Journal of Psychophysiology by Hogrefe

Volume 33 / Number 1 / 2019

Journal of

Psychophysiology

Editor-in-Chief Michael Falkenstein Editorial Board Markus Breimhorst Tavis Campbell Ritobrato Datta Nicola Ferdinand Patrick Gajewski Edward Golob David R. Herring Sien Hu Julian Koenig Cristina Ottaviani Patrick Papart Daniel S. Quintana Walter Sannita Henrique Sequeira Franck Vidal Lin Wang Juliana Yordanova

An International Journal

Contents Articles

Journal of Psychophysiology (2019), 33(1)

Respiratory Sinus Arrhythmia During Sleep and Waking: Stability and Relation to Individual Differences in Waking Affective Style Elizabeth M. Stoakley, Karen J. Mathewson, Louis A. Schmidt, and Kimberly A. Cote

Rumination Moderates the Association Between Resting High-Frequency Heart Rate Variability and Perceived Ethnic Discrimination DeWayne P. Williams, Kinjal D. Pandya, LaBarron K. Hill, Andrew H. Kemp, Baldwin M. Way, Julian F. Thayer, and Julian Koenig

Habitual Use of Cognitive Reappraisal Is Associated With Decreased Amplitude of the Late Positive Potential (LPP) Elicited by Threatening Pictures Neil R. Harrison and Philippe Chassy

Job Satisfaction Among Mental Health Workers: Associations With Respiratory Sinus Reactivity to, and Recovery From Exposure to Mental Stress William H. O’Brien, Paul W. Goetz, Heather McCarren, Eileen Delaney, William F. Morrison, Tanya S. Watford, and Kristin A. Horan

Autonomic Cardiovascular Dysregulation at Rest and During Stress in Chronically Low Blood Pressure Stefan Duschek, Alexandra Hoffmann, Casandra I. Montoro, and Gustavo A. Reyes del Paso

Neural Correlates of Empathy for Physical and Psychological Pain Vera Flasbeck and Martin Brüne

Ó 2019 Hogrefe Publishing

Article

Respiratory Sinus Arrhythmia During Sleep and Waking Stability and Relation to Individual Differences in Waking Affective Style Elizabeth M. Stoakley,1 Karen J. Mathewson,2 Louis A. Schmidt,2 and Kimberly A. Cote1 1

Psychology Department, Brock University, St. Catharines, Ontario, Canada

Department of Psychology, Neuroscience & Behaviour, McMaster University, Ontario, Canada

Abstract: Resting respiratory sinus arrhythmia (RSA) is related to individual differences in waking affective style and self-regulation. However, little is known about the stability of RSA between sleep/wake stages or the relations between RSA during sleep and waking affective style. We examined resting RSA in 25 healthy undergraduates during the waking state and one night of sleep. Stability of cardiac variables across sleep/ wake states was highly reliable within participants. As predicted, greater approach behavior and lower impulsivity were associated with higher RSA; these relations were evident in early night Non-REM (NREM) sleep, particularly in slow wave sleep (SWS). The current research extends previous findings by establishing stability of RSA within individuals between wake and sleep states, and by identifying SWS as an optimal period of measurement for relations between waking affective style and RSA. Keywords: heart rate, RSA, sleep, impulsivity, affective style

Resting heart rate variability (HRV) has been associated with individual differences in cognitive, affective, and physiological regulation (Thayer, Hansen, Saus-Rose, & Johnsen, 2009). High frequency heart rate variability (HF HRV), often referred to as respiratory sinus arrhythmia (RSA), is a parasympathetic nervous system (PNS) phenomenon that is primarily driven by brainstem activation of the sinoatrial node of the heart via the vagus nerve (Berntson, Quigley, & Lozano, 2007). The PNS maintains moment-to-moment control of heart rate (HR), increasing or decreasing inhibitory efferent activity to the heart in response to situational demands (Porges, 2007). Brainstemdriven parasympathetic activity, however, is modulated by projections from higher brain regions (e.g., the amygdala, hypothalamus, and prefrontal neural structures; Benarroch, 1993). Because it is modulated by higher brain centers, vagal control of the heart is thought to reflect the ability of an individual to respond adaptively to his or her environment (Beauchaine, 2001; Grossman & Taylor, 2007; Porges, 2007; Thayer & Lane, 2000), with widespread implications for physical health (Thayer & Lane, 2007), and psychological well-being (Appelhans & Luecken, 2006).

Ó 2017 Hogrefe Publishing

Stability of Heart Rate Variability In the waking state, resting HRV measures have been shown to be reliable in adult, adolescent, and childhood samples. For example, El-Sheikh (2005) reported stable individual differences in resting baseline RSA, as well as RSA reactivity to a stressor, in children aged 6–13 who were tested on two occasions, two years apart. Resting HRV measures in the waking state have also been shown to be stable over time in healthy adults when measured after intervals of 30 (Borges, Mathewson, & Schmidt, 2017) and 65 days (Kleiger et al., 1991), 7 months (Pitzalis et al., 1996), and 5 years (Bornstein & Seuss 2000). Further, HF HRV recorded over three consecutive nights of sleep was found to be highly reliable in both good sleepers and patients with insomnia (Israel et al., 2012). Thus, RSA appears to be stable over time in both waking and sleep states, supporting its utility as a trait-like measure of psychophysiological regulation. The stability of HRV in individuals across different contexts and over time, along with consistent relations with psychological measures such as emotional responding and physical health, provides a foundation from which to further our understanding of individual differences in affective style

Journal of Psychophysiology (2019), 33(1), 1–12 https://doi.org/10.1027/0269-8803/a000200

(Appelhans & Luecken, 2006; Thayer & Lane, 2007). Consideration of the context of sleep in relation to HRV and affective style is of interest due to the well-known comorbid nature of insomnia and depression (Baglioni et al., 2011), and evidence linking poor sleep with emotional dysregulation, including aggression and violence (Kamphius, Meerlo, Koolhaas, & Lancel, 2012).

Cardiac Measures in Waking and Sleep States Cardiac behavior varies by time of day due to both circadian and sleep processes which have independent yet interacting influences (for reviews, see Chouchou & Desseilles, 2014; Lanfranchi & Somers, 2011; Tobaldini et al., 2013; Trinder, Waloszek, Woods, & Jordan, 2012; Stein & Pu, 2012). Non-REM (NREM) sleep constitutes about 75–80% of the sleep period, and relative to wake, it is a period of reduced consciousness and increased brain wave synchronization. Consistent with this notion, the PNS is more active and the SNS is less active in NREM sleep compared to the waking state. During NREM sleep, R-R intervals, HF HRV power, and pre-ejection period tend to increase, while blood pressure and muscle sympathetic nerve activity tend to decrease relative to wake (Lanfranchi & Somers, 2011; Trinder et al., 2012). REM sleep is a paradoxical state marked by an active brain in an atonic body; it is a period of fluctuating PNS and SNS activity, and thus cardiac instability (Lanfranchi & Somers, 2011; Tobaldini et al., 2013). During sleep, HR is under strong circadian influence, reaching a nadir in its circadian oscillation about two-thirds of the way through a sleep period at night. Heart rate also decreases systematically during the sleep onset period, is higher in REM sleep compared to NREM sleep, and increases during arousals from sleep. HRV measures have been shown to be sleep-stage dependent, being greater in REM states, and similar between wake and REM. Their values within discrete sleep states are fairly constant over time (Stein & Pu, 2012; Trinder et al., 2012). Sex differences have also been identified in HRV measures recorded during waking and sleep in healthy adults (e.g., Elsenbruch, Hamish, & Orr, 1999). Resting RSA was shown to be lower in men than women during wake, Stage 2, and REM sleep in a healthy adult sample, even after controlling for hormone levels, body mass index (BMI), and degree of habitual physical activity (Valladares, Eljammal, Motivala, Ehlers, & Irwin, 2008). If RSA is a trait-like physiological measure of selfregulation and affective style, then individual differences in RSA should be stable across different contexts, including waking and sleep states. There appear to be, however, no studies that have examined the relation between Journal of Psychophysiology (2019), 33(1), 1–12

E. M. Stoakley et al., RSA, Sleep, Affective Style

sleep-wake stability in RSA and individual differences in self-reported waking affective style.

Heart Rate Variability and Self-Regulation Associations between RSA and behavioral and emotional functioning have been documented across the life span (Beauchaine, 2001). In one line of research, higher versus lower resting RSA has been linked to positive emotional style (e.g., Oveis et al., 2009; Wang, Lu, & Qin, 2013). Relatively higher resting RSA has been associated with higher scores on measures of Extraversion (Big Five Inventory; Oveis et al., 2009), Trait Optimism (Life Orientation Test; Oveis et al., 2009), and Positive Affect (PANAS; Wang et al., 2013). Complementary to studies linking higher RSA to positive emotional functioning, research has also noted relations between lower RSA and psychopathology, including antisocial (Mezzacappa et al., 1997; Sloan et al., 1994), depressed (reviewed in Rottenberg, 2007), and anxious conditions (Brosschot, Van Dijk, & Thayer, 2007; Friedman & Thayer, 1998). These studies provide support for RSA as a nonspecific marker of emotion regulation, and simultaneously suggest that emotion dysregulation may be a risk factor for psychopathology (Beauchaine, 2001; Beauchaine, GatzkeKopp, & Mead, 2007; Porges, 2007). Resting RSA also has been linked to behavioral control. Impulsivity, or the tendency to act without adequate consideration of consequences, has been described as a heritable trait associated with maladaptive response to reward (Beauchaine, 2012). Its factors include behavioral activation, compulsivity, increased sensitivity to reward/ punishment, temporal discounting, and risk taking (Meda et al., 2009). In high-risk environments that promote emotional lability, impulsive behavior is highly likely to develop into externalizing behavior. In these situations, however, RSA may serve as a marker of emotion regulation capacity that buffers the effects of high-risk environments, thereby moderating the developmental trajectory of externalizing disorders (Beauchaine, 2012; Beauchaine et al., 2007).

The Present Study There were two goals for the present study. First, we investigated the stability of HR and RSA within individuals across sleep and waking states. Based on previous research demonstrating stability in HR and HRV measures over time (Borges et al., 2017; Bornstein & Seuss, 2000; Kleiger et al., 1991; Pitzalis et al., 1996), and within sleep states (e.g., Israel et al., 2012; Trinder et al., 2012), we predicted that HR and RSA in the current sample of healthy, young adults would be highly correlated across waking and sleep states. Ó 2017 Hogrefe Publishing

E. M. Stoakley et al., RSA, Sleep, Affective Style

Second, we explored associations between cardiac measures during sleep states and individual differences in waking affective style. Given that sleep is a relatively quiescent time, free from intrusive influences that might affect cardiac variables, measurement of HR and RSA during sleep was considered an opportune time from which to sample cardiovascular activity as an individual difference marker of waking affective style. Sampling RSA during sleep was also of interest due to relations between poor sleep and emotional dysregulation (Baglioni et al., 2011; Kamphius et al., 2012). We therefore examined HR and RSA in discreet sleep stages including early night Stage 2 (EST2), slow wave sleep (SWS), late night Stage 2 (LST2), and REM sleep (REM). Specifically, we examined HR and RSA in wake and sleep states as predictors of variation in waking affective responding using the Behavioral Inhibition-Activation Scale (BIS/BAS; Carver & White, 1994) and impulsivity using Barratt’s Impulsiveness Scale (BIS-11; Patton, Stanford, & Barratt, 1995). We predicted that RSA in sleep would be associated with positive emotional style, that is, with behavioral approach tendencies as measured by the BIS/BAS, and also with lower impulsivity as measured by Barratt’s Impulsiveness Scale.

Method Participants and Overview Participants were recruited as part of a larger study on the effects of sleep deprivation on attention and emotion regulation in young adults. Undergraduate students were recruited through an online recruitment system hosted by the Psychology Department of a midsize university in Ontario, Canada, and through classroom presentations and posters. Participants were considered eligible for participation if they were aged 18–30, in good health (free of medication, psychiatric history, or traumatic brain injury), and good sleepers (free of sleep disorders, regular sleep schedule). The present study comprised 25 participants (Men: n = 13; Mage = 19.23 years, SD = 1.48; Women: n = 12; Mage = 19.25 years, SD = 1.29) who were randomly assigned to a rested control group as part of the larger study. Detailed descriptions of the full study sample can be found elsewhere (Cote, McCormick, Geniole, Renn, & MacAulay, 2013; Cote, Mondloch, Sergeeva, Taylor, & Semplonius, 2014; Renn & Cote, 2013).

Procedures All study procedures were cleared by the institutional Research Ethics Board. A $110 honorarium or $90 plus Ó 2017 Hogrefe Publishing

credit toward a university course was provided upon study completion. In order to determine if volunteers met all inclusion criteria, a screening interview was conducted by telephone, followed by completion of online questionnaires. The screening interview and questionnaires were reviewed for indications of sleep habits or difficulties that would present as confounds in this study, such as a shifted sleep schedule or symptoms of insomnia. Eligible participants were then scheduled for an overnight polysomnography (PSG) screening in order to screen for sleep disordered breathing and periodic limb movement. Overnight sleep electrophysiological recordings were scored by experienced individuals according to standard procedures (Rechtschaffen & Kales, 1968), by categorizing each 30-s epoch into a discrete sleep stage (wake, 1, 2, slow wave sleep, or REM). Eligible participants were scheduled to return to the laboratory to complete the main study which consisted of two consecutive nights and one day in the sleep laboratory. Participants were instructed to avoid vigorous physical activity and to refrain from consuming caffeinated or alcoholic beverages for the duration of the study. Control participants completed a baseline night of sleep (Night 1) and were permitted to leave the laboratory in the morning (Day 1) and return that evening at 21:00 (Night 2) when they were permitted another 8-hr sleep opportunity (23:00–07:00). PSG recordings were obtained for each night of sleep and cardiac data were extracted from Night 2 recordings. During Day 2, electrode caps were applied at 09:30 and performance assessment batteries were administered at 10:30–12:00 and at 14:00–15:30. Waking cardiac recordings were obtained during a resting period at 14:05 on Day 2, during which participants were asked to sit quietly for five min, alternating eyes open and eyes closed for 30-s epochs. The purpose of the alternating eyes open and eyes closed instructions was related to measuring alertness during EEG recordings which were part of a larger study and are not germane to the HR data reported here. During this task, participants were seated, alone, in front of a computer in a private bedroom. See Figure 1 for a graphic of the study design.

Affective Style Measures Affective style measures were assessed as part of a questionnaire set: test measures included the Behavioral InhibitionActivation (BIS/BAS) scales (Carver & White, 1994), and the Barratt Impulsiveness Scale (Patton et al., 1995). Data were missing for one participant on the latter scale. Behavioral Inhibition-Activation (BIS/BAS) Scales Participants completed the BIS/BAS scales, which assess self-reported behavioral tendencies related to activation Journal of Psychophysiology (2019), 33(1), 1–12

E. M. Stoakley et al., RSA, Sleep, Affective Style

SCREENING

Phone interview, online screening questionnaires and PSG Night

Night 1

Day 1

Sleep in Lab

23:00

7:00

Night 2

Day 2

Sleep in Lab

Performance Assessment

23:00

23:00-01:00 Early Stage 2 Sleep Slow Wave Sleep

7:00

05:00-07:00 Late Stage 2 Sleep REM Sleep

14:05 Waking

Figure 1. Study Design for rested control participants from pre-study screening for physical and mental health including sleep characteristics, through to the performance assessment day. Participants spent two consecutive nights in the laboratory environment during data collection with electrophysiological recordings made each night. ECG data were extracted from Night 2 during the first 2 hr of the sleep opportunity for 5-min epochs of Stage 2 and slow wave sleep and during the last 5-min epochs of Stage 2 and REM sleep prior to awakening. Waking ECG was extracted during a seated recording during the performance assessment on Day 2.

(BAS) and inhibition (BIS). The questionnaire consists of 20 items rated on a 4-point rating scale, ranging from “strongly disagree” (1) to “strongly agree” (4). Behavioral activation is assessed in three subscales: Drive (e.g., “When I want something, I usually go all-out to get it,” four items), Reward-sensitivity (e.g., “When I get something I want, I feel excited and energized,” five items), and Fun-seeking (e.g., “I will often do things for no other reason than that they might be fun,” four items). Drive indexes a preference to strive for appetitive goals, Reward-Responsiveness assesses sensitivity to reward cues and escape from punishment, and Fun-seeking indexes impulsivity and attraction to novel, rewarding events. The behavioral inhibition subscale comprises seven items assessing sensitivity to punishment (e.g., “I feel pretty worried or upset when I think or know somebody is angry at me”). Higher scores indicate greater drive, reward sensitivity, fun-seeking, or inhibition, respectively (Carver & White, 1994). Validity and reliability for each of the subscales is reported to be good (e.g., Cronbach’s α Drive: .80; Reward-sensitivity: .83; Funseeking: .70; BIS: .76; Jorm et al., 1999). Barratt Impulsiveness Scale The 30 items of this self-report instrument assess individual differences in trait impulsivity in nonselected and clinical populations (Patton et al., 1995). Higher scores indicate greater impulsivity. Factor analysis of the Barratt Impulsiveness scale identified six first-order factors, labeled Attention (e.g., “I am restless at the theatre or lectures,” five items), Motor Impulsiveness (e.g., “I act on the spur of the moment,” seven items), Self-control (e.g., “I plan tasks carefully, six items), Cognitive Complexity (e.g., “I like to think about complex problems,” five items), Perseverance (e.g., “I change jobs,” four items), and Cognitive Instability (e.g., “I often have extraneous thoughts when thinking,” three items). The total score was significantly correlated Journal of Psychophysiology (2019), 33(1), 1–12

with each of the single factor scores, suggesting it was an internally consistent measure of impulsivity suitable for applied settings.

Psychophysiological Data Collection All electrophysiological signals during PSG screening, baseline, and experimental nights, and waking electrophysiology were recorded using Neuroscan Synamps II amplifiers and version 4.5 software (Neuroscan, Inc., El Paso, TX). Electrodes recorded electrocardiography (ECG; below right and left clavicle), electromyography (EMG; submental), electrooculography (EOG; outer canthus of each eye), and electroencephalography (EEG; O1, O2, C3, and C4 for sleep recordings; a Neuroscan 64-channel Ag/AgCl Quikcap with a central site reference between Cz and CPz was used for waking recordings). Impedances for PSG and waking data were maintained at 10 KΩ or less. Prior to sleep scoring, recordings were re-referenced offline to the contralateral mastoid site, A1 or A2. Hardware filters used to record electrophysiology were direct current (DC) to 100 Hz. Only the ECG measures are reported below. The other measures were collected as part of the larger study and are not discussed further. In both waking and sleep recordings, respiration was spontaneous, and respiration measures were not collected during waking rest or across the experimental night. Participants were asked to sit quietly during waking recordings, which is considered acceptable for withinparticipants comparisons of HRV (Allen, Chambers, & Tower, 2007; Houtveen, Rietveld, & de Geus, 2002). It has also been suggested that similar to seated rest, the relatively quiescent nature of sleep effectively minimizes any confound between respiration rate and RSA. In resting conditions, spontaneous respiration typically does not Ó 2017 Hogrefe Publishing

E. M. Stoakley et al., RSA, Sleep, Affective Style

produce changes in RSA that are independent of centrallymodulated vagal efferent activity (Allen et al., 2007; Trinder et al., 2001). Rather, in quiescent settings, any changes in RSA are likely to be driven by centrally-modulated vagal efferent activity, and are not simply the result of respiratory changes. Thus, it is not necessary to control for respiration rate during sleep when RSA is a component of interest in research (Trinder et al., 2001). Further support for not controlling for respiration was summarized in a recent review which emphasized the function of the vagus as a bidirectional pathway associated with both cardiac and respiratory functioning (Thayer, Loerbroks, & Sternberg, 2011). The authors cautioned against controlling for respiration due to the complex interrelations between these physiological systems and posited that the likely consequence of such control would present as lost variability in cardiac measures associated with vagal tone, and thus, skewed results with respect to neural influences on the cardiovascular system.

Data Reduction and Quantification A 4-min waking baseline measure of ECG was extracted for each participant from a seated resting period as described above. The PSG recording of the experimental night (Night 2) in the laboratory was sleep scored, and 5-min epochs of ECG were extracted from early night Stage 2 sleep (EST2), slow wave sleep (SWS), late night Stage 2 sleep (LST2), and REM sleep. Early and late Stage 2 epochs were selected as the first or last continuous artifact-free 5-min segment identified in the record. Segments of SWS were selected from the period containing the most distinct delta wave activity during the first 2 hr of the sleep opportunity, and REM was taken from the final REM period at the end of the night where REM pressure is highest. These epochs were chosen in order to sample both NREM and REM sleep states, to examine both light and deep NREM sleep, and to assess the influence of time of night (EST2 vs. LST2) due to changing arousal thresholds. Each individual file was exported as an ASCII (text) file which was converted and analyzed using a commercial software package (MindWare Heart Rate Variability Scoring Module 3.0.22, MindWare Technologies Ltd., Columbus, OH). R-waves were visually checked in the MindWare software, and manually edited where appropriate, according to the recommendations outlined by Berntson and Stowell (1998). Artifact occasionally obscured the ECG record, resulting in removal of 246.6 s of data, which accounted for less than 1% of the data analyzed. When a single R-wave was obscured, a mid-beat function was utilized to approximate the location of the missing beat; this method was used for 31 R-waves. Sixteen of the thirtyone beats were from a single file (case 25, in SWS); statistical comparisons were run including and excluding this specific case. It was determined that inclusion or exclusion Ó 2017 Hogrefe Publishing

of this case did not alter the pattern of results in the relations tested, and as such, the case was included in the final data set. Spectral analysis of the heart beat series provides a noninvasive technique for indirect measurement of autonomic modulation during sleep. Using a Fast Fourier Transformation (FFT), we calculated HRV based on 1-min epochs, which were then averaged for each waking or sleep state. High frequency values were defined as occurring between 0.12 and 0.40 Hz; these values underwent a natural log transform in order to normalize the distribution, yielding measures of RSA. The resulting values in the high frequency range will hereafter be referred to as RSA.

Statistical Analyses Examination of the HR and log-transformed RSA data revealed that these data were normally distributed. Any violations to assumptions of normality in the subscales of personality measures are reported in Results section. Repeated-measures analyses of variance (ANOVAs) were conducted to test the effect of sleep/wake State (wake, EST2, SWS, LST2, and REM) and Sex by State interactions on the cardiac measures. Effect size is given by eta squared (η2). There were no violations to the assumption of sphericity in any of the ANOVA test results. Pearson correlations and Intraclass Correlation Coefficients (ICC; Shrout & Fleiss, 1979) were calculated to test the stability of HR and RSA across waking and sleep states. Next, a series of regression analyses was carried out to determine the unique contributions of cardiovascular activity during slow wave sleep to affective style, focusing on variables where bivariate correlations were largest. Sex and age were investigated as covariates and are reported in the Results only if they accounted for significant variance in the models.

Results Stability of Cardiac Measures in Sleep and Waking Mean HR and RSA values by state are depicted in Figure 2 and Table 1. Heart rate differed significantly between states, F4,92 = 21.67, p < .001, η2 = 0.49. Pairwise comparisons using the Bonferroni correction revealed that waking HR was significantly faster than HR during each sleep stage (all ps .003). As well, HR during SWS (p = .001), and during EST2 sleep (p = .044), was significantly faster than HR in LST2 sleep. There was no effect of Sex on HR or interaction between Sex and State. Journal of Psychophysiology (2019), 33(1), 1–12

E. M. Stoakley et al., RSA, Sleep, Affective Style

Figure 2. Mean Heart Rate (bpm) by condition (A), error bars represent standard error. HRwake was significantly higher than HREST2, HRSWS, HRLST2, and HRREM (all ps .003), within sleep stages HRSWS was higher than HRLST2 (p = .001) and HREST2 was higher than HRLST2 (p = .044). RSA (in ms2) by condition (B), error bars represent standard error. RSASWS was lower than RSA(p = .024) and RSALST2 EST2 (p = .018).

(A) 80.00

Heart Rate (bpm)

75.00

70.00

65.00

60.00

55.00 Wake

Early Stage 2

Slow Wave

Late Stage 2

REM

State

(B) 7.50 7.30

RSA (ln ms²)

7.10 6.90 6.70 6.50 6.30 6.10 5.90 5.70 5.50 Wake

Early Stage 2

Slow Wave

Late Stage 2

REM

State

Table 1. Mean (SD) and intraclass correlation coefficients by waking/sleep stage Wake

Early stage 2

SWS

Late stage 2

REM

Mean (SD)

ICC

71.99 (10.53)

63.62 (10.02)

64.87 (11.88)

59.68 (9.42)

62.70 (8.46)

0.945***

RSA

6.83 (0.72)

6.99 (1.35)

6.23 (1.39)

7.08 (1.07)

6.80 (1.10)

0.841***

Note. ***Intraclass Correlation Coefficient (ICC) significant at p < .001.

