
PLUNGE
COLLAR
342
Recognition List 1
LAWYER
BASKET
MOTIVE
REMEDY
MOMENT
AUTHOR
PRIEST
PALACE
NATION
HAZARD
MARVEL
CLOVER
VESSEL
GARLIC
OBJECT
PERSON
AMOUNT
COTTON
SHOVEL
RATTLE
COFFIN
PRINCE
REASON
KITTEN
MASTER
PICKLE
WINDOW
ISLAND
CANNON
HONOUR
BEAUTY
SADDLE
NECTAR
SPIDER
CAVERN
POCKET
BREEZE
WEIGHT
DESIGN
ARTIST
BREATH
RECORD
SHAKER
SHRIEK
GROWTH
MOTION
DANCER
NEPHEW
GRUDGE
ORIGIN
MATTER
COURSE
BORDER
ANSWER
BOTTLE
GASKET
SUNSET
SPIRIT
SAFETY
RABBIT
COPPER
THRONG
RESULT
BRONZE
POWDER
FELLOW
MINUTE
MAGNET
GROUND
STABLE
GROCER
APPEAL
DECREE
NUMBER
SQUEAK
WEALTH
DAMAGE
CARROT
DANGER
GALAXY
OYSTER
SISTER
REGRET
LAMENT
MOTHER
FOREST
INSECT
WISDOM
NAPKIN
CATTLE
WINTER
SPONGE
GIRDLE
PIGEON
FACTOR
ORANGE
CHERRY
ERRAND
BRANDY
THRILL
UMPIRE
NATURE
CIRCUS
ENGINE
NEEDLE
PRISON
SIGNAL
CORNER
SPHERE
POSTER
HATRED
BLOUSE
JUNIOR
FELINE
RESORT
Study List 2
DENIAL
FABRIC
MARKET
SKETCH
ANCHOR
BALLOT
BRANCH
SALARY
SORROW
RUBBER
Recognition List 2
TUMBLE
PRAISE
MORTAL
PARADE
EMPIRE
SPEECH
MAIDEN
SEASON
CINDER
DEBATE
Study List 3
CELLAR
NUTMEG
PASTOR
VANITY
POETRY
343
KETTLE
FAMILY
ESTATE
ENAMEL
SCREAM
VELVET
DEVICE
CHURCH
STROKE
THREAT
TEMPER
ROCKET
IMPACT
MUTTON
PEOPLE
REPAIR
HAMLET
WIZARD
VIRTUE
INCOME
RITUAL
HUNGER
HOCKEY
MEADOW
GUTTER
ARMOUR
TEMPLE
GINGER
CIRCLE
LENGTH
GRAVEL
GENIUS
CHAPEL
SPRING
METHOD
SHIVER
STRIDE
LOCKER
CEREAL
TENNIS
PARISH
HELMET
DELUGE
MIRROR
MANURE
LABOUR
MONKEY
STREET
WALLET
EXCUSE
TIMBER
PEPPER
SATIRE
NARROW
ADVICE
LUMBER
EFFECT
HAMMER
IMPORT
TREATY
INJURY
CHILLY
AERIAL
POLLEN
CANDLE
GOSPEL
BULLET
DEFEAT
CRADLE
PILLOW
PLIERS
VICTIM
PROFIT
LEADER
NATIVE
PUZZLE
THEORY
CEMENT
GENDER
TICKET
DECEIT
PATENT
SUMMER
THROAT
BUTTER
TOMATO
SECOND
ESCAPE
CHANCE
DRIVER
SEQUEL
TONGUE
SQUINT
BRIDGE
WEAPON
POTATO
JACKET
PRAYER
MEMORY
RESCUE
Recognition List 3
TRAVEL
SYSTEM
MENACE
BISHOP
RELIEF
STARCH
SUPPER
TALENT
TURTLE
RACKET
Study List 4
GALLON
LESSON
SQUARE
BANKER
PHRASE
BUDGET
MARGIN
WALNUT
APATHY
VIOLIN
Recognition List 4
WICKET
CUSTOM
RIDDLE
MISERY
DEGREE
CHROME
MEMBER
BARREL
REVOLT
BUCKET
WALRUS
HURDLE
FIGURE
TREMOR
SHOWER
COFFEE
HEALTH
KENNEL
CARPET
SULTAN
344
Calibration List