Similarly, RSA differed significantly between states, F4,92 = 4.16, p = .004, η2 = 0.15. Pairwise comparisons of RSA revealed significantly lower RSA in SWS compared to both EST2 (p = .024) and LST2 sleep (p = .018). Waking RSA did not differ significantly from RSA in any sleep state. There was no effect of Sex on RSA or interaction between Sex and State. Pearson correlations indicated high stability of HR and RSA across sleep and waking, and between sleep stages Journal of Psychophysiology (2019), 33(1), 1–12

(see Table 2). Intraclass Correlation Coefficients were also significant for both cardiac measures: HR, F24,96 = 18.04, p < .001, CI [.90, .97], and RSA, F24,96 = 6.30, p < .001, CI [.72, .92], (see Table 1). The ICC is a measure of relative reliability, and provides information about the tendency for an individual to maintain their rank order position within a group on a measure over repeated measurements. An ICC = 1 indicates absolute consistency or stability, whereas an ICC 0 indicates that individuals perform at Ó 2017 Hogrefe Publishing

E. M. Stoakley et al., RSA, Sleep, Affective Style

Table 2. Pearson correlations for Heart Rate (HR) and Respiratory Sinus Arrhythmia (RSA) by waking/sleep stage (n = 25) HR

Wake

RSA

Wake

EST2

SWS

LST2

REM

Wake

EST2

SWS

LST2

REM

–

.718**

.748**

.779**

.657**

–

.517**

.426*

.556**

.514**

–

.918**

.794**

.686**

–

.685**

.589**

.458*

–

.884**

.793**

–

.573**

.437*

–

.838**

–

.616**

EST2 SWS LST2

–

REM

–

Notes. **Correlation is significant at .01 level (two-tailed). *Correlation is significant at .05 level (two-tailed).

Table 3. Correlations between Behavioral Inhibition Scale and Behavioral Activation Scale (BIS-BAS) scores and cardiac measures (HR and RSA) Early stage 2 Subscale

Mean (SD)

SWS

RSA

Late stage 2 RSA

†

RSA

REM HR

RSA

BAS-Drive

11.12 (1.96)

0.498*

0.214

0.345

0.508**

0.148

0.108

0.001

0.203

BAS Fun-Seeking

12.42 (1.56)

0.178

0.303

0.161

0.074

0.161

0.073

0.035

0.031

BAS-Reward Responsiveness

17.84 (1.46)

0.381†

0.508**

0.400*

0.609**

0.221

0.292

0.194

0.157

BIS

18.96 (4.83)

0.345†

0.392†

0.234

0.272

0.057

0.155

0.143

0.012

†

Notes. **Correlation is significant at .01 level (two-tailed). *Correlation is significant at the .05 level (two-tailed). Trend, p .10.

chance on the measure of interest (Bruton, Conway, & Holgate, 2000). It has been suggested that ICC 0.6 is indicative of a reliable measurement tool (Chinn, 1991).

Relations Between Cardiac Measures During Sleep and Waking Affective Style Means and standard deviations for the subscales on the BIS/BAS and Barratt Impulsiveness Scale are reported for the entire sample in Tables 3 and 4, respectively. Separate hierarchical regression analyses were carried out for the Drive, Reward Responsiveness, and Behavioral Inhibition subscales. The BAS Fun-Seeking was negatively skewed due to a single outlier; no significant relations were identified with cardiovascular activity with or without the outlier. In the regression analyses, participant Age and Sex were entered on the first step and RSA in SWS was entered on the second step (see Table 5). The BAS-Drive model approached significance with all predictors entered, F3,21 = 3.2, p = .04, and accounted for 31.4% of the variance in Drive scale scores. On entry into the model at the second step, RSA in SWS was the only significant predictor of BASDrive (t24 = 3.06, p = .006). The BAS-Reward Responsiveness model was significant with all predictors entered, F3,21 = 7.32, p = .002, which accounted for 51.1% of the variance in Reward Responsiveness scale scores. On entry into the model, Sex (t24 = 2.57, p = .02), Age (t24 = 2.20, p = .04), and RSA in SWS (t24 = 2.71, p = .01) were significant predictors. The Behavioral Inhibition Scale (BIS) model was not significant, F3,21 = 0.78, p = .521. Ó 2017 Hogrefe Publishing

Zero-order Pearson correlation coefficients for the BIS/BAS scales with cardiovascular activity are detailed in Table 3. All subscales of the Barratt Impulsiveness Scale were normally distributed, except the Perseverance subscale. Responses on the Perseverance subscale were limited in range and lacked variability and therefore were not transformed or interpreted further. In preliminary regression analyses, there was no influence of Age or Sex, and no interactions with Sex for any of the impulsiveness measures, and thus simple bivariate regressions were carried out to examine the associations between HR and RSA in SWS and the Barratt Impulsiveness Scale Total Score, and the Attention and Self-Control subscales. RSA in SWS accounted for 19.7% of the variance in the Barratt Total score, F1,22 = 5.39, p = .03, 21.4% of the variance in Attention subscale scores, F1,22 = 5.99, p = .02, and 19.4% of the variance in Self-Control subscale scores, F1,22 = 5.30, p = .03. Zero-order correlations between Barratt Impulsiveness scales and cardiovascular activity are reported in Table 4.

Discussion The novel aspects of this study were investigation of the stability of RSA across discrete sleep/wake states and the relation between RSA in sleep and waking affective style. RSA is considered a trait-like measure that is stable in individuals over time and relates to cognitive, affective, and physiological self-regulation (Beauchaine, 2001; Journal of Psychophysiology (2019), 33(1), 1–12

E. M. Stoakley et al., RSA, Sleep, Affective Style

Table 4. Correlations between Barratt Impulsiveness Scale scores and cardiac measures (HR and RSA) Early stage 2

Total score

SWS

RSA

Late stage 2

REM

Mean (SD)

RSA

63.37 (10.60)

0.219

0.054

0.152

0.444*

0.029

0.074

0.090

0.056

10.00 (3.08)

0.206

0.111

0.256

0.463*

0.150

0.123

0.179

0.224 0.091

First-order factors Attention Motor

16.46 (3.54)

0.204

0.224

0.125

0.336

0.037

0.209

0.012

Self-control

12.17 (3.40)

0.106

0.000

0.070

0.440*

0.003

0.113

0.001

0.105

Cognitive complexity

11.58 (2.87)

0.140

0.126

0.000

0.082

0.045

0.017

0.103

0.049

Cognitive instability

6.17 (2.10)

0.027

0.158

0.028

0.112

0.042

0.357†

0.091

0.085

Notes. *Correlation is significant at the .05 level (two-tailed). †Trend, p < .90.

Table 5. Regression statistics for BAS-drive and BAS-reward responsiveness Scale

Step

Drive

BIS

B (SE)

sr2

R2Δ

F change

df1

df2

0.007

0.076

.927

Sex

0.229 (0.82)

0.059

Age

0.083 (0.31)

0.057

RSA in SWS

0.870 (0.28)

0.554

0.307

9.390

.006

Sex

1.277 (0.50)

0.445

0.340

5.675

.010

Age

0.408 (0.19)

0.380

RSA in SWS

0.484 (0.18)

0.413

0.171

7.342

.013

Sex

2.007 (1.96)

0.212

0.056

0.656

.529

Age

0.382 (0.73)

0.108

RSA in SWS

0.805 (0.80)

0.209

0.044

1.015

.325

Porges, 2007). The stability of RSA over time has been established in the waking state (Bornstein & Seuss, 2000; Kleiger et al., 1991; Pitzalis et al., 1996), as well as within sleep states, albeit fewer studies have examined RSA in sleep (Israel et al., 2012; Trinder et al., 2012). Accordingly, we examined the reliability of individual differences in cardiac measures across wake, REM, and NREM sleep in our sample. As predicted, these individual differences were stable between waking and sleep stages. Patterns of cardiac activity were such that HR was higher in the waking state than in any sleep stage, and RSA was lower in SWS than in either early or late night Stage 2 sleep. Higher resting RSA in the waking state is known to be associated with positive emotional style (Oveis et al., 2009; Wang et al., 2013), while lower resting RSA is associated with a range of psychopathology encompassing both internalizing and externalizing disorders (Beauchaine, 2012; Brosschot et al., 2007; Friedman & Thayer, 1998; Mezzacappa et al., 1997; Rottenberg, 2007; Sloan et al., 1994). We therefore expected to observe positive associations between sleep RSA and approach behaviors, and negative associations between sleep RSA and maladaptive behavior (e.g., impulsivity). These relations were present with RSA recorded during early night NREM sleep, particularly deep SWS. As predicted, RSA in early night SWS was associated with higher approach motivation and lower impulsivity. Journal of Psychophysiology (2019), 33(1), 1–12

Cardiac Measures in Waking and Sleep States Consistent with previous reports (Chouchou & Desseilles, 2014; Lanfranchi & Somers, 2011; Tobaldini et al., 2013; Trinder et al., 2012), mean HR was found to be lower during sleep relative to the waking state, and lower in late Stage 2 than in early night NREM sleep, in line with the well-established circadian modulation of HR. In the current sample, HR was generally not higher in REM sleep compared to NREM sleep as others have reported (Stein & Pu, 2012; Trinder et al., 2012); the expected difference may have been present if REM periods had been sampled throughout the night. Also consistent with previous reports, mean RSA was found to be comparable between REM and wake states, and although not significant, RSA appeared to be higher in Stage 2 compared to wake. Most notably, RSA was significantly lower during SWS than Stage 2, whether Stage 2 was sampled from either the early or late part of the night, consistent with reports that RSA varies with sleep stage (Trinder et al., 2001, 2012). RSA has been reported to be relatively constant over time within sleep states (Trinder et al., 2012); however, the difference in RSA between Stage 2 and SWS shown here warrants separation of these NREM states, particularly when investigating relations with waking affective style. The division of NREM states has Ó 2017 Hogrefe Publishing

E. M. Stoakley et al., RSA, Sleep, Affective Style

recently been supported by findings of positive correlations between RSA and NREM delta EEG power that were significantly stronger in the first period of NREM sleep (Rothenberger et al., 2015). The patterns for HR and RSA by sleep state did not differ between men and women. Correlation analyses demonstrated that both HR and RSA values were highly stable across sleep-wake states as hypothesized. As well, intra-class correlation coefficients illustrated that individual RSA measures were consistent in their rank order within the sample across wake and sleep states. Taken together, these data support the utility of cardiac measures during sleep as stable, trait-like variables and establishes their viability as markers of individual differences in affective style. These results expand on previous findings supporting the stability of HRV within the waking state over varying time periods (Bornstein & Seuss, 2000; Kleiger et al., 1991; Pitzalis et al., 1996), as well as stability within the sleep state over time (Israel et al., 2012).

Individual Differences in Affective Style and RSA Associations between positive affective style and higher RSA during sleep in our sample of healthy young adults were consistent with previous research (Brenner, Beauchaine, & Sylvers, 2005; Oveis et al., 2009; Wang et al., 2013) and extend associations between behavior and waking RSA to the context of sleep (cf. Brosschot et al., 2007). Specifically, higher RSA during early night NREM sleep was associated with greater approach motivation and lower overall impulsivity. Consistent with work by Brenner et al. (2005) suggesting a relation between approach motivation and waking RSA, positive relations between self-reported BAS-Drive and BAS-Reward Responsiveness and relatively higher RSA were identified during early night sleep (EST2 and SWS). These findings suggest that individuals with higher RSA have greater approach motivation, which is indicative of a more positive affective style. Approach motivation was positively related to RSA only during sleep, particularly during early night NREM sleep when sleep drive is strongest, a finding that merits further exploration. NREM sleep is a period of quiescence characterized by parasympathetic activation and may provide a more robust estimate of RSA than is possible with waking resting baseline recordings. With respect to impulsivity, higher overall Barratt Impulsivity Scores were inversely associated with RSA during early night SWS, suggesting that RSA, an indicator of effective self-regulation, was related to lower levels of impulsiveness. Closer evaluation of impulsivity scores revealed that the Attentional and Self-control subscales were most strongly related to RSA in SWS, consistent with previous research relating waking RSA to greater executive control Ă&#x201C; 2017 Hogrefe Publishing

via prefrontal neural structures (Beauchaine, 2012; Thayer et al., 2009). The results reported here extend these findings by identifying SWS as an optimal period for measurement of RSA to examine relations with waking affective style. Again, slow wave sleep may represent the best time to sample RSA because of the relative quiescence and parasympathetic dominance characteristic of this stage of sleep (Lanfranchi & Somers, 2011; Trinder et al., 2012). These results may be considered in relation to large population-based studies that have established relations between low resting HR and antisocial behavior. In one large study, resting HR in late adolescence was predictive of criminal convictions in 700,000 Swedish men (Latvala, Kuja-Halkola, Almqvist, Larsson, & Lichtenstein, 2015); another study showed relations between resting HR and both self-reported and official records of criminal activity in 3,600 Brazilian adolescents (Murray et al., 2016). In a recent review, Portnoy and Farrington (2015) presented two theoretical explanations for the association between low resting HR and antisocial behavior (e.g., aggression and psychopathy). The first theory suggests that those who demonstrate low ANS arousal at rest (as indexed by low HR), find their physiological state to be uncomfortable and seek stimulation in order to elevate their arousal level and activate the ANS. This is the stimulation-seeking theory, which is estimated to be more likely than the contrasting theory of fearlessness also proffered by the authors. In the fearlessness theory, low resting HR is interpreted as a lack of fear which affects the developmental processes whereby societal standards for behavior are established in an individual, in part by condition with use of punishments. In our healthy sample of undergraduate students, we would not expect to see a high degree of antisocial or criminal behavior; however, our findings of higher impulsivity and lower approach motivation with lower RSA fit within the larger framework relating low ANS arousal to emotional style and behavioral outcomes.

Limitations and Future Directions There were a few limitations in the study, including the different postures used in the wake (seated) and sleep (lying) ECG recording. In future studies, it would be ideal to use the same posture for both wake and sleep measures of ECG. Another limitation of the current study was the small sample of undergraduate students. Nonetheless, this small, homogenous sample was useful in terms of minimizing confounds due to age, fitness, health factors, socioeconomic factors, etc. Future studies should include a larger sample size and participants from the community in order to ensure the reliability and generalizability of the present findings. In order to clarify the influence of circadian versus homeostatic influences on HR and RSA and the impact of Journal of Psychophysiology (2019), 33(1), 1â&#x20AC;&#x201C;12

these factors on relations between RSA and affective style, sampling HR during multiple periods of SWS during nighttime sleep and daytime napping would prove informative. Given the large number of sleep studies run in clinical settings every night, collaborative efforts to collect these types of data would be both practical and informative.

Conclusion The results from the present study support the stability of individual differences in HR and RSA across sleep-wake states. RSA in early night NREM sleep was predictive of individual differences in waking affective style. Findings suggest that early night NREM sleep may be a particularly opportune time to sample resting RSA. Thus, cardiac measures in NREM sleep appear analogous to waking state measures as markers of individual differences in waking self-regulatory capacity and affective style. Acknowledgments This research was supported by a Natural Sciences and Engineering Research Council (NSERC) of Canada grant awarded to K. A. Cote. Ethics and Disclosure Statements All participants of the study provided written informed consent and the study was approved by the Brock University Research Ethics Board. All authors disclose no actual or potential conflicts of interest including any financial, personal, or other relationships with other people or organizations that could inappropriately influence (bias) their work.

References Allen, J. J. B., Chambers, A. S., & Tower, D. N. (2007). The many metrics of cardiac chronotropy: A pragmatic primer and a brief comparison of metrics. Biological Psychology, 74, 243–262. doi: 10.1016/j.biopsycho.2006.08.005 Appelhans, B. M., & Luecken, L. J. (2006). Heart rate variability as an index of regulated emotional responding. Review of General Psychology, 10, 229–240. doi: 10.1037/1089-2680.10.3.229 Baglioni, C., Battaglises, G., Feige, B., Spiegelhalder, K., Nissen, C., Voderholzer, U., . . . Reimann, D. (2011). Insomnia as a predictor of depression: A meta-analytic evaluation of longitudinal epidemiological studies. Journal of Affective Disorders, 135, 10–19. doi: 10.1016/j.jad.2011.01.011 Beauchaine, T. (2001). Vagal tone, development, and Gray’ motivational theory: Toward an integrated model of autonomic nervous system functioning in psychopathology. Developmental Psychopathology, 13, 183–214. doi: 10.1017/S0954579401002012 Beauchaine, T. (2012). Physiological markers of emotion and behavior dysregulation in externalizing psychopathology. Monographs of the Society for Research in Child Development, 77, 79–86. doi: 10.1111/j.1540-5834.2011.00665.x

Journal of Psychophysiology (2019), 33(1), 1–12

E. M. Stoakley et al., RSA, Sleep, Affective Style

Beauchaine, T., Gatzke-Kopp, L., & Mead, H. K. (2007). Polyvagal theory and developmental psychopathology: Emotion dysregulation and conduct problems from preschool to adolescence. Biological Psychology, 74, 174–184. doi: 10.1016/j.biopsycho. 2005.08.008 Benarroch, E. E. (1993). The central autonomic network: Functional organization, dysfunction and perspective. Mayo Clinic Proceedings, 68, 988–1001. doi: 10.1016/S0025-6196(12) 62272-1 Berntson, G. G., Quigley, K. S., & Lozano, D. (2007). Cardiovascular psychophysiology. In J. T. Cacioppo, L. G. Tassinary, & G. G. Berntson (Eds.), Handbook of Psychophysiology (pp. 182–210). New York, NY: Cambridge University Press. Berntson, G. G., & Stowell, J. R. (1998). ECG artifacts and heart period variability: Don’t miss a beat!. Psychophysiology, 35, 127–132. doi: 10.1111/1469-8986.3510127 Borges, C., Mathewson, K. J., & Schmidt, L. A. (2017). Short-term test-retest reliability of respiratory sinus arrhythmia (RSA) in young adults. Journal of Psychophysiology. doi: 10.1027/02698803/a000187 Bornstein, M. H., & Seuss, P. E. (2000). Physiological selfregulation and information processing in infancy: Cardiac vagal tone and habituation. Child Development, 71, 273–287. doi: 10.1111/1467-8624.00143 Brenner, S. L., Beauchaine, T. P., & Sylvers, P. D. (2005). A comparison of psychophysiological and self-report measures of BAS and BIS activation. Psychophysiology, 42, 108–115. doi: 10.1111/j.1469-8986.2005.00261.x Brosschot, J. F., Van Dijk, E., & Thayer, J. F. (2007). Daily worry is related to low heart rate variability during waking and the subsequent nocturnal sleep period. International Journal of Psychophysiology, 63, 39–47. doi: 10.1016/j.ijpsycho.2006. 07.016 Bruton, A., Conway, J. H., & Holgate, S. T. (2000). Reliability: What is it and how is it measured? Physiotherapy, 86, 94–99. doi: 10.1016/S0031-9406(05)61211-4 Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS/BAS scales. Journal of Personality and Social Psychology, 67, 319–333. doi: 10.1037/ 0022-3514.67.2.319 Chinn, S. (1991). Repeatability and method comparison. Thorax, 46, 454–456. doi: 10.1136/thx.46.6.454 Chouchou, F., & Desseilles, M. (2014). Heart rate variability: A tool to explore the sleeping brain? Frontiers in Neuroscience, 8, 402. doi: 10.3389/fnins.2014.00402 Cote, K. A., McCormick, C. M., Geniole, S. N., Renn, R. P., & MacAulay, S. D. (2013). Sleep deprivation lowers reactive aggression and testosterone in men. Biological Psychology, 92, 249–256. doi: 10.1016/j.biopsycho.2012.09.011 Cote, K. A., Mondloch, C. J., Sergeeva, V., Taylor, M., & Semplonius, T. (2014). Impact of total sleep deprivation on behavioural neural processing of emotionally expressive faces. Experimental Brain Research, 232, 1429–1442. doi: 10.1007/ s00221-013-3780-1 El-Sheikh, M. (2005). Stability of respiratory sinus arrhythmia in children and young adolescents: A longitudinal examination. Developmental Psychobiology, 46, 66–74. doi: 10.1002/dev. 20036 Elsenbruch, S., Hamish, M. J., & Orr, W. C. (1999). Heart rate variability during waking and sleep in healthy males and females. Sleep, 22, 1067–1071. doi: 10.1093/sleep/22.8.1067 Friedman, B. H., & Thayer, J. F. (1998). Autonomic balance revisited: Panic anxiety and heart rate variability. Journal of Psychosomatic Research, 44, 133–151. doi: 10.1016/S00223999(97)00202-X

Ó 2017 Hogrefe Publishing

E. M. Stoakley et al., RSA, Sleep, Affective Style

Grossman, P., & Taylor, E. W. (2007). Toward understanding respiratory sinus arrhythmia: Relations to cardiac vagal tone, evolution and biobehavioral functions. Biological Psychology, 74, 263–285. doi: 10.1016/j.biopsycho.2005.11.014 Houtveen, J. H., Rietveld, S., & de Geus, E. J. (2002). Contribution of tonic vagal modulation of heart rate, central respiratory drive, respiratory depth, and respiratory frequency to respiratory sinus arrhythmia during mental stress and physical exercise. Psychophysiology, 39, 427–436. doi: 10.1017. S0048577202394022 Israel, B., Buysse, D. J., Krafty, R. T., Begley, A., Miewald, J., & Hall, M. (2012). Short-term stability of sleep and heart rate variability in good sleepers and patients with insomnia: For some measures, one night is enough. Sleep, 35, 1285–1291. doi: 10.5665/sleep.2088 Jorm, A. F., Christensen, H., Henderson, A. S., Jacomb, P. A., Korten, A. E., & Rodgers, B. (1999). Using the BIS/BAS scales to measure behavioural inhibition and behavioural activation: Factor structure, validity and norms in a large community sample. Personality and Individual Differences, 26, 49–58. doi: 10.1016/S0191-8869(98)00143-3 Kamphius, J., Meerlo, P., Koolhaas, J. M., & Lancel, M. (2012). Poor sleep as a potential causal factor in aggression and violence. Sleep Medicine, 13, 327–334. doi: 10.1016/j.sleep.2011.12.006 Kleiger, R. E., Bigger, J. T., Bosner, M. S., Chung, M. K., Cook, J. R., Rolnitzky, L. M., . . . Fleiss, J. L. (1991). Stability over time of variables measuring heart rate variability in normal subjects. American Journal of Cardiology, 68, 626–630. doi: 10.1016/ 0002-9149(91)90355-O Lanfranchi, P. A., & Somers, V. K. (2011). Cardiovascular physiology: Autonomic control in health and in sleep disorders. In M. H. Kryger, T. Roth, & W. C. Dement (Eds.), Principles and Practices of Sleep Medicine (5th ed., pp. 226–236). MO: Elsevier. Latvala, A., Kuja-Halkola, R., Almqvist, C., Larsson, H., & Lichtenstein, P. (2015). A longitudinal study of resting heart rate and violent criminality of more than 700 000 men. JAMA Psychiatry, 72, 971–978. doi: 10.1001/jamapsychiatry.2015.1165 Meda, S. A., Stevens, M. C., Potenza, M. N., Pittman, B., Gueorguieva, R., Andrews, M. M., . . . Pearlson, G. D. (2009). Investigating the behavioral and self-report constructs of impulsivity domains using principal component analysis. Behavioral Pharmacology, 20, 390–399. doi: 10.1097/FBP. 0b013e32833113a3 Mezzacappa, E., Tremblay, R. E., Kindlon, D., Saul, J. P., Arseneault, L., Sequin, J., . . . Earls, F. (1997). Anxiety, antisocial behavior and heart rate regulation in adolescent males. Journal of Child Psychology and Psychiatry, 38, 457–469. doi: 10.1111/ j.1469-7610.1997.tb01531.x Murray, J., Hallal, P. C., Mielke, G. I., Raine, A., Wehrmeister, F. C., Anselmi, L., & Barros, F. C. (2016). Low resting heart rate is associated with violence in late adolescence: A prospective birth cohort study in Brazil. International Journal of Epidemiology. Advance online publication. doi: 10.1093/ije/dyv340 Oveis, C., Cohen, A. B., Gruber, J., Shiota, M. N., Haidt, J., & Keltner, D. (2009). Resting respiratory sinus arrhythmia is associated with tonic positive emotionality. Emotion, 9, 265–270. doi: 10.1037/a0015383 Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor structure of the Barratt impulsiveness scale. Journal of Clinical Psychology, 51, 768–774. doi: 10.1002/1097-4679(199511) 51:6<768::AID-JCLP2270510607>3.0.CO;2-1 Pitzalis, M. V., Mastropasqua, F., Massari, F., Forleo, C., Di Maggio, M., Passantino, A., . . . Rizzon, P. (1996). Short- and long-term reproducibility of time and frequency domain heart rate variability measurements in normal subjects. Cardiovascular Research, 32, 226–233. doi: 10.1016/0008-6363(96)00086-7

Ó 2017 Hogrefe Publishing

Porges, S. W. (2007). The polyvagal perspective. Biological Psychology, 74, 116–143. doi: 10.1016/j.biopsycho.2006.06.009 Portnoy, J., & Farrington, D. P. (2015). Resting heart rate and antisocial behavior: An updated systematic review and meta-analysis. Aggression and Violent Behavior, 22, 33–45. doi: 10.1016/j.avb.2015.02.004 Rechtschaffen, A., & Kales, A. (1968). Manual of standardized terminology, techniques, and scoring systems for sleep stages of human subjects. Los Angeles, CA: Brain Information/Brain Research Institute UCLA. Renn, R. P., & Cote, K. A. (2013). Performance monitoring following total sleep deprivation: Effects of task type and error rate. International Journal of Psychophysiology, 88, 64–73. doi: 10.1016/j.ijpsycho.2013.01.013 Rothenberger, S. D., Krafty, R. T., Taylor, B. J., Cribbet, M. R., Thayer, J. F., Buysse, D. J., . . . Hall, M. H. (2015). Time-varying correlations between delta EEG power and heart rate variability in midlife women: The SWAN sleep study. Psychophysiology, 52, 572–584. doi: 10.1111/psyp.12383 Rottenberg, J. (2007). Cardiac vagal control in depression: A critical analysis. Biological Psychology, 74, 200–211. doi: 10.1016/j.biopsycho.2005.08.010 Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin, 86, 420–428. doi: 10.1037/0033-2909.86.2.420 Sloan, R. P., Shapiro, P. A., Bigger, J. T. Jr., Bagiella, E., Steinman, R. C., & Gorman, J. M. (1994). Cardiac autonomic control and hostility in healthy subjects. American Journal of Cardiology, 74, 298–300. doi: 10.1016/0002-9149(94)90382-4 Stein, P. K., & Pu, Y. (2012). Heart rate variability, sleep and sleep disorders. Sleep Medicine Reviews, 16, 47–66. doi: 10.1016/ j.smrv.2011.02.005 Thayer, J. F., Hansen, A. L., Saus-Rose, E., & Johnsen, B. H. (2009). Heart rate variability, prefrontal neural function, and cognitive performance: The neurovisceral integration perspective on self-regulation, adaptation, and health. Annals of Behavioral Medicine, 37, 141–153. doi: 10.1007/s12160-0099101-z Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61, 201–216. doi: 10.1016/S0165-0327(00) 00338-4 Thayer, J. F., & Lane, R. D. (2007). The role of vagal function in the risk for cardiovascular disease and mortality. Biological Psychology, 74, 224–242. doi: 10.1016/j.biopsycho. 2005.11.013 Thayer, J. F., Loerbroks, A., & Sternberg, E. (2011). Inflammation and cardiorespiratory control: The role of the vagus nerve. Respiratory Physiology & Neurobiology, 178, 387–394. doi: 10.1016/j.resp.2011.05.016 Tobaldini, E., Nobili, L., Strada, S., Casali, K. R., Braghiroli, A., & Montano, N. (2013). Heart rate variability in normal and pathological sleep. Frontiers in Physiology, 4, 294. doi: 10.3389/ fphys.2013.00294 Trinder, J., Kleiman, J., Carrington, M., Smith, S., Breen, S., Tan, N., & Kim, Y. (2001). Autonomic activity during human sleep as a function of time and sleep stage. Journal of Sleep Research, 10, 253–264. doi: 10.1046/j.1365-2869.2001.00263.x Trinder, J., Waloszek, J., Woods, M. J., & Jordan, A. S. (2012). Sleep and cardiovascular regulation. Pflugers Arch: European Journal of Physiology, 463, 161–168. doi: 10.1007/s00424-0111041-3 Valladares, E. M., Eljammal, S. M., Motivala, S., Ehlers, C. L., & Irwin, M. R. (2008). Sex differences in cardia sympathovagal balance and vagal tone during nocturnal sleep. Sleep Medicine, 9, 310–316. doi: 10.1016/j.sleep.2007.02.012

Journal of Psychophysiology (2019), 33(1), 1–12

Wang, Z., Lu, W., & Qin, R. (2013). Respiratory sinus arrhythmia is associated with trait positive affect and positive emotional expressivity. Biological Psychology, 93, 190–196. doi: 10.1016/ j.biopsycho.2012.12.006 Received July 28, 2016 Revision received January 21, 2017 Accepted March 8, 2017 Published online September 25, 2017

Journal of Psychophysiology (2019), 33(1), 1–12

E. M. Stoakley et al., RSA, Sleep, Affective Style

Louis A. Schmidt Department of Psychology, Neuroscience & Behaviour McMaster University 1280 Main Street West Hamilton, Ontario L8S 4K1 Canada schmidtl@mcmaster.ca

Ó 2017 Hogrefe Publishing

Article

Rumination Moderates the Association Between Resting High-Frequency Heart Rate Variability and Perceived Ethnic Discrimination DeWayne P. Williams,1 Kinjal D. Pandya,1,2 LaBarron K. Hill,3,4 Andrew H. Kemp,5 Baldwin M. Way,1 Julian F. Thayer,1 and Julian Koenig1 1

Department of Psychology, The Ohio State University, Columbus, OH, USA

Department of Psychology, University of South Carolina, Columbia, SC, USA

Center for the Study of Aging and Human Development, Duke University Medical Center, Durham, NC, USA

Department of Psychiatry, Duke University Medical Center, Durham, NC, USA

Department of Psychology, and Health and Wellbeing Academy, Swansea University, United Kingdom

Abstract: Ethnic discrimination (ED) is both an unfortunate and uncontrollable phenomenon that uniquely impacts African Americans (AAs) and other individuals of ethnic minority status. Perceived ethnic discrimination (PED), defined as the degree to which an individual consciously perceives a negative event as discriminatory and threatening, largely determines the impact that ED can have on target individuals. However, research has not yet considered how individual differences in both emotion regulation abilities, as indexed by resting high-frequency heart rate variability (HF-HRV), and rumination, a maladaptive emotion regulation strategy, may predict PED in AAs. The following investigation examined this relationship in a sample of 101 college-aged students (45 AAs and 56 Caucasian Americans). Resting HF-HRV was assessed via electrocardiogram during a 5-minute-resting period. Rumination was assessed using the ruminative responses scale and everyday PED was assessed using the perceived ethnic discrimination questionnaire. Results showed a significant negative relationship between resting HF-HRV and PED in AAs only. Rumination significantly moderated this relationship, such that lower HF-HRV was related to higher PED only in AAs who reported moderate to higher, β = 0.417 (0.125), p < .01, levels of trait rumination. These results suggest that greater HF-HRV and lesser ruminative tendencies are key factors in reducing PED and therefore possibly, negative consequences associated with ED. Keywords: heart rate variability, perceived ethnic discrimination, rumination, emotion regulation

Ethnic discrimination (ED), defined as the negative treatment of an individual based on their ethnic background, remains a major societal concern and can produce negative outcomes for health in the target group. For example, converging evidence links ED with physiological outcomes such as poorer autonomic function (e.g., blood pressure, BP; Merritt, Bennett, Williams, Edwards, & Sollers, 2006), psychological outcomes such as depression (e.g., Noh & Kaspar, 2003) and self-esteem (Major, Quinton, & Schmader, 2003), and health status such as cardiovascular disease (see Williams & Mohammed, 2009, for review). As ED is both an unfortunate and uncontrollable Ó 2017 Hogrefe Publishing

phenomenon that uniquely impacts African Americans (AAs) and other individuals of ethnic minority status (Kessler, Mickelson, & Williams, 1999; Landrine & Klonoff, 1996), converging evidence suggests that ED is associated with poorer health in these individuals particularly (Pascoe & Smart Richman, 2009; Sellers & Shelton, 2003; Todorova, Falcón, Lincoln, & Price, 2010; Williams & Mohammed, 2009). In comparison to Caucasian American (CAs), AAs are at elevated risk for morbidity and mortality from the leading causes of death in America, including cardiovascular and other diseases (Karlamangla, Merkin, Crimmins, & Seeman, 2010; Journal of Psychophysiology (2019), 33(1), 13–21 https://doi.org/10.1027/0269-8803/a000201

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

Mozaffarian et al., 2016). Given the aforementioned negative impact ED can have on health and well-being in AAs, many propose that ED contributes to such health disparities (see Williams & Mohammed, 2009, for review). Perceived ED (PED) is defined as the degree to which an individual consciously perceives a negative event as discriminatory and threatening (Sellers & Shelton, 2003). It is important to note that ED can be considered detrimental to the target, even if it is not consciously perceived (Allison, 1998; Clark, Anderson, Clark, & Williams, 1999; Landrine & Klonoff, 1996; Sellers & Shelton, 2003). For example, one study demonstrated increased blood pressure in AAs following manipulated ED under both blatant (explicit and conscious) and subtle (ambiguous and unconscious) experimental conditions (Merritt et al., 2006). Nevertheless, everyday PED may be characterized by individual differences independent of ED, that is, the same ED event may be perceived as either threatening or nonthreatening depending on the individual (i.e., more or less PED; Sellers & Shelton, 2003). In this regard, research has primarily focused on social psychological factors as individual differences in PED, such as racial identity and stigma sensitivity (see Major et al., 2003). Interestingly, Berger and Sarnyai (2015) reviewed articles that provided both direct and indirect evidence that chronic exposure to ED may impair executive brain region (e.g., the prefrontal cortex; PFC) function. Executive brain regions, particularly the PFC, are responsible for proper emotion regulation (ER), defined as a process by which individuals can modify their emotional experiences and expressions (for review, see Etkin, Egner, & Kalisch, 2011; Lane et al., 2009). Thus, the researchers proposed that the negative impact ED can have on executive brain region function may lead to a subsequent heightened stress response for additional ED (i.e., PED) or other general threat (Berger & Sarnyai, 2015). Therefore, given the role of executive function in regulating emotions, it would be important to consider how ER abilities, as determined by executive brain function, may serve as an individual difference factor in PED.

Vagally Mediated Heart Rate Variability as a Psychophysiological Indicator of Emotion Regulation Abilities A key mechanism for successful ER is inhibitory control – individuals must inhibit inappropriate emotional responses and instead encourage more acceptable, appropriate, and desirable ones (Lane et al., 2009; Thayer, Åhs, Fredrikson, Sollers, & Wager, 2012). Executive brain regions including

Journal of Psychophysiology (2019), 33(1), 13–21

the PFC exert an inhibitory influence on subcortical brain structures such as the amygdala, allowing the individual to adaptively respond to demands from the environment, and organize their emotional and behavioral responses effectively (Etkin et al., 2011; Lane et al., 2009). These core sets of brain structures are also structurally and functionally linked with autonomic nervous system (ANS) regulation. The ANS dually innervates peripheral organs including the heart, and in a resting state, ANS influence is characterized by a relative dominance of the parasympathetic nervous system (PNS) over influences of the sympathetic nervous system (SNS; Thayer & Lane, 2009; Thayer et al., 2012). PNS activity is thought to reflect executive brain activity, whereas SNS activity is thought to reflect amygdala activity (see Thayer et al., 2012, for review). The vagus nerve is the primary nerve of the PNS responsible for regulating physiological functions (e.g., immune, inflammatory, and cardiac function; Thayer & Sternberg, 2006; Weber et al., 2010) via inhibitory control. Therefore, resting high-frequency heart rate variability (HF-HRV), defined as variability between heartbeats mediated by the vagus, is considered an index of both (cardiac) PNS activity and executive brain function (Thayer et al., 2012), in addition to overall ER abilities. This idea is not without behavioral evidence, as many studies have linked decreased resting HF-HRV with poorer ER (e.g., Appelhans & Luecken, 2006; Melzig, Weike, Hamm, & Thayer, 2009; for review, see Thayer & Lane, 2009; Williams et al., 2015). Overall, resting HF-HRV is a measure of the degree to which the brain’s “integrative” system for adaptive regulation provides flexible control over both the periphery (Thayer et al., 2012) and self-regulatory processes (e.g., ER; Kemp & Quintana, 2013). Understanding the relationship between ED/PED and resting HF-HRV is both warranted and important, but this relationship has not been studied extensively. A handful of investigations have shown the impact of experimentally manipulated ED on phasic changes in HF-HRV, having shown decreased HF-HRV in individuals following the experience of ED (e.g., Hoggard, Hill, Gray, & Sellers, 2015; Neblett & Roberts, 2013). However to our knowledge, only one study has examined the direct relationship between resting HF-HRV and everyday PED (Hill et al., 2017). This study showed higher self-reported PED was associated with lower resting HF-HRV, concluding that repeated exposure of ED may lead to decreased PNS activity overtime (Hill et al., 2017). However, research has not yet considered how resting HF-HRV, as an index of ER abilities, potentially influences the likelihood that an individual perceives everyday negative events as both discriminatory and threatening (i.e., PED).

Ó 2017 Hogrefe Publishing

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

Vagally Mediated Heart Rate Variability, Perceived Ethnic Discrimination, and Rumination Rumination is a factor thought to prolong the negative impact ED can have on physiological arousal and psychological distress, particularly in AAs (Bennett, Merritt, Edwards, & Sollers, 2004; Merritt et al., 2006). Rumination can be defined as the perseverative thinking over stressors, and often predicts negative mental states such as depression and anxiety – making rumination a largely maladaptive coping strategy (Nolen-Hoeksema, Wisco, & Lyubomirsky, 2008). However, little research has considered how the tendency to employ particular ER strategies such as rumination may influence everyday PED. Given the definition of PED (i.e., a past perception of ED), it is possible that individuals with a general tendency to ruminate, may create a “running dialog” associated with their experiences of ED and thus, the negative threat of ED can remain subjectively active/present (i.e., increased PED). Additionally, individuals with lower resting HF-HRV typically employ poorer ER strategies when regulating negative emotions compared to those with higher resting HF-HRV (see Brosschot, Gerin, & Thayer, 2006, for review); indeed, rumination is considered a poor ER strategy characteristic of individuals with lower ER abilities, as indexed by lower resting HF-HRV (Brosschot, et al., 2006). However, ER abilities and strategies are conceptually different; ER strategies are thought to be context dependent, that is, some strategies may be more or less adaptive depending on both the individual (e.g., abilities) and the environment (e.g., motivations; Aldao & Nolen-Hoeksema, 2012). In contrast, ER ability is a more stable factor across situations (Thayer & Lane, 2000), and thus it would be helpful to understand how both ER abilities and strategies interact to determine PED. However to date, no study has examined how an individual’s trait rumination can alter or moderate the association between resting HF-HRV and everyday PED.

The Present Study Research on the relationship between resting HF-HRV and PED is warranted as to our knowledge, only one other study has investigated this link (Hill et al., 2017). From an ER perspective, research has yet to consider resting HF-HRV as an individual difference factor in everyday PED. Furthermore, it would be important to investigate if trait rumination, a maladaptive ER strategy, moderates the link between ER abilities, as indexed by resting HF-HRV, and everyday PED. Thus, the present study sought to both (i) replicate previous findings that showed a negative association between resting HF-HRV and everyday Ó 2017 Hogrefe Publishing

self-reported PED (Hill et al., 2017) and (ii) assess how trait rumination may moderate this relationship. We hypothesized that (i) resting HF-HRV would be inversely related to PED, such that AAs with lower HF-HRV would report higher PED and that (ii) this relationship would be moderated by rumination, such that this relationship should be strongest in AAs who report greater trait rumination. We expected to observe no meaningful relationships in CAs. These hypothesized results would suggest that both ER abilities (i.e., HF-HRV) and ER strategies (i.e., rumination) are important individual difference factors in reducing PED and thereby potentially minimizing the impact of actual ED. Finally, directionality is important, as we conceptualize resting HF-HRV as the independent variable rather than an outcome (dependent) variable as in previous work (Hill et al., 2017). Therefore, we also test the reverse of our proposed moderation model above, with PED as the independent variable, rumination as the moderating variable, and resting HF-HRV as the outcome variable.

Methods General Procedure We recruited 101 college-aged individuals (45 AAs, 72 female, mean age = 19.48, SD = 2.26). The experiment was conducted at the Emotions and Quantitative Psychophysiology (EQP) Lab at the Ohio State University. Subjects were recruited from the Research Experience Program (REP) pool at The Ohio State University, allowing students to participate in research for partial class credit in an introductory level psychology course. Participants outside of the REP pool were also recruited and paid for their participation. We asked all participants not to smoke, undergo vigorous physical activity, or drink caffeine 6 hr prior to the experiment. The study was approved by the institutional review board, and all participants signed written informed consent. All participants were placed in a soundproof experimental room, equipped with a camera and a microphone for safety and instructional reasons, and a high definition TV for stimulus presentation. Participants were given a detailed explanation of the procedures that would take place without indicating the specific hypothesis under study or manipulations applied. Electrocardiogram (ECG) leads were attached to the subjects and while in a separate control room, the experimenter led the subjects to the initial phases of the experiment. First, participants completed a 5-minuteresting baseline period, where participants sat in a resting (spontaneous breathing) position, and viewed a blank gray screen. Following this period, participants completed a series of self-report questionnaires. Journal of Psychophysiology (2019), 33(1), 13–21

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

Vagally Mediated Heart Rate Variability Cardiac data was recorded continuously throughout each experiment via a 3-lead electrocardiogram (ECG) at a 1,000 Hz sampling rate using a MindwareTM 2000D (MW2000D, Ó 2009–2017 MindWare Technologies) Impedance Cardiograph package. Electrodes were placed (1) below the right clavicle, (2) on the left side of the abdomen (below the heart), and (3) on the right side of the abdomen. Variability between R-spikes in milliseconds was collected to calculate baseline HF-HRV for the full 5-minute-resting period. Participants’ successive interbeat intervals (IBIs; in milliseconds) were extracted using MindwareTM HRV Analysis software. IBIs were written in a text file and analyzed using Kubios HRV analysis package 2.0 (Tarvainen, Niskanen, Lipponen, Ranta-Aho, & Karjalainen 2014), allowing for the calculation of frequency-domain indices of resting HF-HRV. Artifacts within the R-to-R series were visually detected, and we applied an artifact correction level that would differentiate and remove artifacts (differing abnormal IBIs from the mean IBI; smoothing priors as a detrend method; see Tarvainen et al., 2014, for review) using a piecewise cubic spline interpolation method. Autoregressive estimates were calculated, yielding high-frequency power HRV (HF-HRV, 0.15–0.4 Hz; Thayer, Hansen, & Johnsen, 2010; Task Force of the European Society of Cardiology, 1996). HF-HRV is a reliable and valid measure of cardiac vagal activity (i.e., HF-HRV; Thayer et al., 2010). HF-HRV values were natural log transformed (ln) to fit assumptions of linear analyses. Additionally, high-frequency peak (HF peak) values were obtained from the autoregressive analysis as a measure of respiration rate to control for potential bias (Thayer, Sollers, Ruiz-Padial, & Vila, 2002).

Self-Report Questionnaires Perceived Ethnic Discrimination Perceived Ethnic Discrimination (PED) was assessed using the Perceived Ethnic Discrimination Questionnaire (PEDQ). The PEDQ is a 17-item questionnaire designed to assess subjective feelings of everyday ED (i.e., PED) and contains four subscales, including discrimination via exclusion and rejection (PED-Exclusion; α = .70; source of reliability from the current data), discrimination via stigmatization and/or devaluation (PED-Stigma; α = .74), discrimination at work and/or school (PED-Work; α = .69), and discrimination via threat and/or aggression (PED-Threat; α = .80; Brondolo et al., 2005). Participants rate the frequency with which they experienced particular situations (sample item: “Because of your ethnicity, have others threatened to hurt you”) from 1 (= never) to Journal of Psychophysiology (2019), 33(1), 13–21

7 (= very often). Within the current sample, the PEDQ showed good overall internal consistency (α = .87). Trait Rumination Rumination was assessed using the 22-item Ruminative Responses Scale (RRS; Treynor, Gonzalez, & NolenHoeksema, 2003). Participants answered on a scale from 1 (= almost never) to 4 (= almost always; sample item: “How often do you think about how alone you feel”), with higher values representing higher trait rumination (Cronbach’s α = .92).

Statistics All statistical tests were conducted using SPSS (version 19, IBM Chicago, IL). Independent sample t-tests were also used to examine differences between CAs and AAs on all psychological and physiological variables. Split by ethnicity, Pearson’s r zero-order correlation coefficients were calculated in order to illustrate the relationships between all variables. An SPSS macro PROCESS was used (Hayes, 2012) to test if rumination moderated the relationship between resting HF-HRV and PEDQ scores in AAs only, as we did not expect to observe a significant relationship in CAs. In the program PROCESS, “Model 1” was used to test a main effect of the independent variable (IV; resting HF-HRV), a main effect of the moderator (M; RRS scores), and an interaction effect of the two on the dependent variable (DV; PEDQ scores). We also tested an alternative version of this model that includes PEDQ scores as the IV, RRS scores as the M, and resting HF-HRV as the DV (see Figure 1A for hypothesized moderation model, and Figure 1B for alternative moderation model). If the 2-way interaction is significant, it suggests that the relationship between the IV and DV changes at different levels of M (see Hayes, 2012, for review). The nature of the interaction was determined using PROCESS’ conditional effects, that is, how the IV-DV relationship changes at different levels of M and W. High and low values for the predictor variables are derived using ±1 SD from the mean, allowing the program to yield predicted values of the DV at varying levels of the predictor variables via regions of significance and simple slope analyses (Hayes, 2012). Statistics reported include unstandardized beta (β) coefficients, standard errors (SE; in parentheses), 95% confidence intervals, partial correlation coefficients (for interactions), and p values. Lastly, potential covariates of resting HF-HRV were controlled for in each model. These variables included respiration rate (HF peak values; Thayer et al., 2002), sex (Koenig & Thayer, 2016), body mass index (BMI; Koenig et al., 2014), and age (Jensen-Urstad et al., 1997). All tests were two-tailed and significance levels were evaluated using an α of .05. Ó 2017 Hogrefe Publishing

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

(A)

(B)

Figure 1. Conceptual proposed moderation model. (A) Hypothesized model: The independent variable is regarded as resting high-frequency heart rate variability (HF-HRV; natural log transformed), the moderator as rumination (Ruminative Responses Scale [RRS] scores), and the dependent variable as perceived ethnic discrimination (PED; indexed by Perceived Ethnic Discrimination Questionnaire [PEDQ] scores). (B) Alternative model: The independent variable is regarded as resting PED, the moderator as rumination, and the dependent variable as HF-HRV.

Table 1. Means and standard deviations of all variables split by ethnicity n

Age

BMI

HF-HRV

PED

Rumination

19.82 (2.48)

25.16 (5.42)

76.21 (8.16)

.27 (0.04)

6.74 (.85)

30.02 (7.29)

41.82 (11.63)

19.20 (2.05)

23.92 (3.16)

72.03 (9.85)

.26 (0.05)

6.57 (1.05)

20.36 (4.05)

41.61 (10.65)

.168

.155

.024

Respiration

.112

.370

.001

.924

Notes. This table shows mean (standard deviation in parentheses) values on baseline measures split between CAs and AAs. Age was calculated in years, heart rate (HR) in beats per minute, body mass index (BMI) was calculated in kg/m2, and natural log high-frequency heart rate variability (HF-HRV) was calculated in ms2. Perceived ethnic discrimination PED was indexed using the Perceived Ethnic Discrimination Questionnaire (PEDQ) with higher scores reflecting higher PED. Trait Rumination was indexed using the Ruminative Response Scales (RRS), with higher scores reflecting higher trait rumination (significant p values bolded).

Results Descriptive statistics, including age, BMI, baseline HR, baseline HF-HRV, PED, and rumination split by ethnicity, are given in Table 1. The AA sample showed significantly higher PED in comparison to CAs, t(99) = 8.44, p < .001, and greater resting HR, t(99) = 2.28, p < .05, but did not differ significantly on any other variable (Table 1). Within the AA group, results showed a significant negative association between resting HF-HRV and total PED scores (r = .303, p = .041). Subscale results revealed a significant negative relationship between HF-HRV and PED-Stigma (r = .402, p < .01) while the other subscales were not significant, but trending in the same direction (PED-Exclusion: r = .241, p = .107; PED-Work: r = .197, p = .190; PED-Threat: r = .142, p = .246). Total rumination was significantly positively associated with total PED (r = .299, p = .025). Total rumination was also significantly positively associated with PED-Threat (r = .442, p = .002). In CAs, no significant relationship between HF-HRV and PED (including all subscales) was found. Additionally, no relationship between PED and total rumination was found in CAs (refer to Table 2 for correlations between all variables in both AAs and CAs; Figure 2). Controlling for aforementioned covariates, moderation results showed that rumination significantly moderated Ó 2017 Hogrefe Publishing

the relationship between resting HF-HRV and PED in the hypothesized model [Figure 1A; β = 0.26 (.12), rpartial = .350, p = .028]. Conditional effects analyses showed a significant relationship between resting HF-HRV and PED in AAs with higher [β = 5.16 (1.61), p = .003] to moderate [β = 2.37 (1.14), p = .04] levels of trait rumination, but not in those with lower trait rumination [β = 0.42 (1.74), p = .813] suggesting that AA individuals with lower trait rumination report similar levels of PED despite levels of resting HF-HRV (Figure 3). Likewise, AA individuals with higher resting HF-HRV report similar levels of PED despite levels of trait rumination [β = 0.03 (0.14), p = .835]. Moderation tests were not significant using the alternative model presented in Figure 1B [β = 0.002 (0.001), rpartial = .228, p = .164; Figure 3].

Discussion The current investigation sought to examine the relationship between resting HF-HRV, a psychophysiological index of ER abilities, and PED in AAs. Additionally, we sought to investigate how ruminative tendencies may moderate this association. In line with an earlier report (Hill et al., 2017), there was a significant and negative association between resting HF-HRV and PED in AAs but not CAs. Journal of Psychophysiology (2019), 33(1), 13–21

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

Table 2. Correlation coefficients between variables split by ethnicity 1

African Americans 1. HF-HRV

–

2. Rumination

.188

–

3. PED-total

.342*

.299*

–

4. PED-exclusion

.258

.018

.711**

5. PED-stigma

.419**

.095

.745**

.549**

–

6. PED-work

.213

.291

.849**

.507**

.443**

–

7. PED-threat

.193

.442**

.713**

.151

.391**

.536**

–

Caucasian Americans 1. HF-HRV

–

2. Rumination

.045

–

3. PED-total

.173

.084

–

4. PED-exclusion

.057

.217

.605**

5. PED-stigma

.203

.299*

.670**

.595**

6. PED-work

.048

.090

.598**

.101

.177

–

7. PED-threat

.184

.063

.819**

.293*

.317*

.422**

– – –

Notes. HF-HRV = high-frequency heart rate variability (natural log transformed); Rumination = indexed using the ruminative responses scale; PEDtotal = perceived ethnic discrimination total scores; PED-exclusion = discrimination via exclusion subscale; PED-stigma = discrimination via stigma subscale; PED-work = discrimination at work/school subscale; PED-threat = discrimination via threat/aggression subscale. *p < .05; **p < .01. Bolded p values represent significant correlation coefficients.