HUNTER
BANANA
ANIMAL
TERROR
OFFICE
NOTICE
CORPSE
PESTLE
SEARCH
STATUE
DOCTOR
STRAIN
BURROW
FLOWER
PUDDLE
BOTHER
PERIOD
AGENCY
SHRIMP
FATHER
SILVER
PUBLIC
WILLOW
LIQUID
DOLLAR
TUNNEL
HEIGHT
VALLEY
INSULT
FINISH
PENCIL
PLANET
THREAD
CRISIS
MORTAR
HUMOUR
SHIELD
SHADOW
RATION
PARDON
C. Appendix: Experiments 3 and 4
Word List A
OPINION
BALANCE
DELIGHT
CONTACT
INSIGHT
TRIBUTE
SUCCESS
VEHICLE
PERFECT
SHELTER
PRODUCE
EXHIBIT
PRODUCT
EMOTION
VICTORY
ROUTINE
DESPAIR
DISPUTE
MYSTERY
PASSAGE
VARIETY
ARTICLE
CONTEXT
NEUTRAL
BOATING
COMFORT
FAILURE
PASSION
OPENING
BENEFIT
CONCERN
LIBRARY
VETERAN
TRIUMPH
SILENCE
HARMONY
WELCOME
ANXIETY
PARTNER
PROMISE
MESSAGE
EXTREME
UNIFORM
LOYALTY
EXPRESS
ARRIVAL
COMMAND
ATTEMPT
CONCEPT
OUTCOME
PRIMARY
ABILITY
VVTRTTVM
XMXRTVTM
XMXRVTRVM
XXRVTRTVM
VVTRVTRVM
XMXRTTTVM
XMMXRTVTM
Word List B
WITNESS
REALITY
CIRCUIT
REMOVAL
MEASURE
PAYMENT
DIGNITY
EDITION
Grammar Study List A
XMMXM
VTTVTM
XMXRVM
VVTRTVM
XXRTTVM
VTTTTVM
XMMXRTVM
XMMMXRVM
345
Grammar Study List B
XMTRM
VVRMTM
XMTRRM
VTRRRRM
XMVTRXM
VVTRXRM
VVTRXRRM
XMTRRRRM
XMVRMTRM
VVTTRXRM
VVRMVRXRM
XXRRRRRRM
VVTTTTRXM
VVTTTRMTM
VTRRRRRRM
XMMMMMXM
VTTTTVTM
XXRVTRVM
VTVTRVTM
VTTTTTVTM
VTTVTRTVM
XMMXRTTVM
VTVTRTVTM
XMMMMMMXM
VTTTVTRVM
VVTRTTVTM
VTTTTTTVM
XXRTTTTVM
XXRTVTRVM
VVTRMTRM
XMVTRMTM
VVRMTRRM
VVTTRMTM
XMVTRMTRM
VVTTTRXRM
XMVRMVRXM
XMVRMTRRM
VVTRMVRXM
VVTRXRRRM
VVTRMTRRM
XMVTTTRXM
VVTTRXRRM
VVRXRRRRM
Grammar Recognition List A
VTVTM
XXRVM
XMMMXM
XXRTVM
XXRVTM
VVTRVM
VTTTVTM
XMMMMXM
VVTRVTM
XXRTVTM
XMXRVTM
XMMXRVM
XMXRTTVM
VTTVTRVM
VTVTRTVM
VTTTTTVM
Grammar Recognition List B
VTRRM
XXRRM
VVRXRM
VVTRXM
XMVRXM
XXRRRM
XMVRXRM
VVTTRXM
XMVRMTM
XXRRRRM
VVRMTRM
VVRXRRM
XXRRRRRM
VVTTTRXM
VTRRRRRM
XMVTRXRM
D. Appendix: Experiment 5
Study List 1
UNIFORM
BLISTER
ANTIQUE
SKYLARK
DEPOSIT
POVERTY
SUCCESS
LEAFLET
OUTCOME
COMPANY
STATION
TROUBLE
CITIZEN
MERCURY
DIAMOND
LEATHER
MONSOON
PRODUCT
ESSENCE
DREAMER
EDITION
MUSTARD
TRACTOR
MIRACLE
SPEAKER
WITNESS
DUNGEON
WARRIOR
BRAVERY
MANSION
GRAMMAR
LOYALTY
FORTUNE
PATIENT
MINERAL
VEHICLE
PALETTE
EMERALD
FEELING
PROBLEM
346
Recognition List 1
COLONEL
PAYMENT
STUDENT
HONESTY
BALANCE
BLOSSOM