(A)

(B)

Figure 2. Scatterplot of resting HF-HRV and perceived ethnic discrimination. (A) Scatterplot between resting natural log transformed (ln) highfrequency heart rate variability (HF-HRV) and Perceived Ethnic Discrimination Questionnaire (PEDQ). This correlation was significant in African American participants only (r = .303, p < .05). (B) Correlation between PEDQ and Ruminative Response Scale (RRS) scores (r = .299, p < .05).

Results also showed a significant negative association between trait rumination and PED in AAs only. Subscale analyses showed resting HF-HRV to be most related to the perception of discrimination via stigmatization as indicated by the respective subscale (PED-Stigma), however all subscales’ correlations trended (although not significant) in a negative direction. Importantly, trait rumination significantly moderated the association between resting HF-HRV

Journal of Psychophysiology (2019), 33(1), 13–21

and PED, such that this relationship was only significant in AA individuals with moderate to higher levels of trait rumination. AAs with both lower resting HF-HRV and higher trait rumination showed higher PED compared to all other AAs. Overall, these data both (i) support the link between resting HF-HRV and PED in AAs, and (ii) present trait rumination as an important moderating factor in this association.

Ó 2017 Hogrefe Publishing

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

Figure 3. Conditional effects of rumination as a moderation variable. This figure represents the conditional effects of rumination on the association between HF-HRV and Perceived Ethnic Discrimination Questionnaire (PEDQ) scores. Higher and lower estimates of resting natural log transformed (ln) high-frequency heart rate variability derived from ±1 SD from the mean (see Methods section for details). Those who scored lower on the Ruminative Responses Scale (RRS) showed no association between resting HF-HRV and PEDQ scores. However, resting HF-HRV was significantly associated with PEDQ scores in those with higher trait rumination, such that lower resting HF-HRV was associated with greater PEDQ scores. AAs with lower HFHRV and higher trait rumination reported the highest PED scores.

Implications It is important to note that moderation tests were only significant with resting HF-HRV as the independent variable and PED as the dependent variable (hypothesized model; Figure 1A), and not vice versa (alternative model; Figure 1B). This lends direct support for our novel idea that ER abilities, as indexed by resting HF-HRV, may differentiate AA individuals in everyday PED. Nevertheless, evidence has shown that following experimentally manipulated ED, AAs can also show decreased HF-HRV (e.g., Neblett & Roberts, 2013). Therefore, it is plausible to consider that in an environment where ED often occurs (Sellers & Shelton, 2003), repeated exposure may diminish the integrity of executive brain regions necessary to inhibit the effects of further ED or threat more generally (Berger & Sarnyai, 2015). Such decrements may lead to lower resting HF-HRV in AAs over time (Hill et al., 2017; Hoggard et al., 2015; Neblett & Roberts, 2013). Finally, as we suggest in the current report, chronic lower resting HF-HRV,

Ó 2017 Hogrefe Publishing

especially when coupled with rumination, may further exaggerate PED in AAs thereby perpetuating a detrimental cycle of stress and disease. Therefore, we propose that resting HF-HRV and/or trait rumination are potential “first step” factors in minimizing both the impact of PED on psychophysiological processes, and psychophysiological processes on PED. From a health standpoint, it is interesting to consider research showing that in comparison to CAs, AAs often show greater total peripheral resistance (TPR) and decreased BP at rest. However, a recent meta-analysis by our group showed that AAs have higher resting HF-HRV compared to CAs (Hill et al., 2015) – a paradoxical pattern that we named the “cardiovascular conundrum.” Here, we proposed that greater HF-HRV in AAs serves as a compensatory mechanism, such that AAs may need more ER abilities, and thus higher HF-HRV, to compensate for unique day-to-day stressors such as ED. In support of this idea a recent investigation showed that in 11,989 individuals, Black Brazilians showed greater resting HF-HRV in comparison to both White Brazilians and mixed (Brown Brazilians) individuals, and this relationship was mediated by experiences of ED (i.e., darker skin tone associated with greater experiences of ED associated with higher resting HF-HRV; Kemp et al., 2016). Whereas Kemp et al. (2016) showed ED to be associated with higher resting HF-HRV between ethnic groups, the current results showed that within AAs only, greater PED was associated with lower resting HF-HRV but only in those AAs with a ruminative coping strategy. Overall, prior work suggests that ED serves as a mechanism underlying differences in resting HF-HRV between, and the current study suggests that higher resting HF-HRV within the AA group is especially important in minimizing PED. Limitations and Future Directions One major limitation of the current investigation is that it is correlational and thus, causation cannot be determined. Future research should use longitudinal techniques in an attempt to better understand causality. A second limitation of the current study is that socioeconomic status (SES) information was not collected. SES is proposed to be an influential variable in the experience of ED and thus, future studies should examine the current relationship while considering SES. A third limitation of the current study is that the sample consists of apparently healthy, young students. While we were able to provide evidence for an association of PED and HRV in this sample, future studies should examine this relationship on those with preexisting health problems and older subjects in general. Finally, although we required participants not to smoke, undergo vigorous physical activity, or drink caffeine 6 hr prior to

Journal of Psychophysiology (2019), 33(1), 13–21

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

the experiment, we did not verify that participants complied, and future investigations should ensure that this information is collected and considered.

Conclusions The present study is the first to suggest that lower resting HF-HRV and trait rumination interact to negatively influence PED in AAs. We do not propose that higher resting HF-HRV and/or lower trait rumination can solve the core issues associated with ED. We are, however, proposing that these factors are of particular importance in AAs, as lower PED in a society where ED often occurs may potentially buffer the negative consequences of ED on health and well-being (Pascoe & Smart Richman, 2009; Sellers & Shelton, 2003).

Acknowledgments This research was supported by funding from The Ohio State University College of Social and Behavioral Sciences, The Ohio State University Graduate School, and The Ohio State University College of Social, Behavioral and Economic Sciences to the first author (D. P. Williams) and second author (K. D. Pandya). A. H. Kemp and J. F. Thayer would also like to acknowledge the financial support of FAPESP, a Brazilian research funding institution in the state of São Paulo, and that of Ohio State University, which has helped to initiate collaborative activities between the authors of the current manuscript. This research was supported by funding from the National Institute of Aging (5T32AG000029) and the National Heart, Lung, and Blood Institute of the National Institutes of Health (R01HL121708) to L. K. Hill.

References Aldao, A., & Nolen-Hoeksema, S. (2012). The influence of context on the implementation of adaptive emotion regulation strategies. Behaviour Research and Therapy, 50, 493–501. doi: 10.1016/j.brat.2012.04.004 Allison, K. W. (1998). Stress and oppressed social category membership. In J. Swim & C. Stangor (Eds.), Prejudice: The target’s perspective (pp. 149–170). New York, NY: Academic Press. doi: 10.1016/B978-012679130-3/50042-9 Appelhans, B. M., & Luecken, L. J. (2006). Heart rate variability as an index of regulated emotional responding. Review of General Psychology, 10, 229. doi: 10.1037/1089-2680.10.3.229 Bennett, G. G., Merritt, M. M., Edwards, C. L., & Sollers, J. J. (2004). Perceived racism and affective responses to ambiguous interpersonal interactions among African American men. American Behavioral Scientist, 47, 963–976. doi: 10.1177/ 0002764203261070

Journal of Psychophysiology (2019), 33(1), 13–21

Berger, M., & Sarnyai, Z. (2015). “More than skin deep”: Stress neurobiology and mental health consequences of racial discrimination. Stress, 18, 1–10. doi: 10.3109/10253890.2014. 989204 Brondolo, E., Kelly, K. P., Coakley, V., Gordon, T., Thompson, S., Levy, E., . . . Contrada, R. J. (2005). The perceived ethnic discrimination questionnaire: Development and preliminary validation of a community version. Journal of Applied Social Psychology, 35, 335–365. doi: 10.1111/j.1559-1816.2005. tb02124.x Brosschot, J. F., Gerin, W., & Thayer, J. F. (2006). The perseverative cognition hypothesis: A review of worry, prolonged stressrelated physiological activation, and health. Journal of Psychosomatic Research, 60, 113–124. doi: 10.1016/ j.jpsychores.2005.06.074 Clark, R., Anderson, N. B., Clark, V. R., & Williams, D. R. (1999). Racism as a stressor for African Americans: A biopsychosocial model. The American Psychologist, 54, 805. doi: 10.1037/0003066X.54.10.805 Etkin, A., Egner, T., & Kalisch, R. (2011). Emotional processing in anterior cingulate and medial prefrontal cortex. Trends in Cognitive Sciences, 15, 85–93. doi: 10.1016/j.tics.2010.11.004 Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling [white paper]. Retrieved from http://www. afhayes.com/public/process2012.pdf Hill, L. K., Hoggard, L. S., Richmond, A. S., Gray, D. L., Williams, D. P., & Thayer, J. F. (2017). Examining the association between perceived discrimination and heart rate variability in African Americans. Cultural Diversity and Ethnic Minority Psychology, 23, 5. doi: 10.1037/cdp0000076 Hill, L. K., Hu, D. D., Koenig, J., Sollers, J. J. III, Kapuku, G., Wang, X., . . . Thayer, J. F. (2015). Ethnic differences in resting heart rate variability: A systematic review and metaanalysis. Psychosomatic Medicine, 77, 16–25. doi: 10.1097/ PSY.0000000000000133 Hoggard, L. S., Hill, L. K., Gray, D. L., & Sellers, R. M. (2015). Capturing the cardiac effects of racial discrimination: Do the effects “keep going? International Journal of Psychophysiology, 97, 163–170. doi: 10.1016/j.ijpsycho.2015.04.015 Jensen-Urstad, K., Storck, N., Bouvier, F., Ericson, M., Lindbland, L. E., & Jensen-Urstad, M. (1997). Heart rate variability in healthy subjects is related to age and gender. Acta Physiologica Scandinavica, 160, 235–241. doi: 10.1046/j.1365-201X.1997. 00142.x Landrine, H., & Klonoff, E. A. (1996). The schedule of racist events: A measure of racial discrimination and a study of its negative physical and mental health consequences. Journal of Black Psychology, 22, 144–168. doi: 10.1177/00957984960222002 Karlamangla, A. S., Merkin, S. S., Crimmins, E. M., & Seeman, T. E. (2010). Socioeconomic and ethnic disparities in cardiovascular risk in the United States, 2001–2006. Annals of Epidemiology, 20, 617–628. doi: 10.1016/j.annepidem.2010. 05.003 Kemp, A. H., Koenig, J., Thayer, J. F., Bittencourt, M. S., Pereira, A. C., Santos, I. S., . . . Benseñor, I. M. (2016). Race and restingstate heart rate variability in Brazilian civil servants and the mediating effects of discrimination: An ELSA-Brasil cohort study. Psychosomatic Medicine, 78, 950–958. doi: 10.1097/ PSY.0000000000000359 Kemp, A. H., & Quintana, D. S. (2013). The relationship between mental and physical health: Insights from the study of heart rate variability. International Journal of Psychophysiology, 89, 288–296. doi: 10.1016/j.ijpsycho.2013.06.018 Kessler, R. C., Mickelson, K. D., & Williams, D. R. (1999). The prevalence, distribution, and mental health correlates of

Ó 2017 Hogrefe Publishing

D. P. Williams et al., Resting Heart Rate Variability, Rumination, and Ethnic Discrimination

perceived discrimination in the United States. Journal of Health and Social Behavior, 40, 208–230. Koenig, J., Jarczok, M. N., Warth, M., Ellis, R. J., Bach, C., Hillecke, T. K., & Thayer, J. F. (2014). Body mass index is related to autonomic nervous system activity as measured by heart rate variability – a replication using short term measurements. The Journal of Nutrition, Health & Aging, 18, 300–302. doi: 10.1007/ s12603-014-0022-6 Koenig, J., & Thayer, J. F. (2016). Sex differences in healthy human heart rate variability: A meta-analysis. Neuroscience & Biobehavioral Reviews, 64, 288–310. doi: 10.1016/j.neubiorev. 2016.03.007 Lane, R. D., McRae, K., Reiman, E. M., Chen, K., Ahern, G. L., & Thayer, J. F. (2009). Neural correlates of heart rate variability during emotion. Neuroimage, 44, 213–222. doi: 10.1016/ j.neuroimage.2008.07.056 Major, B., Quinton, W. J., & Schmader, T. (2003). Attributions to discrimination and self-esteem: Impact of group identification and situational ambiguity. Journal of Experimental Social Psychology, 39, 220–231. doi: 10.1016/S0022-1031(02)00547-4 Melzig, C. A., Weike, A. I., Hamm, A. O., & Thayer, J. F. (2009). Individual differences in fear-potentiated startle as a function of resting heart rate variability: Implications for panic disorder. International Journal of Psychophysiology, 71, 109–117. doi: 10.1016/j.ijpsycho.2008.07.013 Merritt, M. M., Bennett, G. G. Jr., Williams, R. B., Edwards, C. L., & Sollers, J. J. III (2006). Perceived racism and cardiovascular reactivity and recovery to personally relevant stress. Health Psychology, 25, 364. doi: 10.1037/0278-6133.25.3.364 Mozaffarian, D., Benjamin, E. J., Go, A. S., Arnett, D. K., Blaha, M. J., Cushman, M., . . . Howard, V. J. (2016). Heart disease and stroke statistics – 2016 Update. Circulation, 133, e38–e360. doi: 10.1161/CIR.0000000000000366 Neblett, E. W., & Roberts, S. O. (2013). Racial identity and autonomic responses to racial discrimination. Psychophysiology, 50, 943–953. doi: 10.1111/psyp.12087 Noh, S., & Kaspar, V. (2003). Perceived discrimination and depression: Moderating effects of coping, acculturation, and ethnic support. American Journal of Public Health, 93, 232–238. doi: 10.2105/AJPH.93.2.232 Nolen-Hoeksema, S., Wisco, B. E., & Lyubomirsky, S. (2008). Rethinking rumination. Perspectives on Psychological Science, 3, 400–424. Pascoe, E. A., & Smart Richman, L. (2009). Perceived discrimination and health: A meta-analytic review. Psychological Bulletin, 135, 531–554. doi: 10.1037/a0016059 Sellers, R. M., & Shelton, J. N. (2003). The role of racial identity in perceived racial discrimination. Journal of Personality and Social Psychology, 84, 1079. doi: 10.1037/0022-3514.84.5.1079 Tarvainen, M. P., Niskanen, J. P., Lipponen, J. A., Ranta-Aho, P. O., & Karjalainen, P. A. (2014). Kubios HRV – heart rate variability analysis software. Computer Methods and Programs in Biomedicine, 113, 210–220. doi: 10.1016/j.cmpb.2013.07.024 Task Force of the European Society of Cardiology. (1996). Heart rate variability standards of measurement, physiological interpretation, and clinical use. European Heart Journal, 17, 354–381. Thayer, J. F., Åhs, F., Fredrikson, M., Sollers, J. J., & Wager, T. D. (2012). A meta-analysis of heart rate variability and neuroimaging studies: Implications for heart rate variability as a marker of

Ó 2017 Hogrefe Publishing

stress and health. Neuroscience & Biobehavioral Reviews, 36, 747–756. doi: 10.1016/j.neubiorev.2011.11.009 Thayer, J. F., Hansen, A. L., & Johnsen, B. H. (2010). The noninvasive assessment of autonomic influences on the heart using impedance cardiography and heart rate variability. In A. Steptoe (Ed.), Handbook of behavioral medicine (pp. 723–740). New York, NY: Springer. doi: 10.1007/978-0-387-09488-5_47 Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61, 201–216. doi: 10.1016/S0165-0327(00) 00338-4 Thayer, J. F., & Lane, R. D. (2009). Claude Bernard and the heart – brain connection: Further elaboration of a model of neurovisceral integration. Neuroscience & Biobehavioral Reviews, 33, 81–88. doi: 10.1016/j.neubiorev.2008.08.004 Thayer, J. F., Sollers, J. J., Ruiz-Padial, E., & Vila, J. (2002). Estimating respiratory frequency from autoregressive spectral analysis of heart period. IEEE Engineering in Medicine and Biology Magazine, 21, 41–45. doi: 10.1109/MEMB.2002. 1032638 Thayer, J. F., & Sternberg, E. (2006). Beyond heart rate variability: Vagal regulation of allostatic systems. Annals of the New York Academy of Sciences, 1088, 361–372. doi: 10.1196/annals. 1366.014 Treynor, W., Gonzalez, R., & Nolen-Hoeksema, S. (2003). Rumination reconsidered: A psychometric analysis. Cognitive Therapy and Research, 27, 247–259. doi: 10.1023/A:1023910315561 Todorova, I. L., Falcón, L. M., Lincoln, A. K., & Price, L. L. (2010). Perceived discrimination, psychological distress and health. Sociology of Health & Illness, 32, 843–861. doi: 10.1111/j.14679566.2010.01257.x Weber, C. S., Thayer, J. F., Rudat, M., Wirtz, P. H., Zimmermann-Viehoff, F., Thomas, A., . . . Deter, H. C. (2010). Low vagal tone is associated with impaired post stress recovery of cardiovascular, endocrine, and immune markers. European Journal of Applied Physiology, 109, 201–211. doi: 10.1007/ s00421-009-1341-x Williams, D. P., Cash, C., Rankin, C., Bernardi, A., Koenig, J., & Thayer, J. F. (2015). Resting heart rate variability predicts selfreported difficulties in emotion regulation: A focus on different facets of emotion regulation. Frontiers in Psychology, 6, 261. doi: 10.3389/fpsyg.2015.00261 Williams, D. R., & Mohammed, S. (2009). Discrimination and racial disparities in health: Evidence and needed research. Journal of Behavioral Medicine, 32, 20–47. doi: 10.1007/s10865-0089185-0

Received February 9, 2016 Revision received February 28, 2017 Accepted March 20, 2017 Published online September 25, 2017 DeWayne P. Williams Department of Psychology The Ohio State University 1835 Neil Avenue Columbus, OH, 43210 USA williams.2917@buckeyemail.osu.edu

Journal of Psychophysiology (2019), 33(1), 13–21

Article

Habitual Use of Cognitive Reappraisal Is Associated With Decreased Amplitude of the Late Positive Potential (LPP) Elicited by Threatening Pictures Neil R. Harrison and Philippe Chassy Department of Psychology, Liverpool Hope University, UK

Abstract: In contrast to our knowledge about instructed emotion regulation, rather little is known about the effects of habitual (or “spontaneous”) emotion regulation on neural processing. We analyzed the relationship between everyday use of cognitive reappraisal (measured by the Emotion Regulation Questionnaire, ERQ-R), and the amplitude of the late positive potential (LPP), which is sensitive to downregulation of negative emotions via reappraisal. Participants viewed a series of neutral and threatening images, and rated them for level of threat. We found increased LPP amplitude for threatening compared to neutral pictures between 500 and 1,500 ms. Crucially, we found smaller LPP amplitudes to threatening versus neutral pictures for participants who used reappraisal more often in everyday life. This relationship between LPP amplitude and the ERQ-R was observed in the 1,000–1,500 ms interval of the LPP, over right centro-parietal electrodes. The current findings indicate that habitual tendency to use reappraisal is associated with reduced amplitude of the LPP in response to threatening pictures, in the absence of any explicit instruction to regulate emotions. Keywords: emotion, emotion regulation, late positive potential (LPP), spontaneous reappraisal

Emotions are vital for ensuring adaptive responses to events and situations that require immediate action, such as the sudden appearance of a threatening person or animal in the vicinity. However in certain situations, emotional responses may be maladaptive, for example, in the case of an imagined threat or danger. The ability to appropriately control and regulate one’s emotional reactions is therefore of great importance for healthy psychological and social functioning (Gross, 2002). Based on the process model of emotion regulation (Gross, 1998, 2015), cognitive reappraisal is an antecedent-focused regulation strategy that aims to alter emotional responses before they become activated in full, by reinterpreting the meaning or selfrelevance of a situation or event. Cognitive reappraisal is an effective strategy for regulating affective responses which has been shown to successfully decrease subjective negative emotional experience (e.g., Ray, McRae, Ochsner, & Gross, 2010), and is a core aspect of psychotherapeutic techniques such as cognitive behavioral therapy (CBT). Journal of Psychophysiology (2019), 33(1), 22–31 https://doi.org/10.1027/0269-8803/a000202

Cognitive reappraisal can be implemented in two conceptually distinct ways: either under instruction or spontaneously. The vast majority of research studies into cognitive reappraisal have employed an instructed approach, where participants are given explicit instructions about how and/or when to employ the strategy of reappraisal, and participants are usually given an opportunity to practice applying the strategy before the experimental task begins. The advantage of the instructed approach is that the causal effects of the emotion regulation strategy can be readily assessed, and this approach has been successful in providing evidence about the behavioral benefits and neural processes related to reappraisal (e.g., Goldin, McRae, Ramel, & Gross, 2008; Kim & Hamann, 2007; McRae et al., 2010; for review, see Cutuli, 2014). However, emotion regulation under experimentally instructed conditions is rather artificial compared to the typical mode of employment of emotion regulation strategies outside the laboratory. In everyday life it is frequently Ó 2017 Hogrefe Publishing

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

necessary to regulate and modulate the strength and/or duration of emotions in the absence of specific instructions to do so; in other words, emotions are regulated spontaneously (also known as “habitual emotion regulation” Gyurak, Gross, & Etkin, 2011). Moreover, it is known that individuals differ in the extent to which they spontaneously employ emotion regulation strategies; more frequent use of cognitive reappraisal in everyday life (i.e., under noninstructed conditions) has been shown to be associated with a number of favorable psychological outcomes such as decreased levels of emotional distress, enhanced social functioning, and greater psychological and physical health (Garnefski, Kraaij, & Spinhoven, 2001; Gross & John, 2003; for review see Cutuli, 2014). It is also known that the dispositional tendency to use emotion regulation strategies is associated with the strength of neural responses elicited by emotionally valenced stimuli, as measured by functional magnetic resonance imaging (fMRI; for review, see Cutuli, 2014). For example, decreased activity in the amygdala, as measured by fMRI, was observed following presentation of unpleasant facial expressions, in participants who reported more frequent use of cognitive reappraisal in everyday life (Drabant, McRae, Manuck, Hariri, & Gross, 2009). However, very little is known about the influence of habitual emotion regulation strategies under non-instructed viewing conditions on electrophysiological measures of affective processing, which can provide precise temporal information concerning the different stages of processing of an emotional stimulus. Using event-related potentials (ERPs), electrophysiological studies of instructed cognitive reappraisal have commonly focused on modulation of the late positive potential (LPP), a sustained positive deflection in the event-related potential elicited by affective cues, with a peak latency of around 500 ms over centro-parietal cortex (Hajcak, Weinberg, MacNamara, & Foti, 2012; Paul, Simon, Kniesche, Kathmann, & Endrass, 2013; Thiruchselvam, Blechert, Sheppes, Rydstrom, & Gross, 2011) and commonly lasting up to 1,500 ms or beyond (Hajcak & Nieuwenhuis, 2006; Weinberg & Hajcak, 2011). The LPP is thought to reflect extensive processing related to stimulus salience (for reviews, see Hajcak, ManNamara, & Olvet, 2010; Hajcak et al., 2012; Olofsson, Nordin, Sequeira, & Polich, 2008), and is commonly assessed in early (e.g., 500–1,000 ms) and later (e.g., > 1,000 ms) time-windows (e.g., Hajcak & Dennis, 2009; Sarlo, Übel, Leutgeb, & Schienle, 2013). The early portion of the LPP is thought to index enhanced attention to motivationally relevant stimuli, whereas the later portion may reflect deeper processing and the appraisal of stimulus meaning (Hajcak et al., 2010, 2012; MacNamara, Foti, & Hajcak, 2009). The LPP has been shown to be sensitive to regulation of Ó 2017 Hogrefe Publishing

emotion via cognitive reappraisal, with studies typically showing decreased amplitude of the LPP in response to negative pictures when participants are instructed to reappraise the meaning of the images (e.g., Hajcak & Nieuwenhuis, 2006; Paul et al., 2013; Thiruchselvam et al., 2011; although see Baur, Conzelmann, Wieser, & Pauli, 2015). We are aware of only one study that has investigated the effects of individual differences in habitual emotion regulation on electrophysiological indices of affective processing in the context of passive viewing of pictures (i.e., in the absence of explicit instructions to regulate). In this study, Zhang and Zhou (2014) investigated modulation of the LPP in relation to individual differences in automatic emotion regulation, which was defined as the goal-driven regulation of affect in the absence of conscious decision or deliberate control. Participants were divided into two groups, based on their scores on the emotion-regulation Implicit Association Test (Mauss, Evers, Wilhelm, & Gross, 2006): One group consisted of participants who tended to automatically control their emotions, and the other group consisted of participants who tended to automatically express their emotions. The ERP data showed that participants in the automatic emotion control group had reduced right-sided posterior LPP amplitude differences between high and low arousal emotional pictures, compared to the group with automatic emotion express tendencies. While Zhang and Zhou’s (2014) study provided evidence that individual differences in emotion regulation tendencies modulated the LPP, it was not clear from the study which specific emotion regulation technique the participants habitually used, for example, participants could have used repression, or distraction, as automatic techniques to control emotions. In other words, their study could not shed light on the specific effects of habitual cognitive reappraisal on the LPP. The goal of the current study was therefore to use eventrelated potentials (ERPs) to test whether, in the absence of explicit instructions to regulate emotions, the habitual tendency to use cognitive reappraisal was associated with the strength of cortical responses to threatening pictures, as measured by the LPP. Participants’ habitual use of cognitive reappraisal was assessed using the reappraisal scale of the Emotion Regulation Questionnaire (ERQ-R; Gross & John, 2003). All valenced cues in the current study belonged to a single emotional category (threat) that has high intrinsic motivational relevance. We expected amplitude differences in the early posterior negativity (EPN) and the LPP components, between threatening versus neutral images, in line with previous research (Hajcak et al., 2010; Lang & Bradley, 2010; Van Strien, Eijlers, Franken, & Huijding, 2014; Van Strien, Franken, & Huijding, 2009). Participants’ subjective ratings of the threat value of the presented stimuli were collected after Journal of Psychophysiology (2019), 33(1), 22–31

picture offset. Participants’ trait anxiety levels were assessed via self-report to control for emotional reactivity (Kashdan, 2002), and we controlled for the use of expressive suppression to ensure that the results would not be due to individual differences in regulating emotion via a different regulation strategy. We expected to observe a decreased amplitude of the LPP in response to threatening pictures in participants who more frequently used cognitive reappraisal in their daily lives. Secondly, following Zhang and Zhou (2014), we expected to observe cortical asymmetry in the association between the LPP and the self-reported use of cognitive reappraisal.

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

Questionnaires The 10-item Emotion Regulation Questionnaire (ERQ; Gross & John, 2003) was used to measure emotion regulation strategy. The ERQ uses ratings from 1 (= strongly disagree) to 7 (= strongly agree) and contains six items measuring individual differences in use of cognitive reappraisal (e.g., “When I’m faced with a stressful situation, I make myself think about it in a way that helps me stay calm”), and four items related to use of expressive suppression (e.g., “I control my emotions by not expressing them”). Participants also completed the 20-item trait version of the State-Trait Anxiety Questionnaire (STAI; Spielberger, 1968).

Methods Procedure Participants Sixteen participants (11 males and 5 females) voluntarily took part in the experiment. Mean age was 29.0 years (SD = 7.9). All participants had normal or corrected-tonormal vision, and 15 participants were right-handed. All participants gave written informed consent to the study. The experiment was approved by the Liverpool Hope Psychology Ethics Committee.

Stimuli Thirty neutral and 30 threatening images were selected on the basis of valence and arousal norms from the International Affective Picture System1 (IAPS; Lang, Bradley, & Cuthbert, 2008). The threat pictures depicted actual or potential physical threat or harm and were rated low on pleasure (mean = 2.28, SD = 0.75) and high on arousal (mean = 6.41, SD = 0.62) according to the standardized affective rating system (Lang et al., 2008) and included scenes of physical attacks, dead bodies, and accidents. Neutral pictures were rated near the midpoint of the valence scale (mean = 5.19, SD = 0.55) and low on arousal (mean = 3.52, SD = 0.62) and included pictures of people and objects, landscapes, and animals. The freqspat.m Matlab function from Delplanque, N’diaye, Scherer, and Grandjean (2007) was used to confirm that the two picture categories did not differ in spatial frequencies (all ps > .611). The mean and standard deviation luminance was equalized for all 60 images using the lumMatch.m function from the SHINE toolbox for Matlab (Willenbockel et al., 2010). 1

Participants completed the ERQ and the trait STAI prior to the Electroencephalograph (EEG) experiment. For the EEG experiment, participants were seated at a distance of 60 cm from a computer screen. Each trial began with a central fixation cross lasting 1,500 ms, immediately followed by presentation of either a neutral or threatening image for 1,500 ms (e.g., Mogg, Bradley, Miles, & Dixon, 2004). Next, a Likert scale appeared in the center of the screen for participants to rate the preceding image for threat on a 1–9 scale (1 = not at all threatening; 9 = extremely threatening). Participants were instructed at the start of the experiment that threat was defined as “the degree of physical harm or danger to others which the picture depicts and/or the degree of uneasiness or fear which the picture makes you feel” (Mogg et al., 2000, p. 388). After the participant had entered a number between 1 and 9, a blank screen appeared for 1,000 ms, and then the next trial began. Prior to the main experiment, participants completed a practice block of six trials, with visual images that were not included in the main experiment. In the main experiment 180 images were displayed (90 threatening images, 90 neutral images), in three blocks of 60 trials. The order of trials was randomized. The experiment was controlled using E-Prime 2.0.

EEG Data Acquisition and Preprocessing EEG data was recorded from 64 scalp electrodes using an Active Two amplifier system (BioSemi, Amsterdam, The Netherlands). Electrodes were placed according to the extended 10–20 system (Nuwer et al., 1998). Four additional

The threatening IAPS pictures were: 1052, 1120, 1300, 1932, 3010, 3015, 3060, 3064, 3068, 3069, 3168, 3530, 6230, 6244, 6260, 6312, 6313, 9040, 9042, 9301, 9325, 9405, 9410, 9413, 9433, 9584, 9630, 9635.1, 9901, and 9940. The neutral IAPS pictures were: 1350, 1121, 1670, 1947, 2104, 2107, 2214, 2220, 2305, 2377, 2382, 2383, 2393, 2396, 2397, 2400, 2411, 2441, 2484, 2489, 2500, 2595, 7009, 7025, 7026, 7190, 7513, 7547, 7920, and 7950.

Journal of Psychophysiology (2019), 33(1), 22–31

Ó 2017 Hogrefe Publishing

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

leads were placed above and below the left eye and on the outer canthi of the left and right eyes, to record the vertical and horizontal electrooculogram (EOG; VEOG and HEOG, respectively). Electroencephalograph (EEG) signals from all channels were acquired with respect to the common mode sense (CMS) electrode at a sampling rate of 512 Hz, and were digitally filtered (second-order zero-phase-lag bandpass filter, 0.1–30 Hz). The continuous EEG was divided into epochs offline, beginning 1,500 ms prior to stimulus onset and ending 1,500 ms after stimulus onset. EEG artifacts were rejected using the spontaneous coronary artery dissection (SCADS) procedure with standard parameters (Junghöfer, Elbert, Tucker, & Rockstroh, 2000). This procedure first detected individual channel artifacts, then transformed the data to the average reference and then identified global artifacts. Epochs that contained more than 10 unreliable electrodes were excluded from analysis on the basis of the distribution of their amplitude, standard deviation, and gradient. For the remaining epochs, data from artifact-contaminated sensors was replaced by a statistically weighted spherical interpolation using the complete electrode set. With respect to the spatial distribution of the approximated electrodes, it was ensured that the rejected channels were not localized within one region of the scalp, as this would make interpolation for this area unreliable. Therefore the standard deviation of the spherical splines used for approximation was computed for each epoch and epochs that represented outliers from this distribution were rejected. Across all participants and all conditions the procedure rejected an average of 36.9% of epochs as artifacts [there was no difference in rejection rate per condition, t(15) = .284, p = .780].

Event-Related Potentials (ERPs) ERPs were averaged separately for each stimulus condition (threat and neutral), to produce two ERPs per participant. ERP amplitudes were aligned to a 100 ms pre-stimulus baseline period. The early posterior negativity (EPN) was derived from mean activity in the 200–300 ms timewindow at left (O1, PO3, PO7) and right (O2, PO4, PO8) lateral occipital electrode locations (Van Strien et al., 2009, 2014). The late positive potential (LPP) was maximal at around 550 ms over centro-parietal electrodes, and lasted for the duration of the stimulus (i.e., 1,500 ms), consistent with results from previous studies (e.g., MacNamara & Hajcak, 2009). Analysis of the LPP was conducted during two time intervals (500–1,000, and 1,000–1,500 ms, following stimulus onset), in close agreement with a number of previous studies (e.g., Hajcak & Dennis, 2009; Sarlo et al., 2013; Solomon, DeCicco, & Denis, 2012). Ó 2017 Hogrefe Publishing

(A)

(B)

(C)

Figure 1. ERP plots of the early posterior negativity (EPN) and late positive potential (LPP). (A) Grand-averaged ERPs for the threat (solid line) and neutral (dashed line) conditions, averaged over occipitoparietal locations, show an EPN between 200 and 300 ms after stimulus onset, where the waveform for the threat condition is more negative than the neutral condition. (B) The LPP, averaged across left and right centro-parietal electrode clusters, for the threat (solid line) and neutral (dashed line) conditions. The LPP has a peak latency of around 550 ms after stimulus onset and is evident for the whole duration of the stimulus presentation (i.e., 1,500 ms). (C) Topographic maps (back view) of the earlier (left) and later (right) late positive potential (LPP).

In the 500–1,000 ms time-window, the LPP displayed a broad bilateral distribution over posterior electrode sites (Figure 1C). A cluster of three electrodes was selected based on the sensors showing maximum LPP amplitude (P1, P3, and PO3). Equivalent electrodes in the right hemisphere were selected (P2, P4, and PO4). In the 1,000–1,500 ms Journal of Psychophysiology (2019), 33(1), 22–31

time-window, the LPP showed a bilateral distribution over slightly more superiorly located centro-parietal sensor positions (Figure 1C). A cluster of three electrodes was selected based on the sensors showing maximum LPP amplitude within this time-window (CP2, CP4, and C4). Equivalent electrodes in the left hemisphere were selected (CP1, CP3, and C3). These electrode locations are very similar to those reported in previous investigations of the LPP (e.g., Gerdes et al., 2013; Leleu et al., 2015; MacNamara & Hajcak, 2009). The mean amplitudes for each electrode cluster within each time-window were submitted to a repeated-measures analysis of variance (ANOVA) with the factors Condition (threat and neutral) and Scalp Laterality (left and right).

Results Participant Characteristics and Threat Ratings There was a marginally significant negative correlation between dispositional cognitive reappraisal and trait anxiety (r = .457, p = .075). Confirming the experimental design, participants rated the threatening pictures as more threatening (mean = 6.84, SD = 1.11) than the neutral pictures (mean = 1.56, SD = 0.51; t(15) = 19.217, p < .001). A Pearson’s correlation analysis revealed no relationship between cognitive reappraisal and threat ratings (r = .343, p = .193).

Event-Related Potentials Early Posterior Negativity (EPN) To assess the effectiveness of the threatening images to elicit affective responses, we analyzed the EPN, which is known to be related to emotional processing (e.g., Van Strien et al., 2009, 2014). Between 200 and 300 ms ERPs evoked by unpleasant pictures showed a relative negative potential difference over occipito-parietal sites, compared to neutral pictures (see Figure 1A), characteristic of the EPN component (Van Strien et al., 2009, 2014). Mean amplitudes in the 200–300 ms interval were submitted to a repeated-measures ANOVA with the factors Condition (threat and neutral) and Scalp Laterality (left and right). This analysis revealed a significant main effect of Condition, F(1, 15) = 8.99, p = .009, where amplitudes to threatening images (mean = 7.33, SD = 3.81 μV) were less positive than amplitudes to neutral images, (mean = 8.58, SD = 3.71 μV). There was also a significant interaction between Condition and Scalp Laterality, F(1, 15) = 7.20, p = .017. Follow-up analyses revealed that the EPN was less positive for threatening than for neutral pictures in the left cluster,

Journal of Psychophysiology (2019), 33(1), 22–31

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

t(15) = 3.411, p = .004, but the effect was only marginally significant in the right cluster, t(15) = 2.113, p = .052. Late Positive Potential (LPP) Grand-averaged ERP waveforms displayed a late positive potential (LPP), consisting of a sustained positive deflection with a peak amplitude occurring at around 500 ms over posterior electrodes (Figures 1B and 1C). Mean amplitudes in the 500–1,000 ms and the 1,000–1,500 ms intervals were submitted to separate repeated-measures ANOVAs with the factors Condition (threat and neutral) and Scalp Laterality (left cluster and right cluster). Between 500 and 1,000 ms the analysis revealed a significant main effect of Condition, F(1, 15) = 23.6, p < .001, where amplitudes were higher in the threat condition (mean = 3.86, SD = 2.04 μV) compared to the neutral condition (mean = 2.40, SD = 1.82 μV). There was no main effect of Scalp Laterality (p = .404) and no Condition by Scalp Laterality interaction (p = .973). Between 1,000 and 1,500 ms, there was a significant main effect of Condition, F(1, 15) = 7.20, p = .017, where amplitudes were higher in the threat condition (mean = .97, SD = 1.24 μV) compared to the neutral condition (mean = .48, SD = 1.13 μV). There was no main effect of Scalp Laterality (p = .307) and no Condition by Scalp Laterality interaction (p = .686). The primary purpose of the study was to investigate whether habitual use of cognitive reappraisal (as assessed by the ERQ-R) was associated with reduced amplitude of the LPP for the threatening pictures. Between 500 and 1,000 ms poststimulus onset a Pearson’s correlation analysis (one-tailed) revealed no relationship between LPP amplitude in the threat condition and the reappraisal score in either the right (r = .020, p = .471) or the left (r = .311, p = .120) parietal clusters. Between 1,000 and 1,500 ms poststimulus onset a Pearson’s correlation analysis revealed a significant inverse relationship between LPP amplitude in the threat condition and the reappraisal score in the right centro-parietal cluster (r = .614, p = .005; see Figures 2A and 2B), but not in the left cluster (r = .199, p = .231). In other words, over right centro-parietal electrodes, between 1,000 and 1,500 ms after stimulus onset, the amplitude of the LPP was more attenuated for those participants who used cognitive reappraisal more frequently in everyday life. To exclude the potential influence of emotional reactivity, as assessed by the trait STAI which indexes individual differences in proneness to anxiety, the influence by a different emotion regulation technique (expressive suppression), and gender, for the right centro-parietal electrode cluster we ran a partial correlation between LPP amplitude in the threat condition (1,000–1,500 ms) and the reappraisal score, with STAI trait, expressive suppression, and gender as control variables. The correlation between LPP Ó 2017 Hogrefe Publishing

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

(A)

(B)

amplitude in the threat condition and the reappraisal score remained significant, even after excluding the potential influence of emotional reactivity (i.e., trait anxiety), expressive suppression, and gender (r = .731, p = .003, df = 11). To test whether the relationship between habitual use of reappraisal and the LPP amplitude between 1,000 and 1,500 ms over right central-parietal scalp was specific to the threat condition, we carried out a Pearson’s correlation between the ERQ-R and the LPP in the neutral condition, and this showed no significant association (r = .255, p = .171). Further, we carried out a Pearson’s correlation between the ERQ-R and the difference between the LPP amplitude in the threat versus the neutral condition (i.e., threat minus neutral LPP amplitude), which revealed a significant inverse relationship (r = .523, p = .019). The relationship between the ERQ-R and the threat minus neutral LPP remained significant when controlling for the potential influence of emotional reactivity, expressive suppression, and gender (r = .555, p = .025, df = 11). Finally, we investigated whether the use of expressive suppression was related to the LPP amplitude between 500–1,000 ms and 1,000–1,500 ms, and no significant associations were observed (500–1,000 ms: left cluster, r = .163, p = .55; right cluster, r = .14, p = .60; 1,000– 1,500 ms: left cluster, r = .01, p = .99; right cluster, r = .04, p = .88).

Discussion The current experiment aimed to investigate the association between individual differences in the habitual use of cognitive reappraisal and the emotion-related late positive potential (LPP) component of the event-related potential. Our results showed that participants who used cognitive reappraisal more often in their daily life (as assessed by the ERQ-R) displayed decreased amplitude of the LPP over right centro-parietal scalp between 1,000 and 1,500 ms Ó 2017 Hogrefe Publishing

Figure 2. Plot of LPP and scatterplot of relationship between LPP amplitude and reappraisal score. (A) Grand-averaged ERPs for the threat (solid line) and neutral (dashed line) conditions, over right centro-parietal locations. (B) Scatterplot of the reappraisal scale of the ERQ (ERQ-R) and the amplitude of the threat-related LPP between 1,000 and 1,500 ms over right centro-parietal scalp.

after image onset. The attenuation in LPP amplitude was specific to threat-related stimuli and was not present in response to emotionally neutral pictures. Our results could not be explained by individual differences in emotion reactivity (as assessed by trait anxiety), or by the use of another common method of regulating emotions, namely expressive suppression. As expected, we found enhanced amplitudes of the EPN in response to threatening versus neutral images over occipito-parietal regions, in accordance with previous studies (Van Strien et al., 2009, 2014), providing strong evidence that the threatening images evoked the intended emotional response. Likewise, we observed greater LPP amplitude in response to threatening versus neutral images over centro-parietal regions between 500–1,000 ms and 1,000–1,500 ms after picture onset, in general agreement with previous studies (Hajcak et al., 2010; Lang & Bradley, 2010). Our most important finding was that individual differences in the spontaneous use of cognitive reappraisal (as assessed via the ERQ-R) were associated with the amplitude of the LPP in response to threatening images. Specifically, the more frequent the self-reported use of reappraisal, the more the LPP amplitude was attenuated in response to threatening compared to neutral images, between 1,000 and 1,500 ms after stimulus onset, over right centro-parietal scalp. The observed decrease in LPP amplitude is in agreement with the vast majority of previous research that have shown that the LPP is reduced during (instructed) cognitive reappraisal (for reviews, see Hajcak et al., 2010, 2012), but here we show, for the first time, that the LPP is reduced via cognitive reappraisal under more natural conditions, that is, in the absence of experimental instruction. Attenuation of the LPP amplitude during down-regulation of emotion by reappraisal is generally explained as reflecting diminished arousal as a result of changes in stimulus meaning (Hajcak et al., 2010, 2012). This explanation is consistent with the current findings, where the tendency to use cognitive reappraisal in daily life, Journal of Psychophysiology (2019), 33(1), 22–31

and hence to reinterpret the pictures in a way that reduces their affective impact, was associated with diminished amplitude of the LPP. The current findings are also in agreement with fMRI results showing that increased habitual use of cognitive reappraisal is associated with reduced activations in emotion-generative cortical regions such as the amygdala (Drabant et al., 2009). The association between the LPP amplitude and selfreported use of cognitive reappraisal was found only in the 1,000–1,500 ms time-window, and there was no evidence for an association between 500 and 1,000 ms. Several studies have shown LPP modulations by (instructed) reappraisal at comparatively late stages of stimulus processing, for example, after 1,000 ms (Gan, Yang, Chen, & Yang, 2015) or 1,500 ms (Thiruchselvam et al., 2011) poststimulus onset. Modulation of the LPP at this relatively late stage of processing is in accordance with the process model of emotion regulation, in that reappraisal is a relatively timeconsuming process that requires several stages of processing (i.e., attending to and then evaluating the meaning of the stimulus) before successful reinterpretation can be achieved. Indeed, the later portion of the LPP is thought to reflect appraisal of the meaning of the stimulus (Hajcak et al., 2010, 2012; MacNamara et al., 2009). Similarly, Gan et al. (2015) reported that the LPP amplitude was lowered by reappraisal only after 1,000 ms, and found that during the early period (400–1,000 ms) the LPP was increased for cognitive reappraisal, compared to passive viewing. A potential explanation for their finding is that the LPP during the early period is influenced by cognitive processes governing the implementation of the reappraisal strategy. The time-course of LPP modulation in the current study is also in accordance with findings by Moser, Hartwig, Moran, Jendrusina, and Kross (2014) who found that, in the context of instructions to positively reappraise picture content, trait reappraisal modulated the LPP after, but not before, 1,000 ms following picture presentation. Conversely, other studies have reported relatively early effects of reappraisal on the LPP, even beginning at 200 ms (Hajcak & Nieuwenhuis, 2006) to 400 ms (Moser, Krompinger, Dietz, & Simons, 2009) after picture onset. It could be that in the current study where the use of reappraisal was spontaneous rather than instructed, the effects on the LPP were not seen until after 1,000 ms post-picture onset, as implementation of the strategy was more cognitively demanding compared to an instructed reappraisal context. The association between the LPP amplitude and spontaneous use of cognitive reappraisal was found over right, but not left, centro-parietal cortex. The right-lateralized pattern in the LPP is in line with recent findings by Zhang and Zhou (2014), who reported that participants in an automatic emotion control group had reduced right posterior LPP Journal of Psychophysiology (2019), 33(1), 22–31

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

amplitude differences between high and low arousal emotional pictures, compared to a group with automatic emotion express tendencies. Together, this may suggest that the LPP over right centro-parietal scalp is particularly sensitive to individual differences in the use of emotion regulation techniques in the absence of experimental instruction. Moreover, fMRI data has revealed asymmetries in cortical responses as a function of habitual use of cognitive reappraisal, but these asymmetries have been found mainly in the prefrontal cortex (Kim, Cornwell, & Kim, 2012). In any case, it will be important for future studies to better understand the role and function of brain hemispheric asymmetries in the processing of emotional pictures in relation to individual differences in habitual emotion regulation. We found no association between habitual use of expressive suppression and LPP amplitude in the current study, and we suggest two possible explanations. Firstly, the effectiveness of suppression to reduce negative affect has been shown to be reduced compared to reappraisal (Gross & Levenson, 1993), and, unlike reappraisal, it appears not to reduce activation in emotion-related cortical regions such as the amygdala and insula (Goldin et al., 2008). Secondly, suppression (a response-focused strategy) is thought to target different stages in the emotion generation process compared to reappraisal (an antecedent-focused strategy), and suppression likely affects later stages of emotion generation compared to reappraisal. Indeed, Goldin et al. (2008) found that reappraisal activated cortical areas related to emotion control in an early timewindow (0–4.5 s) while suppression activated those regions in a later window (10.5–15 s). Moreover, a recent ERP study (Gan et al., 2015) found that while instructed reappraisal reduced the amplitude of the LPP, suppression did not lower the LPP amplitude, compared to passive viewing. Together, these considerations suggest that the lack of association between habitual use of expressive suppression and the amplitude of the LPP in the current study may be due to the reduced efficacy of suppression as a technique to regulate emotions, and that suppression may influence ERP components other than the LPP (e.g., the N2; Gan et al., 2015). Several limitations of the current study should be acknowledged. Firstly, it is not clear to what extent the participants were using the strategy of cognitive reappraisal while viewing the pictures. Future research could probe the participants’ regulation technique retrospectively after the experiment to more fully elucidate the nature of the participants’ trial-by-trial regulation strategies. In this regard, it would also be useful to ask participants to retrospectively report whether they were using a more deliberate cognitive reappraisal strategy or alternatively a more automatic/implicit strategy, as it is known that spontaneous emotion regulation can encompass both types of strategies (Gyurak et al., 2011), Ó 2017 Hogrefe Publishing

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

depending on, for instance, the length of time that an individual has used a given technique. It is important to note, though, that the current results could not be explained by individual differences in expressive suppression as a strategy to down-regulate emotional reactions. Secondly, we observed no association between selfreported habitual use of cognitive reappraisal and the behavioral outcome of the experiment (i.e., the threat ratings of the pictures). A number of studies have found that instructed forms of cognitive reappraisal led to reduced perceived intensity of negative or unpleasant stimuli (e.g., Hajcak & Nieuwenhuis, 2006; Paul et al., 2013) compared to passive viewing conditions, measured using explicit ratings of the intensity of the participant’s emotional response (Hajcak & Nieuwenhuis, 2006), or the arousal and unpleasantness dimensions of the stimuli (Paul et al., 2013). In the current study the emotional intensity evoked by the pictures was not directly measured; instead, participants were asked to judge the threat value, which may not reflect the judgment of emotional intensity of the picture,2 and could explain why we failed to observe an association between habitual cognitive reappraisal and threat ratings. Future studies should more directly measure the participants’ emotional intensity, to investigate links between habitual reappraisal and self-reported intensity of affect evoked by the images. While we did not explicitly control for emotional reactivity, we instead measured trait anxiety (using the STAI trait version), which is known to be a proxy for emotional reactivity (Kashdan, 2002), with high positive correlations (r = .70) between the STAI trait version and different measures of emotional reactivity (e.g., Fox, Cahill, & Zougkou, 2010; Marshall, Wortman, Vickers, Kusulas, & Hervig, 1994). While the STAI measures general anxiety levels, more specific anxiety measures could be used in future studies, such as those measuring social anxiety, as different types of anxiety are known to influence different ERP components (e.g., Rossignol, Philippot, Bissot, Rigoulot, & Campanella, 2012). A further potential limitation in the study was the relatively small sample size, however our major finding (correlation between ERQ-R and LPP amplitude) was sufficiently strong as to produce statistical significance at the conventional levels and a large effect size. A retrospective power analysis of our main statistical result was carried out using the pwr (Champely, 2012) package in R-statistics (R Core Team, 2015). With N = 16, α set at 0.05, and r = .615 for one-tailed tests, analysis revealed a power (1 β) value of 0.846, indicating a very high – over 85% – chance of detecting genuine effects. In summary, a Type II error was unlikely (Field, 2013). 2

Conclusions The current study aimed to investigate the effect of spontaneous cognitive reappraisal on the LPP, which is sensitive to emotion-related processing. The habitual use of cognitive reappraisal is known to be associated with decreased negative affect, improved interpersonal functioning, and enhanced mental and physical well-being (Gross & John, 2003). We found that a greater tendency to use spontaneous emotion regulation in everyday life was associated with reduced LPP amplitude to threatening pictures between 1,000 and 1,500 ms after stimulus onset, over right centro-parietal electrodes. Most previous research has shown LPP amplitude reductions during instructed cognitive reappraisal, but here we show, for the first time, that the LPP is attenuated via cognitive reappraisal under more ecologically valid conditions. Given the strong association between trait reappraisal and psychological health (Gross & John, 2003), the current findings suggest that the LPP may be a clinically relevant index of adaptive cognitive change as implemented in everyday life, that is, in the absence of explicit experimental instructions. Conflicts of Interest The authors declare that they have no competing interests.