PIONEER
CABINET
BANDAGE
TRUMPET
CAPSULE
GARMENT
RECITAL
DOORWAY
MACHINE
RECEIPT
BOREDOM
FORFEIT
SOLDIER
TRIBUTE
VILLAGE
PATTERN
IMPULSE
STADIUM
COTTAGE
RACQUET
PLATTER
CONTROL
CHARITY
STEEPLE
HADDOCK
EMOTION
PROTEST
CHANNEL
INSIGHT
SOCIETY
EVENING
ABILITY
BEDROOM
HOLIDAY
PEASANT
PARTNER
SILENCE
EMPEROR
TROLLEY
LAUNDRY
FANTASY
GLITTER
PORTION
RESIDUE
RECRUIT
WEATHER
CHUCKLE
BLANKET
FIGMENT
DENTIST
LOBSTER
FLANNEL
MILEAGE
SUBJECT
TOURIST
ARRIVAL
MESSAGE
COUNCIL
SEAWEED
MORNING
INQUIRY
CANTEEN
QUARTER
CUISINE
PASTURE
CENTURY
BOATING
FREEDOM
TRAGEDY
CARAVAN
HARNESS
FOREARM
PROFILE
FASHION
ABDOMEN
ARTICLE
MIXTURE
CAPTAIN
CABBAGE
CEILING
SERVICE
REALITY
CONCERT
APRICOT
DUCHESS
COLLEGE
MONARCH
CULTURE
BROTHER
JUSTICE
BUTCHER
FURNACE
PIANIST
WELCOME
THIMBLE
CREATOR
TEMPEST
OATMEAL
THOUGHT
RESPECT
MALARIA
BOUQUET
SPINACH
SUNBURN
HISTORY
MEETING
FALLACY
ROMANCE
VINEGAR
Study List 2
MERMAID
TYPHOON
COSTUME
HERRING
THUNDER
GALLERY
JOURNAL
HARVEST
SLIPPER
DIGNITY
Recognition List 2
VISITOR
BUILDER
SHUTTER
SULPHUR
SETTLER
REGENCY
SEGMENT
PACKAGE
HATCHET
APOLOGY
Study List 3
RETREAT
ASPIRIN
GLACIER
REFEREE
LANTERN
347
ACCOUNT
GRAVITY
SPANGLE
HAIRPIN
IMPRINT
LIBRARY
PURPOSE
STEAMER
ATHLETE
PHANTOM
PRELUDE
OFFICER
FACTORY
KNUCKLE
VOLCANO
MYSTERY
TEACHER
HYGIENE
TOBACCO
NURSERY
OPINION
HUSBAND
BISCUIT
ECONOMY
TRIUMPH
WHISTLE
MEASLES
ROBBERY
INTERIM
SURFACE
SUSPECT
PASSAGE
FLUTTER
GODDESS
BALLOON
FIELDER
HEROISM
BAGPIPE
VICTORY
SCHOLAR
Recognition List 3
SESSION
DYNASTY
CRYSTAL
PYRAMID
TRAILER
COMMAND
STOMACH
BARGAIN
CIRCUIT
FAILURE
COUNTRY
DISEASE
REVENGE
PICTURE
PAINTER
KINGDOM
FINANCE
CHICKEN
EPISODE
SLUMBER
Instructions
Scenario: We would like you to imagine that you are a person who has recently
been involved in a car accident. You were unconscious for 15 minutes after the
accident, and you had to spend one night in hospital for observation. Gradually
your condition improved over the following months and you have now made a
full recovery. Imagine that the purpose of the test that you are about to
undertake is to determine whether the accident has produced any memoryimpairments due to brain damage.