References Baur, R., Conzelmann, A., Wieser, M. J., & Pauli, P. (2015). Spontaneous emotion regulation: Differential effects on evoked brain potentials and facial muscle activity. International Journal of Psychophysiology, 96, 38–44. Champely, S. (2012). Pwr: Basic functions for power analysis. Retrieved from http://CRAN.R-project.org/package=pwr Cutuli, D. (2014). Cognitive reappraisal and expressive suppression strategies role in the emotion regulation: An overview on their modulatory effects and neural correlates. Frontiers in Systems Neuroscience, 8, 175. Delplanque, S., N’diaye, K., Scherer, K., & Grandjean, D. (2007). Spatial frequencies or emotional effects? A systematic measure of spatial frequencies for IAPS pictures by a discreet wavelet analysis. Journal of Neuroscience Methods, 165, 144–150. Drabant, E. M., McRae, K., Manuck, S. B., Hariri, A. R., & Gross, J. J. (2009). Individual differences in typical reappraisal use predict amygdala and prefrontal responses. Biological Psychiatry, 65, 367–373. Field, A. (2013). Discovering statistics using IBM SPSS statistics (4th ed.). London, UK: Sage. Fox, E., Cahill, S., & Zougkou, K. (2010). Preconscious processing biases predict emotional reactivity to stress. Biological Psychiatry, 67, 371–377. Gan, S., Yang, J., Chen, X., & Yang, Y. (2015). The electrocortical modulation of different emotion regulation strategies. Cognitive Neurodynamics, 9, 399–410.

Following Mogg et al. (2000), part of the threat rating was a judgment of the degree of physical harm or danger to others which the picture depicted, which could be unrelated to judgments about the perceived emotional intensity of the image.

Ó 2017 Hogrefe Publishing

Journal of Psychophysiology (2019), 33(1), 22–31

Garnefski, N., Kraaij, V., & Spinhoven, P. (2001). Negative life events, cognitive emotion regulation and emotional problems. Personality and Individual Differences, 30, 1311–1327. Gerdes, A. B. M., Wieser, M. J., Bublatzky, F., Kusay, A., Plichta, M. M., & Alpers, G. W. (2013). Emotional sounds modulate early neural processing of emotional pictures. Frontiers in Psychology, 4, 1–12. doi: 10.3389/fpsyg.2013.00741 Goldin, P. R., McRae, K., Ramel, W., & Gross, J. J. (2008). The neural bases of emotion regulation: Reappraisal and suppression of negative emotion. Biological Psychiatry, 63, 577–586. Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Reviews of General Psychology, 2, 271–299. Gross, J. J. (2002). Emotion regulation: Affective, cognitive, and social consequences. Psychophysiology, 39, 281–291. Gross, J. J. (2015). Emotion regulation: Current status and future prospects. Psychological Enquiry, 26, 1–26. Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes: Implications for affect, relationships, and well-being. Journal of Personality and Social Psychology, 85, 348–362. Gross, J. J., & Levenson, R. W. (1993). Emotion suppression: Physiology, self-report, and expressive behavior. Journal of Personality and Social Psychology, 64, 970–986. Gyurak, A., Gross, J. J., & Etkin, A. (2011). Explicit and implicit emotion regulation: A dual-process framework. Cognition & Emotion, 25, 400–412. Hajcak, G., & Dennis, T. A. (2009). Brain potentials during affective picture processing in children. Biological Psychology, 80, 333–338. Hajcak, G., ManNamara, A., & Olvet, D. M. (2010). Event-related potentials, emotion, and emotion regulation: An integrative review. Developmental Neuroscience, 35, 129–155. Hajcak, G., & Nieuwenhuis, S. (2006). Reappraisal modulates the electrocortical response to unpleasant pictures. Cognitive, Affective & Behavioral Neuroscience, 6, 291–297. Hajcak, G., Weinberg, A., MacNamara, A., & Foti, D. (2012). ERPs and the study of emotion. In S. J. Luck & E. S. Kappenman (Eds.), The Oxford handbook of event-related potential components (pp. 441–474). New York, NY: Oxford University Press. Junghöfer, M., Elbert, T., Tucker, D. M., & Rockstroh, B. (2000). Statistical control of artifacts in dense array EEG/MEG studies. Psychophysiology, 37, 523–532. Kashdan, T. B. (2002). Social anxiety dimensions, neuroticism, and the contours of positive psychological functioning. Cognitive Therapy and Research, 26, 789–810. Kim, S. H., Cornwell, B., & Kim, S. E. (2012). Individual differences in emotion regulation and hemispheric metabolic asymmetry. Biological Psychology, 89, 382–386. Kim, S., & Hamann, S. (2007). Neural correlates of positive and negative emotion regulation. Journal of Cognitive Neuroscience, 19, 776–798. Lang, P. J., & Bradley, M. M. (2010). Emotion and the motivational brain. Biological Psychology, 84, 437–450. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (2008). International affective picture system (IAPS): Affective ratings of pictures and instruction manual (Technical Report A-8). Gainesville, FL: University of Florida. Leleu, A., Godard, O., Dollion, N., Durand, K., Schall, B., & Baudouin, J.-Y. (2015). Contextual odors modulate the visual processing of emotional facial expressions: An ERP study. Neuropsychologia, 77, 366–379. MacNamara, A., Foti, D., & Hajcak, G. (2009). Tell me about it: Neural activity elicited by emotional pictures and preceding descriptions. Emotion, 9, 531–543.

Journal of Psychophysiology (2019), 33(1), 22–31

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

MacNamara, A., & Hajcak, G. (2009). Anxiety and spatial attention moderate the electrocortical response to aversive pictures. Neuropsychologia, 47, 2975–2980. Marshall, G. N., Wortman, C. B., Vickers, R. R. Jr., Kusulas, J. W., & Hervig, L. K. (1994). The five-factor model of personality as a framework for personality-health research. Journal of Personality and Social Psychology, 67, 278–286. Mauss, I. B., Evers, C., Wilhelm, F. H., & Gross, J. J. (2006). How to bite your tongue without blowing your top: Implicit evaluation of emotion regulation predicts affective responding to anger provocation. Personality and Social Psychology Bulletin, 32, 589–602. McRae, K., Hughes, B., Chopra, S., Gabrieli, J. D. E., Gross, J. J. & Ochsner, K. N. (2010). The neural bases of distraction and reappraisal. Journal of Cognitive Neuroscience, 22, 248–262. Mogg, K., Bradley, B., Miles, F., & Dixon, R. (2004). Time course of attentional bias for threat scenes: Testing the vigilanceavoidance hypothesis. Cognition & Emotion, 18, 689–700. Mogg, K., McNamara, J., Powys, M., Rawlinson, H., Seiffer, A., & Bradley, B. P. (2000). Selective attention to threat: A test of two cognitive models of anxiety. Cognition & Emotion, 14, 375–399. Moser, J. S., Hartwig, R., Moran, T. P., Jendrusina, A. A., & Kross, E. (2014). Neural markers of positive reappraisal and their associations with trait reappraisal and worry. Journal of Abnormal Psychology, 123, 91–105. Moser, J. S., Krompinger, J. W., Dietz, J. D., & Simons, R. F. (2009). Electrophysiological correlates of decreasing and increasing emotional responses to unpleasant pictures. Psychophysiology, 46, 17–27. Nuwer, M. R., Comi, G., Emerson, R., Fuglsang-Frederiksen, A., Guerit, J. M., Hinrichs, H., . . . Rappelsburger, P. (1998). IFCN standards for digital recording of clinical EEG. Electroencephalography and Clinical Neurophysiology, Supplement, 52, 11–14. Olofsson, J. K., Nordin, S., Sequeira, H., & Polich, J. (2008). Affective picture processing: An integrative review of ERP findings. Biological Psychology, 77, 247–265. Paul, S., Simon, D., Kniesche, R., Kathmann, N., & Endrass, T. (2013). Timing effects of antecedent- and response-focused emotion regulation strategies. Biological Psychology, 94, 136–142. R Core Team. (2015). R: A Language and Environment for Statistical Computing. Vienna. Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-project.org/ Ray, R. D., McRae, K., Ochsner, K. N., & Gross, J. J. (2010). Cognitive reappraisal of negative affect: Converging evidence from EMG and self-report. Emotion, 10, 587–592. Rossignol, M., Philippot, P., Bissot, C., Rigoulot, S., & Campanella, S. (2012). Electrophysiological correlates of enhanced perceptual processes and attentional capture by emotional faces in social anxiety. Brain Research, 1460, 50–62. Sarlo, M., Übel, S., Leutgeb, V., & Schienle, A. (2013). Cognitive reappraisal fails when attempting to reduce the appetitive value of food: An ERP study. Biological Psychology, 94, 507–512. Solomon, B., DeCicco, J. M., & Denis, T. A. (2012). Emotional picture processing in children: An ERP study. Developmental Cognitive Neuroscience, 2, 110–119. Spielberger, C. D. (1968). Self-Evaluation Questionnaire. STA Form X2. Palo Alto, CA: Consulting Psychologists Press. Thiruchselvam, R., Blechert, J., Sheppes, G., Rydstrom, A., & Gross, J. J. (2011). The temporal dynamics of emotion regulation: An EEG study of distraction and reappraisal. Biological Psychology, 87, 84–92.

Ó 2017 Hogrefe Publishing

N. R. Harrison & P. Chassy, Habitual Reappraisal and the LPP

Van Strien, J. W., Eijlers, R., Franken, I. H. A., & Huijding, J. (2014). Snake pictures draw more early attention than spider pictures in non-phobic women: Evidence from event-related brain potentials. Biological Psychology, 96, 150–157. Van Strien, J. W., Franken, I. H. A., & Huijding, J. (2009). Phobic spider fear is associated with enhanced attentional capture by spider pictures: A rapid serial presentation event-related potential study. Neuroreport, 20, 445–449. Weinberg, A., & Hajcak, G. (2011). The late positive potential predicts subsequent interference with target processing. Journal of Cognitive Neuroscience, 23, 2994–3007. Willenbockel, V., Sadr, J., Fiset, D., Horne, G. O., Gosselin, F., & Tanaka, J. W. (2010). Controlling low-level image properties: The SHINE toolbox. Behavior Research Methods, 42, 671–684.

Ó 2017 Hogrefe Publishing

Zhang, J., & Zhou, R. (2014). Individual differences in automatic emotion regulation affect the asymmetry of the LPP component. PLoS One, 9, e88261.

Received March 31, 2016 Accepted March 29, 2017 Published online September 25, 2017 Neil R. Harrison Department of Psychology Liverpool Hope University Liverpool L16 9JD UK harrisn@hope.ac.uk

Journal of Psychophysiology (2019), 33(1), 22–31

Article

Job Satisfaction Among Mental Health Workers Associations With Respiratory Sinus Reactivity to, and Recovery From Exposure to Mental Stress William H. O’Brien,1 Paul W. Goetz,2 Heather McCarren,3 Eileen Delaney,4 William F. Morrison,1 Tanya S. Watford,1 and Kristin A. Horan1 1

Bowling Green State University, Bowling Green, OH, USA

Division of Cardiac Surgery, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA

VHA National Center for Organization Development, Cincinnati, OH, USA

Naval Center for Combat and Occupational Stress Control (NCCOSC), San Diego, CA, USA

Abstract: Work characteristics such as job satisfaction have been associated with mental and physical health outcomes in several crosssectional and longitudinal studies. However, meta-analytic reviews indicate that nearly all of the reported relationships between these two sets of constructs rely on self-report measures. Thus, the magnitude of the reported relationships may be inaccurate and inflated due to common method variance (mono-method bias) and negative affectivity. Respiratory sinus arrhythmia (RSA) is an objective measure of risk for adverse mental health and physical health outcomes. To our knowledge, there has been no investigation of the relationship between job satisfaction and respiratory sinus arrhythmia. In this investigation, 29 workers in mental health settings who experience higher than average levels of work stress due to the amount and unpredictability of workload completed sociodemographic measures and a job satisfaction measure. RSA was then collected during a resting baseline, a worry induction stressor condition where they were instructed to think about work stressors, and a post-stress recovery condition. RSA reactivity to the stressor was significantly greater for participants with low job satisfaction. The low job satisfaction participants also demonstrated less RSA recovery after the stressor ended. Alternatively, participants with higher job satisfaction reacted less and recovered more completely from the stressor. Keywords: respiratory sinus arrhythmia, job satisfaction

Workers who provide services to persons with mental health problems such as intellectual disabilities (ID) can experience significant work stress and low levels of job satisfaction which have been linked to turnover rates and intent to quit (Gray-Stanley et al., 2010; Larson, Lakin, & Bruininks, 1998; Lunsky, Hastings, Hensel, Arenovich, & Dewa, 2014; Mutkins, Brown, & Thorsteinsson, 2011; Rose, Home, Rose, & Hastings, 2004). Workplace factors that have been linked to work stress and low job satisfaction among persons working with ID populations include problematic client behaviors (Mitchell & Hastings, 2001; Skirrow & Hatton, 2007), perceived lack of support (Rose, 1993), role ambiguity (Hatton, Rashes, Caine, & Emerson, 1995), and unpredictable workloads. Job satisfaction has been associated with adverse mental and physical health outcomes. For example, Faragher, Cass, and Cooper (2005) conducted a comprehensive metaanalysis of research investigating relationships between job Journal of Psychophysiology (2019), 33(1), 32–38 https://doi.org/10.1027/0269-8803/a000203

satisfaction and mental and physical health outcomes. Their meta-analysis provided important insights into this research domain. First, job satisfaction was moderately correlated with several self-reported mental health states such as burnout (overall unadjusted r = .409), depressed mood (overall unadjusted r = .366), and anxiety (overall unadjusted r = .354). Second, job satisfaction was correlated, to a lesser extent, with self-reported physical symptoms (overall unadjusted r = .235). Third, job satisfaction was very modestly correlated with more objective physical health outcomes such as cardiovascular disease (overall unadjusted r = .121) and musculoskeletal disease (overall unadjusted r = .078). Finally, their results indicated that the evaluation of relationships between job satisfaction and these outcomes overwhelmingly relied on self-report measures. As such, it is difficult to parse out the effects of method bias as well as negative affectivity in the existing literature. Nonetheless, it is important to note, as did Faragher et al. (2005), that job Ó 2017 Hogrefe Publishing

W. H. O’Brien et al., Job Satisfaction and RSA

satisfaction was the largest work-related predictor of mental health symptoms. Considering this, the use of objective physiological measures to examine the impact of job satisfaction is warranted. Respiratory sinus arrhythmia (RSA), for example, is an index of parasympathetic influence on the heart (Porges, 2007). RSA has been associated with risk for a number of adverse health outcomes such as diabetes, obesity, and cardiovascular disease (Masi, Hawkley, Rickett, & Cacioppo, 2007). RSA has also been associated with work stress (Toivanen, Länsimies, Jokela, & Hänninen, 1993; Watanabe et al, 2002; Zanstra, Schellekens, Schaap, & Kooistra, 2006) and depression (Bylsma, Solomon, Taylor-Clifft, Morris, & Rottenberg, 2014; Yaptangco, Crowell, Baucom, Bride, & Hansen, 2015). In the aforementioned research, higher levels of RSA were typically associated with lower levels of risk for adverse outcomes. Respiratory sinus arrhythmia may be particularly relevant in the study of job satisfaction. The detrimental effects of work stress, which has been associated with job satisfaction, have been associated with impaired parasympathetic functioning (Clays et al., 2011; Orsila et al., 2008; Vrijkotte, Van Doornen, & De Geus, 2000). Vrijkotte and colleagues (2000) suggested this may be due to the role of the parasympathetic system in recovery and restoration. For example, workers with poor RSA recovery after work stress exposure may also struggle to relax after work on a daily basis and, over time, may have a substantially higher risk for coronary heart disease (Suadicani, Hein, & Gyntelberg, 1993). A review by Jarczok and colleagues (2013) adduced the Neurovisceral Integration Model (Thayer & Lane, 2000) to explain the link between work stress exposure and cardiovascular risk. They proposed that work stress is associated with heart rate variability which, like RSA, is an index of parasympathetic functioning. The Neurovisceral Integration Model posits that RSA can be thought of as an index of cognitive and emotional processes in the cortical and subcortical brain areas. Further, this model suggests that RSA is an index, not only of heart function, but also the degree of integration between the central autonomic network and the peripheral nervous system. The integration of the central autonomic network and the peripheral nervous system provides an individual with the ability to flexibly respond, both psychologically and physiologically, to emotional experiences and to generate effective behavioral responses to address environmental demands (Thayer, Åhs, Fredrikson, Sollers, & Wager, 2012). In other words, parasympathetic functioning indexed by RSA partially reflects an individual’s ability to function in a complex environment (Jarczok et al., 2013). To our knowledge, however, there are no published investigations of the relationship between job satisfaction and RSA. Ó 2017 Hogrefe Publishing

Given the paucity of research examining the relationship between job satisfaction and objective measures of health, in this exploratory study we evaluated the relationship between job satisfaction and RSA among workers in community agencies that provide services to persons with intellectual disabilities. We predicted that higher levels of job satisfaction would be associated with higher levels of RSA overall. We also predicted that higher levels of job satisfaction would be associated with lower levels of RSA reactivity to an acute stressor.

Method Participants Participants were recruited from two agencies that provide services to ID persons in Northwest Ohio. The participants were informed that the study involved an evaluation of work stress and cardiovascular reactivity to stress. A total of 29 workers participated in the study. The study participants were predominantly women (86%, n = 25) and Caucasian (97%, n = 28, one person reported African American ethnicity). The average age of participants was 42 years (range: 25–60; SD = 10.50). Most of the participants were full time employees (90%, n = 26) and they reported working an average of 41 hr per week (SD = 7.00). The job tenure was 6.9 years (SD = 4.40). Four (14%) participants obtained a postgraduate degree, 16 (55%) obtained an undergraduate degree, 7 (24%) completed some college, and 2 (7%) obtained a high school degree.

Procedure Participants completed the work stress evaluation at their workplace during working hours. To accomplish this, the researchers set up the laboratory equipment in a private office in the organizations. Upon arrival, the project was described in detail. Participants then completed a written informed consent. Following this, participants completed self-report questionnaires and then underwent a stress reactivity protocol while RSA was recorded using portable cardiovascular measurement equipment. The stress reactivity protocol consisted of three conditions: a 10-min resting Baseline condition, a 5-min Stress/ Worry condition, and a 10-min Recovery condition. During the Baseline condition, participants were asked to close their eyes and to relax and focus on their breathing. During the Stress/Worry condition, participants were instructed to pick a topic about which they are currently most stressed or worried and asked to focus on this topic as intensely as they Journal of Psychophysiology (2019), 33(1), 32–38

could. This stress/worry induction method has been widely used and accepted within the stress and worry literature (e.g., Behar, Vescio, & Borkovec, 2005; McLaughlin, Borkovec, & Sibrava, 2007). Finally, during the Recovery condition participants were asked to close their eyes, relax, and focus on their breathing. At the conclusion of the Recovery condition, electrodes were removed, and participants were debriefed.

Measures Demographics An 11-item self-report inventory was used to assess demographic variables, illness variables (e.g., doctor visits and days of work missed due to illness), and work characteristics (job tenure, job title, hours worked per week). Job Satisfaction Job satisfaction was measured using the abridged Job Descriptive Index (JDI; Stanton et al., 2002). This scale contains five subscales that measure satisfaction with five aspects of the job including work, pay, promotion opportunities, coworkers, and supervisor. Respondents are asked to indicate whether short phrases describe their job by checking “Yes,” “No,” or “? Cannot Decide.” The JDI is widely accepted as having sound psychometric properties and it has been noted that the abridged version retains the reliability, validity, and factor consistency of the full-length scale (Stanton et al., 2002). The Cronbach’s α for the total JDI was .89 in this sample (note that while the JDI has five subscales, our analyses indicated it could be treated as a global and unidimensional index of job satisfaction). Thought Rating At the conclusion of the baseline, stress/worry, and recovery conditions, participants were asked to “briefly describe the thoughts you experienced during the last 5 min.” They were then asked to write the thoughts. Trained raters then read and assigned the following values: 0 = no worry-related content, 9 = minimal worry-related content, 2 = moderate worry-related content, and 3 = maximal worry-related content. Inter-rater reliability of these worry ratings was high, r = .81, p < .001, with a 70% agreement rate. Differences between raters were resolved through discussion. Respiratory Sinus Arrhythmia Allen, Chambers, and Towers (2007) reviewed methods for calculating RSA and noted that a separate measure of respiration is not needed because the interaction between heart rate variability and respiration occurs in specific Journal of Psychophysiology (2019), 33(1), 32–38

W. H. O’Brien et al., Job Satisfaction and RSA

frequency band that ranges from .12 to .40 Hz. Allen et al. (2007) also developed a software approach that uses two programs to calculate RSA. The first program, the QRSTool, was used to detect R waves in ECG (electrocardiogram) collected from each participant. The QRSTool has an option that allowed us to manually detect and correct artifacts in ECG data. In most cases, this involved identifying points in an ECG where an R wave was not detected. When this occurred, we manually inserted a maker variable for the missed R wave. In some instances, such as movement artifacts or large T waves, we removed markers that the program misidentified as an R wave. After the R waves were marked using the QRSTool, the CMetX program was used to compute RSA. Using the CMetX program, the interbeat intervals between successive R waves were converted to a time series and then run through a .12–.40 Hz band-pass filter. The Allen et al. (2007) validation study of the CMetX program indicated that the .12–.40 range captured variation in respiration-related heart rate variability and parasympathetic activation. The resultant RSA values were then log transformed. Allen et al. (2007) tested the validity of their QRSTool and CMetX system with the MXedit system developed by Porges and Boher (1990). Using a sample of 96 undergraduate students, they found the correlations between the two systems exceeded .99 during resting and stressor conditions. In terms of hardware, ECG data was collected using a Biopac Systems MP30 (Biopac Systems, Galeta, CA) with the Biopac version 3.7.2 analysis software. Three silver-silver chloride electrodes were positioned in a standard Lead II configuration (negative electrode on the right wrist, positive electrode on the left ankle, ground electrode on the right ankle ground) to record ECG. The ECG was sampled at a rate of 1,000 Hz which is well above recommended 500 Hz sampling rate needed to calculate RSA.

Data Reduction and Analysis Missing values were replaced with an individual’s mean score of endorsed items on a self-report measure if less than 20% of measure items were missed. RSA scores were averaged across each condition of the experiment. Only the last 5 min of the Baseline condition and the first 5 min of the Recovery condition were used in computing RSA scores, to match the 5-min period of the Stress/Worry condition.

Results All predictor and dependent measures were examined for measurement adequacy. Skewness ranged from 0.78 to 0.44 and kurtosis ranged from 1.04 to 0.88. Each Ó 2017 Hogrefe Publishing

W. H. O’Brien et al., Job Satisfaction and RSA

Figure 1. RSA as a function of JDI and condition.

6.2 6.1

6 5.9 5.8 5.7 5.6

5.5 5.4 5.3

5.2 5.1 Baseline

Stressor Low JDI

Recovery High JDI

skewness and kurtosis statistic was divided by the standard error for each statistic in order to generate a z-score and the corresponding probability value associated with each z-score (e.g., Hopkins & Weeks, 1990). The skewness z-scores ranged from 1.24 to 0.50 (p = .22 to p = .62) and the kurtosis z-scores ranged from 1.23 to 0.58 (p = .22 to p = .56). Based on these analyses, none of the variables were characterized by significant skew or kurtosis. An examination of individual participant standardized scores indicated that there were no outliers using the recommended cutoff of 3.0. Finally, all of the analyses of variance (ANOVAs) and follow-up comparisons (described below) were tested for equality of variances and sphericity. None of the analyses indicated a significant violation of the normality assumption. The mean, median, and standard deviation for the total JDI was calculated (M = 53.17, Mdn = 57, SD = 16.11) and found to be consistent with prior reports (Stanton et al., 2002). Participants were divided into high JDI and low JDI groups using a median split. Specifically, participants with JDI scores greater than 57 were assigned to the high group while participants with scores equal to or lower than 57 were assigned to the low JDI group). The two groups were compared on relevant sociodemographic and health variables. No significant differences (all p values were > .10 on t-tests for ratio and ordinal variables and Chi-Square tests for nominal variables) were observed between the two groups on the following measures: age, gender, race, marital status, educational attainment, job tenure, and hours worked. A 2 3 (high JDI, low JDI Baseline, Stress/Worry, and Recovery conditions) repeated-measures ANOVA was then conducted using the thought rating as the dependent variable. A significant main effect was observed for condition, F(2, 26) = 86.66, p < .001, ηp2 = .870. Follow-up Ó 2017 Hogrefe Publishing

comparisons of the main effect for condition indicated that the Baseline thought rating (M = 0.48, SD = 0.83) was significantly, F(1, 28) = 96.76, p < .001, ηp2 = .782, lower than the Stress/Worry thought rating (M = 2.35, SD = 0.77) but not significantly, F(1, 28) = 2.31, p = .14, ηp2 = .079, different from the Recovery thought rating (M = 0.21, SD = 0.41). Additionally, the Stress/Worry thought rating was significantly higher than the Recovery thought rating, F(1, 28) = 168.01, p < .001, ηp2 = .867. The main effect for JDI was nonsignificant, F(1, 27) = 1.32, p = .26, ηp2 = .045. Similarly, the interaction between JDI and condition was nonsignificant, F(2, 26) = .50, p = .61, ηp2 = .037. Using Cohen’s (1988) effect size classification scheme (i.e., ηp2 = .01, .06, and .14 are labeled small, medium, and large, respectively), the effect sizes for the condition main effect, the Baseline-Stress/Worry comparison, and the Stress/ Worry-Recovery comparison were all very large. This is because worrisome thoughts were virtually absent during the Baseline and Recovery conditions whereas they were quite common (as expected given instructions) during the Stress/Worry condition. The effect sizes for the JDI main effect, the interaction between the JDI and condition, and the Baseline-Recovery comparison were very small. A 2 3 (high JDI, low JDI Baseline, Stress/Worry, Recovery) repeated-measures ANOVA was then conducted using the RSA as the dependent variable. A significant interaction between the JDI and condition was observed, F(2, 26) = 3.67, p < .04, ηp2 = .220. Figure 1 provides a graphic representation of the interaction. A significant main effect was also observed for condition, F(2, 26) = 5.28, p = .01, ηp2 = .289. Follow-up comparisons of the main effect for condition indicated that the Baseline RSA (M = 5.86, SD = 1.27) was significantly higher, F(1, 28) = 9.94, p = .004, ηp2 = .262, than the Stress/Worry RSA (M = 5.54, SD = 1.35) but not significantly higher, F(1, 28) = .383, Journal of Psychophysiology (2019), 33(1), 32–38

p = .54, ηp2 = .013, than the Recovery RSA (M = 5.91, SD = 1.18). Additionally, the Stress/Worry RSA was significantly lower than the Recovery RSA, F(1, 28) = 9.64, p = .004, ηp2 = .256. The main effect for JDI was nonsignificant, F(1, 27) = .06, p = .81, ηp2 = .002. The effect sizes for the JDI by condition interaction, the main effect for condition, the Baseline-Stress/Worry comparison, and the Stress/ Worry-Recovery comparison were all large. The effect sizes for the JDI main effect and the Baseline-Recovery comparison were very small. Follow-up analyses of the interaction between JDI and condition were conducted. First, a one-way repeated-measures ANOVA was conducted for each group separately. Following this, pairwise comparisons were conducted using the Bonferroni-corrected probability value (p = .02). For the low JDI group, the one-way repeated measures ANOVA was significant, F(2, 14) = 8.10, p = .005, ηp2 = .536. Follow-up pairwise comparisons indicated that the Baseline RSA (M = 5.91, SD = 1.30) was significantly higher than the Stress/Worry RSA (M = 5.46, SD = 1.22) but not the Recovery RSA (M = 5.78, SD = 1.18); F(1, 15) = 17.35, p = .001, ηp2 = .536 and F(1, 15) = 1.20, p = .29, ηp2 = .074, respectively. The comparison between the Stress/Worry RSA and recovery RSA was nonsignificant using the Bonferroni correction, F(1, 15) = 5.37, p = .04, ηp2 = .264. The effect sizes for the overall one-way ANOVA and the Baseline-Stress/ Worry comparison and Stress/Worry-Recovery comparison were large. For the high JDI group, the one-way repeated-measures ANOVA was not significant F(2, 11) = 3.02, p = .09. Exploratory pairwise comparisons were conducted. These revealed that the Baseline RSA (M = 5.79, SD = 1.27) was not significantly different (using Bonferroni correction) from the Stress/Worry RSA (M = 5.64, SD = 1.54) or Recovery RSA (M = 6.07, SD = 1.20); F(1, 12) = 0.45, p = .40, ηp2 = .059 and F(1, 12) = 5.95, p = .03, ηp2 = .331 respectively. The comparison between the Stress/Worry RSA and Recovery RSA was also nonsignificant, F(1, 12) = 4.25, p = .06, ηp2 = .262. Note that the Baseline-Stress/Worry comparison and the Stress/Worry-Recovery comparison were significant using the conventional probability cutoff of .05 but not when the more conservative Bonferroni corrected cutoff was used. Additionally, both comparisons indicated that the Recovery RSA was higher. In sum, the high JDI participants did not show significant RSA reactivity to the stressor. Further, during the Post-Stress/Worry Recovery period, they demonstrated a higher RSA which indicated greater parasympathetic activation. Because of the small sample size, nonparametric analyses were conducted examine the possibility that a few participants were driving the ANOVA results. The Friedman’s test of ranks was used to evaluate condition effects within each group. For the low JDI group, all comparisons were Journal of Psychophysiology (2019), 33(1), 32–38

W. H. O’Brien et al., Job Satisfaction and RSA

similar with a single exception. Specifically, the Baseline RSA (Mean Rank = 1.81) was significantly higher than the Recovery RSA (Mean Rank = 1.18; w2(1) = 6.25, p = .01). This indicated that the low JDI participants did not fully recover from the stressor. For the high JDI participants, the nonparametric findings were equivalent to the ANOVA results. Bivariate correlations were conducted to further explore the relationship between the JDI and RSA. The relationship between the JDI and baseline levels of RSA was not significant, r(27) = .152, p = .45. We then computed two RSA residual scores by: (a) regressing the Stressor/Worry RSA on the Baseline RSA and saving the standardized residual and (b) regressing the Recovery RSA on the Baseline RSA and saving the standardized residual. These residual scores provided indices of reactivity and recovery that were uncorrelated with Baseline levels of RSA. A significant correlation was observed between the JDI and the Stress/ Worry residual score, r(27) = .32, p = .05 and the Recovery residual score, r(27) = .56, p < .002. Together, these correlations indicated that lower levels of job satisfaction were associated with more reactivity to stress and less complete recovery from stress. In summary, results indicated that the laboratory stressor was able to induce a significant increase in worry and RSA reactivity. Further, participants with lower job satisfaction scores showed great RSA reactivity to the stressor and did not fully recover from it. Alternatively, participants with higher job satisfaction scores did not show significant RSA reactivity to the stressor and their recovery was more complete.

Discussion This investigation was designed to evaluate the relationship between job satisfaction and RSA among persons working in agencies that provide services to persons with ID. We predicted that higher levels of job satisfaction would be associated lower levels of reactivity to a stressor. Consistent with expectations, we found that higher levels of job satisfaction were associated with lower levels RSA reactivity to a stressor and better recovery from a stressor. Alternatively, persons with lower job satisfaction reacted more strongly to the stressor and recovered less completely. As noted in the Introduction, job satisfaction has been associated with mental and physical health using self-report measures. We thus predicted that job satisfaction would be associated with a healthier RSA response to stress. The finding of a relationship between job satisfaction and RSA reactivity results suggests that one mechanism through which job satisfaction may be linked to adverse health outcomes is through impaired parasympathetic functioning. This is consistent with the Neurovisceral Integration Ó 2017 Hogrefe Publishing

W. H. O’Brien et al., Job Satisfaction and RSA

Model (Thayer & Lane, 2000), which states that parasympathetic functioning indexed by RSA provides an individual with a better ability to function, both psychologically and physiologically, in a complex environment (Jarczok et al., 2013). Furthermore, according to the Polyvagal Theory forwarded by Porges (2007), higher levels of RSA reactivity would indicate the presence of poorer parasympathetic “braking.” Poorer parasympathetic braking, in turn, would cause a person to experience both higher and more prolonged cardiovascular reactivity to daily stressors. As outlined by the recurrent cardiovascular reactivity hypotheses (Chida & Steptoe, 2010; Schuler & O’Brien, 1997), this pattern of higher RSA reactivity and incomplete recovery could lead to the development of both adverse mental and physical health outcomes over time. According to this model, when a person generates higher levels of cardiovascular reactivity to daily stressors and/or experiences delayed recovery from these stressors, there is an accumulation of adverse physiological impact that promotes the development of atherosclerosis, hypertension, and cardiovascular disease. There are three limitations that merit discussion. First, the sample size was small which limited our power to reject the null hypothesis with some of the follow-up pairwise comparisons which had Bonferroni adjusted alpha levels. Second, respiration was not measured or controlled for in the analyses. Thus, it is possible that the observed RSA effects stemmed not from variation in parasympathetic activation, but alteration in rates of breathing among the low JDI group relative to the high JDI group. Third, the measures were collected at a single point in time. Thus, it is not possible to rule out a reversed causal account for these findings. For example, it is possible that persons with higher levels of RSA were more able to tolerate work stress and/or view their work as more satisfying. Similarly, at a methodological level, completing questionnaires assessing job satisfaction could have primed persons with lower job satisfaction to generate more distressing thoughts during the worry induction. However, the thought sampling data indicated that participants rarely reported worrisome thoughts during the baseline period suggesting that this effect, if present, was modest. Further research controlling for respiration effects (e.g., instructing that participants do paced breathing or collect respiration measurement and use as a covariate), reactivity of measurement, using larger samples, and longitudinal designs would be beneficial in this research arena. Regardless of the causal direction (i.e., job satisfaction > RSA or RSA > job satisfaction) the current findings are important because it is one of a very few published demonstrations of a reliable link between job satisfaction, physiological reactivity to stress, and risk for disease. As such, it is reasonable to argue, as did Faragher et al. (2005) that work Ó 2017 Hogrefe Publishing

characteristics associated with, and/or work interventions targeting improvement of, job satisfaction can exert important salutatory effects on worker health and resilience. Ethics and Disclosure Statements All participants of the study provided written informed consent and the study was approved by the Ethics Committee at Bowling Green State University. All authors disclose no actual or potential conflicts of interest including any financial, personal, or other relationships with other people or organizations that could inappropriately influence (bias) their work.

References Allen, J. J. B., Chambers, A. S., & Towers, D. N. (2007). The many metrics of cardiac chronotropy: A pragmatic primer and a brief comparison of metrics. Biological Psychology, 74, 243–262. https://doi.org/10.1016/j.biopsycho.2006.08.005 Behar, E., Vescio, T. K., & Borkovec, T. D. (2005). The effects of suppressing thoughts and images about worrisome stimuli. Behavior Therapy, 36, 289–298. https://doi.org/10.1016/ S0005-7894(05)80077-2 Bylsma, L., Solomon, K., Taylor-Clifft, A., Morris, B., & Rottenberg, J. (2014). Respiratory sinus arrhythmia in current and remitted depressive disorder. Psychosomatic Medicine, 76, 66–73. Chida, Y., & Steptoe, A. (2010). Greater cardiovascular responses lab mental stress are associated with poor subsequent cardiovascular risk status: A meta-analysis of prospective evidence. Hypertension, 55, 1026–1032. https://doi.org/ 10.1161/HYPERTENSIONAHA.109.146621 Clays, E., De Bacquer, D., Crasset, V., Kittel, F., De Smet, P., Kornitzer, M., . . . De Backer, G. (2011). The perception of work stressors is related to reduced parasympathetic activity. International Archives of Occupational and Environmental Health, 84, 185–191. https://doi.org/10.1007/s00420-010-0537-z Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Gray-Stanley, J. A., Muramatsu, N., Heller, T., Hughes, S., Johnson, T. P., & Ramirez-Valles, J. (2010). Work stress and depression among direct support professionals: The role of work support and locus of control. Journal of Intellectual Disabilities Research, 54, 749–761. https://doi.org/10.1111/ j.1365-2788.2010.01303.x Faragher, E. B., Cass, M., & Cooper, C. (2005). The relationship between job satisfaction and health: A meta-analysis. Occupational and Environmental Medicine, 62, 105–112. https://doi. org/10.1136/oem.2002.006734 Hatton, C., Rashes, R., Caine, A., & Emerson, E. (1995). Stressors, coping strategies and stress-related outcomes amongst direct care staff in staffed houses for people with learning disabilities. Journal of Applied Research in Developmental Disabilities, 8, 252–271. Hopkins, K. D., & Weeks, D. C. (1990). Tests for normality and measures of skewness and kurtosis: Their place in research reporting. Educational and Psychological Measurement, 50, 717–729. https://doi.org/10.1177/0013164490504001 Jarczok, M. N., Jarczok, M., Mauss, D., Koenig, J., Li, J., Herr, R. M., & Thayer, J. F. (2013). Autonomic nervous system activity and workplace stressors – A systematic review. Neuroscience & Biobehavioral Reviews, 37, 1810–1823. https://doi.org/ 10.1016/j.neubiorev.2013.07.004

Journal of Psychophysiology (2019), 33(1), 32–38

Larson, S. A., Lakin, K. C., & Bruininks, R. H. (1998). Staff recruitment and retention: Study results and intervention strategies. Washington, DC: American Association on Mental Retardation. Lunsky, Y., Hastings, R. P., Hensel, J., Arenovich, T., & Dewa, C. S. (2014). Perceptions of Positive Contributions and Burnout in Community Developmental Disability Workers. Intellectual and Developmental Disabilities, 52, 249–257. https://doi.org/ 10.1352/1934-9556-52.4.249 Masi, C. M., Hawkley, L. C., Rickett, E. M., & Cacioppo, J. T. (2007). Respiratory sinus arrhythmia and diseases of aging: Obesity, diabetes, mellitus, and hypertension. Biological Psychology, 74, 212–223. https://doi.org/10.1016/j.biopsycho.2006.07.006 McLaughlin, K. A., Borkovec, T. D., & Sibrava, N. J. (2007). The effects of worry and rumination on affect states and cognitive activity. Behavior Therapy, 38, 23–38. https://doi.org/10.1016/ j.beth.2006.03.003 Mitchell, G., & Hastings, R. P. (2001). Coping, burnout, and emotion in staff working in community services for people with challenging behaviors. American Journal of Mental Retardation, 106, 448–459. https://doi.org/10.1352/0895-8017(2001)106% 3C0448:CBAEIS%3E2.0.CO;2 Mutkins, E., Brown, E. B., & Thorsteinsson, E. B. (2011). Stress, depression, workplace and social supports and burnout in intellectual disability support staff. Journal of Intellectual Disability Research, 5, 500–510. https://doi.org/10.1111/ j.1365-2788.2011.01406.x Orsila, R., Virtanen, M., Luukkaala, T., Tarvainen, M., Karjalainen, P., Viik, J., & Nygård, C. H. (2008). Perceived mental stress and reactions in heart rate variability – A pilot study among employees of an electronics company. International Journal of Occupational Safety and Ergonomics, 14, 275–283. https://doi. org/10.1080/10803548.2008.11076767 Porges, S. W. (2007). The polyvagal perspective. Biological Psychology, 74, 116–143. https://doi.org/10.1016/j.biopsycho. 2006.06.009 Porges, S. W., & Boher, R. E. (1990). Analyses of periodic processes in psychophysiological research. In J. T. Caccioppo & L. G. Tassinary (Eds.), Principles of psychophysiology: Physical, social, and inferential elements (pp. 708–753). New York, NY: Cambridge University Press. Rose, J. (1993). Stress and staff in residential settings: The move from hospital to the community. Mental Handicap Research, 6, 312–332. https://doi.org/10.1111/j.1468-3148.1993.tb00062.x Rose, D., Home, S., Rose, J. L., & Hastings, R. P. (2004). Negative emotional reactions to challenging behavior and staff burnout: Two replication studies. Journal of Applied Research in Intellectual Disabilities, 17, 219–223. https://doi.org/10.1111/ j.1468-3148.2004.00194.x Schuler, J. L. H., & O’Brien, W. H. (1997). Cardiovascular recovery from stress and hypertension risk factors: A meta-analytic review. Psychophysiology, 34, 649–659. https://doi.org/ 10.1111/j.1469-8986.1997.tb02141.x Skirrow, P., & Hatton, C. (2007). “Burnout” amongst direct care workers in services for adults with intellectual disabilities: A systematic review of research findings and initial normative data. Journal of Applied Research in Intellectual Disabilities, 20, 131–144. https://doi.org/10.1111/j.1468-3148.2006.00311.x

Journal of Psychophysiology (2019), 33(1), 32–38

W. H. O’Brien et al., Job Satisfaction and RSA

Stanton, J., Sinar, E., Balzer, W., Julian, A., Thorensen, P., & Aziz, S. (2002). Development of a compact measure of job satisfaction: The abridged job descriptive index. Educational and Psychological Measurement, 62, 173–191. https://doi.org/ 10.1177/001316440206200112 Suadicani, P., Hein, H., & Gyntelberg, F. (1993). Are social inequalities as associated with the risk of ischaemic heart disease a result of psychosocial working conditions? Atherosclerosis, 101, 165–175. https://doi.org/10.1016/00219150(93)90113-9 Thayer, J. F., Åhs, F., Fredrikson, M., Sollers, J. J., & Wager, T. D. (2012). A meta-analysis of heart rate variability and neuroimaging studies: Implications for heart rate variability as a marker of stress and health. Neuroscience & Biobehavioral Reviews, 36, 747–756. https://doi.org/10.1016/j.neubiorev.2011.11.009 Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61, 201–216. https://doi.org/10.1016/ S0165-0327(00)00338-4 Toivanen, H., Länsimies, E., Jokela, V., & Hänninen, O. (1993). Impact of regular relaxation training on the cardiac autonomic nervous system of hospital cleaners and bank employees. Scandinavian Journal of Work, Environment & Health, 19, 319– 325. https://doi.org/10.5271/sjweh.1468 Vrijkotte, T. G., Van Doornen, L. J., & De Geus, E. J. (2000). Effects of work stress on ambulatory blood pressure, heart rate, and heart rate variability. Hypertension, 35, 880–886. https://doi. org/10.1161/01.HYP.35.4.880 Watanabe, T., Sugiyama, Y., Sumi, Y., Watanabe, M., Takeuchi, K., Kobayashi, F., & Kono, K. (2002). Effects of vital exhaustion on cardiac autonomic nervous functions assessed by heart rate variability at rest in middle-aged male workers. International Journal of Behavioral Medicine, 9, 68–75. https://doi.org/ 10.1207/S15327558IJBM0901_05 Yaptangco, M., Crowell, S., Baucom, B., Bride, D., & Hansen, E. (2015). Examining the relationship between respiratory sinus arrhythmia and depressive symptoms in emerging adults: A longitudinal study. Biological Psychology, 110, 34–41. https:// doi.org/10.1016/j.biopsycho.2015.06.004 Zanstra, Y. J., Schellekens, J. M., Schaap, C., & Kooistra, L. (2006). Vagal and sympathetic activity in burnouts during a mentally demanding workday. Psychosomatic Medicine, 68, 583–590. https://doi.org/10.1097/01.psy.0000228012.38884.49

Received May 16, 2016 Accepted April 7, 2017 Published online November 21, 2017

William H. O’Brien Department of Psychology Bowling Green State University Bowling Green, OH 43403 USA wobrien@bgsu.edu

Ó 2017 Hogrefe Publishing

Article

Autonomic Cardiovascular Dysregulation at Rest and During Stress in Chronically Low Blood Pressure Stefan Duschek,1 Alexandra Hoffmann,1 Casandra I. Montoro,1 and Gustavo A. Reyes del Paso2 UMIT – University of Health Sciences Medical informatics and Technology, Institute of Psychology, Hall in Tirol, Austria

1 2

Department of Psychology, University of Jaén, Spain

Abstract: Chronic low blood pressure (hypotension) is accompanied by symptoms such as fatigue, reduced drive, faintness, dizziness, cold limbs, and concentration difficulties. The study explored the involvement of aberrances in autonomic cardiovascular control in the origin of this condition. In 40 hypotensive and 40 normotensive subjects, impedance cardiography, electrocardiography, and continuous blood pressure recordings were performed at rest and during stress induced by mental calculation. Parameters of cardiac sympathetic control (i.e., stroke volume, cardiac output, pre-ejection period, total peripheral resistance), parasympathetic control (i.e., heart rate variability), and baroreflex function (i.e., baroreflex sensitivity) were obtained. The hypotensive group exhibited markedly lower stroke volume, heart rate, and cardiac output, as well as higher pre-ejection period and baroreflex sensitivity than the control group. Hypotension was furthermore associated with a smaller blood pressure response during stress. No group differences arose in total peripheral resistance and heart rate variability. While reduced beta-adrenergic myocardial drive seems to constitute the principal feature of the autonomic impairment that characterizes chronic hypotension, baroreflex-related mechanisms may also contribute to this state. Insufficient organ perfusion due to reduced cardiac output and deficient cardiovascular adjustment to situational requirements may be involved in the manifestation of bodily and mental symptoms. Keywords: hypotension, blood pressure, autonomic control, baroreflex, heart rate variability, cardiac output

Introduction The term chronic hypotension refers to a persistent state of inappropriately low blood pressure independent of the occurrence of further pathological conditions (De Buyzere, Clement, & Duprez, 1998). According to WHO (1978) criteria, hypotension is diagnosed when systolic blood pressure falls below 100 mmHg in women and 110 mmHg in men. The chronic form is distinguished from orthostatic hypotension (i.e., circulatory problems when assuming an upright position) and symptomatic hypotension, which occurs, for example, due to blood loss or medication (Freeman et al., 2011). The prevalence of chronic hypotension has been estimated at 2–3% in the general population with women being predominantly affected (Pemberton, 1989). In contrast to elevated blood pressure, chronic hypotension is commonly not regarded as a dangerous medical condition requiring treatment (De Buyzere et al., 1998). While hypertension Ó 2017 Hogrefe Publishing

constitutes a major risk factor for cardiovascular diseases, low blood pressure is associated with reduced cardiovascular mortality (Prospective Studies Collaboration, 2002). Nonetheless, affected individuals frequently report complaints including fatigue, reduced drive, dizziness, headaches, and cold limbs (Pilgrim, 1994; Rosengren, Tibblin, & Wilhelmsen, 1993; Wessely, Nickson, & Cox, 1990). A number of studies furthermore demonstrated cognitive impairment in affected individuals, particularly in the fields of attention and memory, which has been ascribed to suboptimal cerebral blood flow regulation and blunted cortical activity (Duschek, Meinhardt, & Schandry, 2006; Duschek & Schandry, 2004, 2006; Duschek, Weisz, & Schandry, 2003). Aberrances in autonomic cardiovascular control were hypothesized to be involved in the etiology of chronic hypotension (Covassin, de Zambotti, Cellini, Sarlo, & Stegagno, 2013). Previous data pointed toward a reduction Journal of Psychophysiology (2019), 33(1), 39–53 https://doi.org/10.1027/0269-8803/a000204

in cardiac sympathetic outflow in hypotensive versus normotensive subjects (Duschek, Heiss, et al., 2009). The estimation of hemodynamic parameters from continuous blood pressure recording (“Modelflow analysis”; Wesseling, Jansen, Settels, & Schreuder, 1993) revealed diminished stroke volume (SV) and cardiac output (CO) in a hypotensive sample at rest and under conditions of mental stress. Considering the beta-adrenergic innervation of the ventricles, the finding suggests lower cardiac contractility due to diminished beta-adrenergic activity (Berntson, Quigley, Norman, & Lozano, 2016; Levy & Pappano, 2007). As a restriction, it should be noted that the reduction in CO – that is given by the product of SV and heart rate – was partly due to lower heart rate, which underlies both sympathetic and parasympathetic control. In addition, the precision of the Modelflow technique in hemodynamic recording is certainly limited (Bogert & Lieshout, 2005). Another line of research defined characteristics of autonomic cardiovascular control in hypotension during sleep. Impedance cardiography revealed lower nocturnal CO and higher values of pre-ejection period (PEP) and left ventricular ejection time (LVET) in affected individuals (Covassin, de Zambotti, Cellini, Sarlo, & Stegagno, 2012; de Zambotti, Covassin, Cellini, Sarlo, Torre, et al., 2012). While LVET cannot be unambiguously related to the sympathetic system, PEP is inversely related to beta-adrenergic activity; as such, its increase is consistent with the notion of reduced myocardial sympathetic outflow (Cacioppo et al., 1994; Hassan & Turner, 1983). By definition, the findings obtained during sleep are not necessarily generalizable to the waking state. Increased parasympathetic tone may also contribute to persistently low blood pressure (Duschek & Schandry, 2007). Vagal cardiac influence is commonly quantified by means of heart rate variability (HRV) analysis (Berntson et al., 2016). Consistent with the present assumption, hypotensive individuals investigated during sleep exhibited higher HRV, as indexed by the root mean square of successive differences in heart cycle duration (RMSSD) and the proportion of cycle lengths differing by more than 50 ms (pNN50) (Covassin et al., 2012). In contrast, nocturnal HRV in the high frequency band, another popular index of vagal cardiac outflow (Berntson et al., 2016), did not differ between hypotensive and normotensive samples (de Zambotti, Covassin, Cellini, Sarlo, Torre, et al., 2012). The same study documented reduced sympathovagal balance in terms of a lower ratio between the power in the low and high frequency bands of the HRV spectrum (LF/HF ratio) in hypotensives. However, the validity of the LF/HF ratio remains highly questionable (Reyes del Paso, Langewitz, Mulder, van Roon, & Duschek, 2013); as such, the findings on alterations in HRV remain equivocal

Journal of Psychophysiology (2019), 33(1), 39–53

S. Duschek et al., Autonomic control in chronic hypotension

and the available evidence base does not allow firm conclusions to be drawn. The cardiac baroreflex is an additional, potentially relevant regulatory mechanism in the present context. This medullary reflex consists of a negative feedback loop, in which activity changes in arterial baroreceptors, resulting from blood pressure fluctuations, precipitate compensatory changes in heart rate and myocardial contractility (Berntson et al., 2016; Duschek, Werner, & Reyes del Paso, 2013). While the baroreflex is regarded as the most important control system in buffering short-term blood pressure fluctuations, recent research also confirmed its involvement in the setting of tonic blood pressure and the origin of essential hypertension (Carthy, 2013; Lohmeier, Irwin, Rossing, Serdar, & Kieval, 2004). Pilot data suggested the occurrence of increased baroreflex sensitivity (BRS) in chronic hypotension (Duschek, Dietel, Schandry, & Reyes del Paso, 2008). It has been suggested that elevated responsiveness of the reflex leads to overcompensation of phasic blood pressure increases and thus stabilization of blood pressure at a lower level. However, to date, enhanced BRS was observed only in a single study; thus, replication of this finding is certainly required. With the goal of increasing our etiological knowledge about chronic hypotension, and taking into account the methodological limitations of the available research, the current study comprehensively quantified autonomic cardiovascular control in chronic hypotension. SV, CO, PEP, and total peripheral resistance (TPR) were obtained using impedance cardiography. While cardiac contractility is controlled by the beta-adrenergic system, TPR is mainly linked to alpha-adrenergic effects (Levy & Pappano, 2007). Even though SV is a well-established contractility index, to some extent its magnitude also depends on heart rate. Higher heart rate is related to lower ventricular preload, which, mediated by the Frank-Starling mechanism, results in lower SV (Kenny, Plappert, Doubilet, Salzman, & Sutton, 1987; Levy & Pappano, 2007). Considering this, heart rate was controlled in the statistical analysis pertaining to SV. The RMSSD derived from electrocardiography (ECG) recordings was taken as an index of HRV. This parameter represents HRV in the high frequency range and its validity as an indicator of vagal cardiac tone has been repeatedly shown (Berntson, Lozano, & Chen, 2005; Berntson et al., 2016). Finally, BRS was extrapolated from continuous blood pressure recordings using sequence analysis. This method allows estimation of BRS in the time domain based on the spontaneous covariation between systolic blood pressure and heart cycle duration (Bertinieri et al., 1985; Reyes del Paso, González, & Hernández, 2010). In order to obtain a more comprehensive picture, cardiovascular recordings were carried out under resting conditions and during

Ó 2017 Hogrefe Publishing

S. Duschek et al., Autonomic control in chronic hypotension

exposure to a mental challenge. This procedure also allowed for evaluation of the magnitude of cardiovascular reactivity during stress conditions, where previous research suggested blunted reactivity in chronic hypotension and thus poorer short-term autonomic cardiovascular adjustment to situational requirements (Covassin et al., 2013; Duschek, Dietel, et al., 2008; Duschek, Matthias, & Schandry, 2005). The following main hypotheses were tested in the study: Hypothesis 1 (H1): Taking into account the research delineated above, reduced beta-adrenergic cardiac tone in chronic low blood pressure was postulated. This was expected to be expressed in lower SV and CO, and in a longer PEP in hypotensive versus normotensive individuals. Hypothesis 2 (H2): Though the available research remains inconsistent, augmented parasympathetic influences on heart rate, indicated by higher expressions of RMSSD, were predicted in hypotension. Hypothesis 3 (H3): Based on our pilot data, greater BRS in hypotensive versus normotensive participants was also predicted. Hypothesis 4 (H4): Finally, reduced autonomic stress reactivity, in terms of smaller modulations in the assessed parameters during mental challenge, was postulated.