Instruction 1: Please perform to the best of your abilities on the recognition
memory task, answering ‘new’ when you think a word is new and ‘old’ when you
think a word is old.
Instruction 2: At your memory test, you decide to exaggerate the effects of your
accident in case there is extra compensation money available. Please produce
responses that would convince an examiner that you still have memory loss.
Impairments should be presented in a “believable” manner, and major
exaggerations, such as not remembering anything should be avoided, because
even if you were performing at chance you would still get about 50% of the
answers correct. £10 worth of book vouchers will be awarded to the individual
who best manages to simulate a believable memory deficit.
Instruction 3: Please say ‘new’ to new words and also say ‘new’ to old words
that you recognise.
348
E. Appendix: Experiment 6
Study List
NUMBER
MARBLE
SPRING
SUPPER
PROFIT
SILVER
MOTHER
TONGUE
ANIMAL
PRISON
FAMILY
MIRROR
SEASON
CORNER
REPORT
BRIDGE
DINNER
CELLAR
PALACE
PENCIL
SYMBOL
GROUND
COUSIN
AUTUMN
DANGER
FIGURE
DANCER
HEIGHT
BORDER
POCKET
CAREER
BOTTLE
PERSON
OFFICE
SIGNAL
NATURE
ESTATE
ORANGE
SPEECH
BEAUTY
ANSWER
BUDGET
MARKET
DOCTOR
CHURCH
SHADOW
SISTER
LETTER
ISLAND
CENTER
WEALTH
COTTON
DRIVER
BARREL
STREET
LESSON
POWDER
DEGREE
SCHOOL
CIRCLE
ARTIST
FRIEND
POETRY
WINDOW
BUTTER
LEADER
PARADE
SUMMER
PRAYER
ENGINE
Recognition List
LIQUID
PLANET
LAWYER
COLUMN
MINUTE
COFFEE
SADDLE
FLOWER
COMEDY
RECORD
F. Appendix: Experiment 7
Study List
BLANKET
DIAMOND
BALLOON
INCENSE
FACTORY
KNUCKLE
ANTIQUE
AMATEUR
BOUQUET
CHICKEN
SPINACH
GALLERY
ATHLETE
ECONOMY
CONCERT
BUTCHER
BEDROOM
ASPIRIN
BROTHER
LAUNDRY
FURNACE
APOLOGY
BISCUIT
LIGHTER
ALGEBRA
ACADEMY
SOLDIER
EPISODE
PYRAMID
CABBAGE
LEATHER
PADLOCK
HAIRPIN
COTTAGE
MINERAL
CARAVAN
COSTUME
MEASLES
BANDAGE
LANTERN
MUSTARD
ARCHERY
BOREDOM
Recognition List
HARVEST
349
GESTURE
RAINBOW
FINANCE
MANSION
SLIPPER
BLISTER
FASHION
COURAGE
HOSTAGE
PEASANT
LOBSTER
FLANNEL
BAGPIPE
SHUTTER
DISPLAY
CITIZEN
BLOSSOM
BAYONET
CAPTAIN
ALCOHOL
EMERALD
GLITTER
ROBBERY
JOURNAL
BUILDER
CEILING
FORTUNE
CANTEEN
CABINET
HOLIDAY
DUNGEON
SPEAKER
NURSERY
MISSILE
CRYSTAL
DENTIST
G. Appendix: Experiments 8 and 9
Weak Items – Block 1
SPLIT
CROWN
ROCKY
RAZOR
CRIME
ALIKE
BLEAK
CABIN
BLOOM
KNOCK
REBEL
LOGIC
ALIEN
CHEEK
RIGID
SPRAY
GUIDE
CREEK
STEAK
PITCH
CHOSE
PUPIL
BRAVE
STICK
SMART
SHEEP
ARISE
NOBLE
STORM
CIGAR
CRAZY
DRIFT
PROSE
SHOCK
MOTEL
ENTRY
FRUIT
GLOBE
CREAM
ADOPT
FAINT
CRAWL
WIRED
YIELD
ANGEL
ELITE
FAULT
CLOUD
SOLAR
BLAST
BENCH
SPARK
UPSET
CHARM
BELLY
CHAOS
FLOCK
AROSE
Strong Items – Block 1
MOUND
NURSE
VOCAL
HANDY
STAKE
FALSE
SMELL
THANK
GUEST
UNITE
PLATE
LOYAL
PRINT
MORSE
WAIST
SWIFT
New Items – Block 1
LABEL
BRIDE
PATCH
MAKER
BLADE
CLUNG
SLATE
STEEP
ORBIT
GLORY
PAINT
ARGUE
TOAST
TENSE
BAKER
STEAM
350
Weak Items – Block 2
ESSAY
BRICK
MOVIE
EXACT
BEARD
GLAZE
FLOOD
FLUSH
SQUAD
RANCH
OCEAN
DOUGH
SHADE
SUNNY
DUSTY
TOWER
ATLAS
STOVE
PENNY
TROOP
SPADE
NERVE
FOCUS
DRAIN
MERCY
SHAME
BADLY
WOUND
BACON
CLERK
CLOCK
TALES
BOOST
VAGUE
TRACE
CURSE
SWISS
QUEST
BATON
MEDIA
ALARM
FOLLY
RIDER
DIRTY
CANDY
ATTIC
ALERT
STRAW
LYRIC
DEBUT
HABIT
REACT
DODGE
DAIRY
SUITE
EAGER
SLAVE
GRACE
PURSE
SAUCE
IVORY
LEMON
LEVER
WIDTH
IRONY
GAUGE
TOOTH
GLOOM
SCOPE
GIANT
BRACE
BUNCH
Strong Items – Block 2
GUILT
CHILL
SEWER
CRUDE
TWIST
TIGHT
RADAR
CREST
DRILL
INPUT
CHASE
CYCLE
BRAND
MAYOR
CIVIC
GRAVE
New Items – Block 3
SPORT
BURST
SKIRT
MERGE
POUND
VIVID
ADULT
DEALT
NOISE
MOUSE
COACH
SAINT
HURRY
GRAPH
OWNER
TRACT
Weak Items – Block 3
GHOST
IMPLY
GROVE
LUNAR
SUGAR
THUMB
LAUGH
QUOTE
ACTOR
SLIDE
HELLO
FORGE
REALM
HONEY
VERSE
ADMIT
351
Strong Items – Block 3
SHIRT
DEVIL
NAVAL
ROAST
PANIC
MAGIC
FLASH
ELDER
STUFF
RALLY
GRAIN
CLIFF
PROOF
WRIST
FERRY
FLUID
TRICK
ARROW
PIANO
STUCK
LEASE
DELAY
GRILL
BLANK
CRASH
DITCH
JUICE
WORST
STARE
PAUSE
BLAME
LAYER
STERN
AMPLE
AWFUL
AWAKE
CHEAP
CLIMB
MASON
BLOND
HARSH
TREAT
PRIZE
FLEET
New Items – Block 3
FEVER
LOBBY
PASTE
ARRAY
TOKEN
FENCE
GRASP
ONION
RIVAL
MERIT
SLOPE
ALOUD
CRAFT
VIRUS
HAVEN
DRAFT