Methods Participants Forty subjects with hypotension, according to WHO (1978) criteria, and 40 normotensive control persons participated (35 women and 5 men in each group). None of the subjects suffered from a relevant physical disease or mental disorder. Health status was assessed by means of an anamnestic interview and a questionnaire covering diseases of the cardiovascular, respiratory, gastrointestinal, and urogenital systems, and of the thyroid and the liver, as well as metabolic diseases and psychiatric disorders. None of the participants used any kind of medication affecting the cardiovascular or central/ peripheral nervous system. In total, 67 of the participants were university students (34 in the hypotensive sample, 33 in the control group); the remaining subjects were in the workforce. Table 1 provides information about blood pressure, as recorded at the beginning of the experimental procedure, as well as age and body mass index (BMI). Sample size was determined based on previous studies comparing parameters of autonomic control between Ó 2017 Hogrefe Publishing

Table 1. Means (M) and standard deviations (SD) of systolic blood pressure, diastolic blood pressure, age, and body mass index in both samples Hypotensive group

Normotensive group

Systolic blood pressure (mmHg)

95.41

6.69

119.18

3.82

Diastolic blood pressure (mmHg)

64.59

6.30

77.05

5.48

Age (years)

24.70

4.90

23.98

4.28

Body mass index (kg/m2)

20.15

1.85

21.95

2.78

hypotensive and normotensive individuals (c.f. Introduction), which revealed effect sizes (Cohen’s d) generally between .3 and .5. Assuming an effect size of .4, an α level of 5%, and a β error of 20%, power analysis revealed a required total sample size of 39 per group. The study was part of a larger project investigating psychophysiological aspects of chronic hypotension. Further results obtained in this sample are presented in Duschek, Hoffmann, and Reyes del Paso (2017) and Duschek, Hoffmann, Reyes del Paso, and Ettinger (2017).

Hemodynamic Recordings For impedance cardiography recording, a CardioScreen 1000 (Medis Inc., Ilmenau, Germany) device was employed (c.f. Berntson et al., 2016; Moshkovitz, Kaluski, Milo, Vered, & Cotter, 2004; Raaijmakers, Faes, Scholten, Goovaerts, & Heethaar, 1999 for technical framework and discussion of method reliability and validity). The impedance signal was acquired using four spot electrodes positioned at the lateral neck and the lateral chest (left side) with alternating current of 1.5 mA and 85 kHz. The ECG was recorded at a sampling rate of 1,000 Hz from two electrodes placed at the left mid-clavicle and lowest right rib using a Biopac system (MP 150, Biopac Systems Inc., Goleta, CA). The back of the left hand served as a ground. Blood pressure was monitored continuously with a Finometer Model-2 (Finapres Medical Systems, Amsterdam, The Netherlands). The cuff of the device was applied to the left index finger and that hand was positioned at the level of the heart. For periodic recalibration, the device’s “Physiocal” feature (Wesseling, De Wit, Van der Hoeven, Van Goudoever, & Settels, 1995) was in operation. Data were recorded by means of the Biopac system (sampling rate 1,000 Hz).

Procedure Assignment of subjects to the two study groups was carried out on the basis of blood pressure readings taken in a Journal of Psychophysiology (2019), 33(1), 39–53

screening session, which was conducted at least one week prior to the main experiment and again at the beginning of the experimental session. Here, after a rest period of 10 min, three sphygmomanometric blood pressure measurements were taken in a sitting position. For this purpose an automatic inflation blood pressure monitor (Omron M400, Omron Healthcare, Vernon Hills, IL) was used. Readings were separated by 5-min rest intervals. The mean value of the three measurements was used for group assignment. Females with a mean systolic blood pressure of less than 100 mmHg, and males with a mean value below 110 mmHg, were assigned to the hypotensive group. The inclusion criterion for the control group was systolic blood pressure between 115 and 140 mmHg. The criteria had to be fulfilled at both the screening and experimental sessions. A total of 12 prospective participants who met the inclusion criteria for the hypotensive group during screening no longer satisfied them at the experimental session. For the same reason, 17 candidates for the control group had to be excluded. Hemodynamic recordings were accomplished at rest and during mental stress induced by a serial subtraction task. During the 7-min resting phase participants were asked to sit still, not to speak and to relax with their eyes open. The subtraction task was conducted over a 3-min period, during which subjects had to count down from 1,000, subtracting 17 each time and saying the numbers out loud. They were asked to perform the task as quickly and accurately as possible. Experimental sessions were conducted in the morning, between 8 and 11 a.m., and in the afternoon between 2 and 5 p.m. In order to control for circadian effects, the same number of participants from both study groups was tested in the morning and afternoon. Participants were requested not to drink alcohol or beverages containing caffeine for 3 hr prior to the screening and experimental sessions. The study was approved by the Board for Ethical Questions in Science of the University of Innsbruck/Austria and all participants provided their written informed consent.

Data Analysis The data revealed by impedance cardiography was processed using Cardio-Vascular-Lab software (Medis Inc.). SV (mL) was obtained by applying the Kubicek equation (Berntson et al., 2016), and CO (L/min) was computed by multiplying SV by heart rate. PEP (ms) was defined as the period between the onset of ventricular depolarization (Q-point in ECG) and the onset of left ventricular ejection (B-point in impedance signal). TPR (dyn s/cm5) was computed as (MAP/CO) 80, where MAP represents mean arterial pressure obtained by continuous finger Journal of Psychophysiology (2019), 33(1), 39–53

S. Duschek et al., Autonomic control in chronic hypotension

recording. The software provides values of the indices on a beat-to-beat basis. The data were processed further in two different ways: firstly, mean values across the resting and mental stress conditions were computed. In addition, to explore the time course of the changes in the parameters during stress induction, mean values were computed for consecutive intervals of 10 s duration each, resulting in 18 values for the mental stress period. To aggregate the ECG data, KARDIA software (Perakakis, Joffily, Taylor, Guerra, & Vila, 2010) was employed. In a first step, values of heart cycle duration – defined as the RR interval – were computed from the raw ECG. The ECG data were visually screened, and artifacts were corrected by linear interpolation. Subsequently, values of the RMSSD (ms) were computed for the entire resting and stress conditions, and for each of the 18 consecutive 10 s intervals of the mental stress period. BRS was quantified using software developed by one of the authors (Reyes del Paso, 1994). The program locates sequences of three to six consecutive heart cycles (reflex sequences) in which systolic blood pressure increases are accompanied by increases in heart cycle duration, and those in which blood pressure decreases are accompanied by decreases in heart cycle duration. Since a time lag of one heartbeat is known to produce the best estimates of BRS (Steptoe & Vögele, 1990), each systolic value was paired with the duration of the cycle immediately following. Values of 1 mmHg and 2 ms were applied as minimal criteria for changes in blood pressure and heart cycle duration, respectively. The interval between systolic points (intersystolic interval) was taken as an index of heart cycle duration (Reyes del Paso et al., 2010). When one of these reflex sequences was detected, the regression line was computed across all heart cycles of the given sequence. BRS was expressed as the change in heart cycle duration (ms) per mmHg blood pressure change, measured by the slope of the regression line. The reliability and validity of this method has been well established (Duschek, Werner, et al., 2013; Reyes del Paso, 1994). Values of BRS (ms/mmHg) were computed for the entire resting and stress conditions and, in addition, for consecutive 30 s intervals of the stress period. Longer intervals were chosen for BRS than for the remaining parameters in order to enable reliable estimation of this parameter. As a main instrument of statistical analysis, analysis of variance (ANOVA) models were computed with experimental group (hypotensive vs. control group) as betweensubjects factor and experimental condition (rest vs. mental stress) as the within-subjects factor. The cardiovascular indices served as dependent variables. To account for the possible dependency of SV on heart rate, an additional model was computed for SV, in which baseline heart rate was applied as a covariate. Post hoc F-tests were conducted Ó 2017 Hogrefe Publishing

S. Duschek et al., Autonomic control in chronic hypotension

for all variables with significant group by condition interactions. Within both groups, values of the resting condition were compared with those of the stress period. To describe the time course of the cardiovascular responses during stress induction, change scores were calculated as the difference between the values computed for the consecutive 10 or 30 s intervals of the stress period and corresponding mean values of the preceding rest phase. For statistical analysis, further ANOVA models were computed. Here, group served as a between-subjects factor and time (i.e., the 6 change scores of BRS and the 18 change scores of the remaining variables) as a withinsubjects factor. Alpha was set at .05 in all analyses.

Results Figure 1 depicts the results for the indices derived from impedance cardiography, as well as heart rate (taken from the ECG recordings). SV, heart rate, and CO were lower in the hypotensive group than in the normotensive control group, whereas PEP was higher. Only a small group difference arose in TPR. While the values of heart rate and CO were higher during stress than at rest, SV and PEP decreased during the mental calculation task; TPR did not systematically change. The ANOVA revealed a significant main effect of group for all parameters except TPR [SV: F(1, 78) = 10.22, p < .01, η2p = .12; heart rate: F(1, 78) = 6.88, p = .010, η2p = .081; CO: F(1, 78) = 47.85, p < .01, η2p = .38; TPR: F(1, 78) = 0.33, p = .57, η2p < .01; PEP: F(1, 78) = 7.69, p < .01, η2p = .090]. The main effect of condition was significant for all variables except TPR [SV: F(1, 78) = 18.13, p < .01, η2p = .19; heart rate: F(1, 78) = 89.80, p < .01, η2p = .54; CO: F(1, 78) = 58.45, p < .01, η2p = .43; TPR: F(1, 78) < .01, p = .98, η2p < .01; PEP: F(1, 78) = 74.42, p < .01, η2p = .49]. Significant group and condition effects were also observed in the additional model for SV, in which baseline heart rate was controlled for [group: F(1, 77) = 29.51, p < .01, η2p = .28; condition: F(1, 77) = 5.82, p = .018, η2p = .070]. The size of the group effect was higher than in the model in which heart rate was not controlled for. The effect of the covariate was also significant, F(1, 77) = 37.30, p < .01, η2p = .33. Significant group by condition effects did not arise in any of these models. In Figure 2 the values of RMSSD, BRS, and blood pressure, taken from continuous finger measurements, 1

are displayed. While only a small group difference arose in the RMSSD, BRS was larger in hypotensive versus normotensive subjects. BRS was lower during the calculation task than at rest and systolic and diastolic blood pressure was higher. However, the extent of the blood pressure increase from rest to task was larger in the normotensive than in the hypotensive group. The group and condition effects were significant for BRS and systolic and diastolic blood pressure, but not for the RMSSD [group: BRS, F(1, 78) = 4.99, p = .028, η2p = .060; RMSSD, F(1, 78) = 0.27, p = .61, η2p < .01; systolic blood pressure, F(1, 78) = 52.62, p < .01, η2p = .40; diastolic blood pressure, F(1, 78) = 21.61, p < .01, η2p = .22; condition: BRS, F(1, 78) = 37.83, p < .01, η2p = .33; RMSSD, F(1, 78) = 0.027, p = .87, η2p < .01; systolic blood pressure, F(1, 78) = 94.57, p < .01, η2p = .55; diastolic blood pressure, F(1, 78) = 126.60, p < .01, η2p = .62]. Among these models, significant interactions between group and condition were only found in those computed for blood pressure [systolic blood pressure: F(1, 78) = 8.48, p < .01, η2p = .098; diastolic blood pressure: F(1, 78) = 8.03, p < .01, η2p = .093]. Post hoc testing conducted separately for the hypotensive and normotensive participants revealed stress-induced increases in systolic and diastolic blood pressure in both groups; however, the effect sizes were larger in the normotensive versus hypotensive group [hypotension: systolic blood pressure, F(1) = 32.51, p < .01, η2p = .46; diastolic blood pressure, F(1) = 53.86, p < .01, η2p = .58; control group: systolic blood pressure, F(1) = 62.08, p < .01, η2p = .61; diastolic blood pressure, F(1) = 73.91, p < .01, η2p = .66]. The time courses of the changes in the cardiovascular parameters during the stress period are depicted in Figures 3 and 4. SV decreased during the first 30 s of the task period in both groups; thereafter, it progressively rose but remained below the baseline level during almost the entire period. The ANOVA revealed a main effect of time for SV, F(17, 76) = 3.09, p = .011, η2p = .039.1 CO increased in both groups, peaked at the end of the first task minute, and decreased during the remaining period [time effect: F(17, 76) = 3.19, p < .01, η2p = .040]. PEP strongly declined in both groups, reached its minimum around 60 s after task onset, and remained at a low level toward the end of the period [time effect: F(17, 76) = 12.51, p < .01, η2p = .14]. TPR gradually rose beginning around 50 s after task onset; the increase was somewhat steeper in the normotensive versus hypotensive group [time effect: F(17, 76) = 17.12, p < .01, η2p = .18]. Both groups exhibited a steep initial heart rate increase, which reached its maximum at 20 s after task

The data revealed by impedance cardiography and HRV analysis from two subjects had to be excluded from the time course analysis due to missing values for certain seconds of the task interval (the noncontinuous analysis presented above was based on the available data; as such, the number of included values was somewhat reduced in these subjects). The BRS data of three subjects had to be excluded because no baroreflex sequences occurred in some of the 30 s intervals (for the analysis pertaining to the entire stress period, the number of baroreflex sequences was sufficient to reliably estimate BRS in all subjects).

Ó 2017 Hogrefe Publishing

Journal of Psychophysiology (2019), 33(1), 39–53

S. Duschek et al., Autonomic control in chronic hypotension

Stroke volume (ml)

Hypotension Control group 100

Cardiac output (l/min) 8 7

6 80

4 3

2 1

50 Rest

1100

Rest

Mental stress

Total peripheral resistance (dyn x s/cm5)

1000

Mental stress

Pre-ejection period (ms) 140 130

900 120 800 110

700 600

100 Rest

Mental stress

Rest

Mental stress

Heart rate (beats/min)

95 90 85 80 75 70 65 60 Rest

Mental stress

Figure 1. Mean values of stroke volume, cardiac output, total peripheral resistance, pre-ejection period, and heart rate (bars denote standard errors of the mean).

onset. Between 40 s and 100 s, heart rate decreased once again, and remained stable until the end of the period [time effect: F(17, 78) = 20.60, p < .01, η2p = .21]. While BRS progressively declined in both groups during the entire task, [time effect: F(5, 75) = 6.87, p < .01, η2p = .084], RMSSD decreased during the initial task phase and thereafter gradually increased [time effect: F(17, 76) = 6.40, p < .01, η2p = .078]. Systolic and diastolic blood pressure progressively rose during the task in both groups [time effect systolic blood pressure: F(17, 78) = 47.26, p < .01, η2p = .38; time effect diastolic blood pressure: F(17, 78) = 44.28, p < .01, η2p = .36]. However, the slope of the increase was markedly higher in the control group versus the hypotensive group, as confirmed by significant group and interaction effects [group effect of systolic blood pressure: F(1, 78) = 8.31, Journal of Psychophysiology (2019), 33(1), 39–53

p < .01, η2p = .096; group effect of diastolic blood pressure: F(1, 78) = 6.24, p = .015, η2p = .074; interaction effect of systolic blood pressure: F(17, 78) = 4.51, p < .01, η2p = .055; and interaction effect of diastolic blood pressure: F(17, 78) = 3.30, p < .04, η2p = .041]. No group or interaction effects were obtained for the remaining variables.

Discussion With the aim of extending our knowledge regarding the relevance of autonomic dysregulation to chronic hypotension, this study compared parameters of sympathetic and parasympathetic cardiovascular control and baroreflex Ó 2017 Hogrefe Publishing

S. Duschek et al., Autonomic control in chronic hypotension

Hypotension Control group 19

Baroreflex senstitivity (ms/mmHg)

RMSSD (ms) 60

0 Rest

Mental stress

Rest

Diastolic blood pressure (mmHg)

Systolic blood pressure (mmHg)

130

120

110

Mental stress

100 50

70 Rest

Mental stress

Rest

Mental stress

Figure 2. Mean values of baroreflex sensitivity, root mean square of successive differences in heart cycle duration (RMSSD), systolic and diastolic blood pressure (bars denote standard errors of the mean).

function between individuals with blood pressure in the hypotensive and normotensive ranges. While hypotension was associated with markedly lower SV, heart rate, and CO, as well as with higher PEP and BRS, no group differences arose in TPR and high frequency HRV indexed by the RMSSD. According to these results, reduced betaadrenergic myocardial drive and increased baroreflex function stand to reason as being the principal components of the autonomic impairment associated with chronic hypotension. The study confirmed previous findings of reductions in SV and CO in chronic hypotension at rest and during mental stress (Duschek, Heiss, et al., 2009). These findings, however, had to be regarded as preliminary, given that SV was derived from blood pressure waveforms using Modelflow analysis (Wesseling, Jansen, Settels, & Schreuder, 1993). While this method allows relatively precise assessment of intra-individual hemodynamic changes, it provides only a rough estimation of absolute values; as such, its suitability for the assessment of interindividual differences in hemodynamic function is confined to tight limits (Bogert & Lieshout, 2005). For the latter purpose, the impedance cardiography presently applied is clearly more appropriate (Moshkovitz et al., 2004). This method provides reliable hemodynamic measures, which are highly correlated with those of echocardiography or invasive techniques Ă&#x201C; 2017 Hogrefe Publishing

(Berntson et al., 2016; Moshkovitz et al., 2004; Raaijmakers et al., 1999). Even though SV varies subject to myocardial contractility, which underlies beta-adrenergic control, to some degree it is also influenced by heart rate (Berntson et al., 2016; Hassan & Turner, 1983). Therefore, heart rate was statistically controlled for in the present analysis, which led to an even larger effect size in the group comparison. PEP is inversely related to beta-adrenergic activity (Levy & Pappano, 2007); as such, its presently observed reduction also reflects lower sympathetic cardiac drive in hypotension. The diminished SV together with the lower heart rate caused a reduction in CO by 20% at rest and 18% during mental challenge in the hypotensive versus control group. It has been argued that diminished organ perfusion due to lower CO accounts for some of the complaints that typically accompany chronically low blood pressure (Covassin et al., 2012; Duschek, Heiss, et al., 2009). This may be the case, for example, for low skin temperatures and the experience of cold limbs. Reduced cerebral blood flow in hypotension may be another consequence (Duschek & Schandry, 2004, 2007). Cerebral hypo-perfusion may also be relevant in terms of the association between hypotension and geriatric cognitive disorders, which has repeatedly been reported (Heijer et al., 2003; Jochemsen et al., 2013; Muller et al., 2014; Qiu, Winblad, & Fratiglioni, 2005). Journal of Psychophysiology (2019), 33(1), 39â&#x20AC;&#x201C;53

S. Duschek et al., Autonomic control in chronic hypotension

Figure 3. Time course of the changes in stroke volume, cardiac output, total peripheral resistance, pre-ejection period, and heart rate during mental stress (absolute changes with respect to baseline; bars denote standard errors of the mean).

However, it should be noted that these studies were based on samples that did not represent the population currently addressed. In most samples in research on late-life hypotension, this state was preceded by mid-life cardiovascular and cerebrovascular disease or hypertension. These conditions may cause brain atrophy and cognitive decline in older age, for example, due to vascular changes or Journal of Psychophysiology (2019), 33(1), 39â&#x20AC;&#x201C;53

impaired cerebral autoregulation. As such, it has to be considered that behavioral and physiological alterations associated with geriatric hypotension are mediated by mechanisms different from those occurring in young individuals with chronic hypotension. Confirming our previous observation (Duschek, Heiss, et al., 2009), the current study groups did not differ in Ă&#x201C; 2017 Hogrefe Publishing

S. Duschek et al., Autonomic control in chronic hypotension

Figure 4. Time course of the changes in baroreflex sensitivity, root mean square of successive differences in heart cycle duration (RMSSD), systolic and diastolic blood pressure during mental stress (absolute changes with respect to baseline; bars denote standard errors of the mean).

TPR. The pattern of reduced SV and CO and normal TPR has been considered to be indicative of the hemodynamic origin of chronic hypotension (Covassin et al., 2012; Duschek, Heiss, et al., 2009). Accordingly, this state may mainly be caused by cardiac factors, whereas vascular regulation may only play a subordinate role. This hypothesis has been partially challenged by the study of Covassin et al. (2012) on nocturnal autonomic cardiovascular control. The comparison between hypotensive and normotensive individuals did not reveal an overall difference in TPR recorded during 7 hr of sleep; however, groups differed in the temporal profile of TPR changes. In contrast to controls, hypotensive participants demonstrated progressive rise in TPR especially during the second half of the night, which was interpreted as a compensatory response to the drop in CO which occurred over time. On this account, aberrances in the regulation of vascular tone mediated by the alpha-adrenergic system in hypotension cannot be completely ruled out. Research on autonomic dysregulation in prehypertension may also be of interest in the present context. Ó 2017 Hogrefe Publishing

Prehypertension has been defined by systolic blood pressure between 120 and 139 mmHg and diastolic blood pressure between 80 and 89 mmHg (Chobanian et al., 2003). It has been implicated in increased risk of cardiovascular disease (Qureshi, Suri, Kirmani, & Divani, 2005; Vasan, Larson, Leip, Kannel, & Levy, 2001), which is supported by meta-analytic evidence showing that the linear relationship between blood pressure and vascular mortality not only exists in hypertension, but also in the prehypertensive range (Prospective Studies Collaboration, 2002). It has been argued that sympathetic cardiac control is involved in the genesis of this state. Observations of increased norepinephrine levels, in addition to elevated CO due to enhanced cardiac contractility and higher heart rate, confirmed this notion (Davis et al., 2012; Drukteinis et al., 2007). It has been shown that prehypertension tends to rapidly progress to chronic hypertension (Vasan et al., 2001). This is consistent with the view that sympathetically mediated augmentation of CO contributes to hypertension during its early phase (Lovallo & al’Absi, 1998; Lund-Johansen, 1991). Available findings on TPR in Journal of Psychophysiology (2019), 33(1), 39–53

prehypertension are still inconsistent (Davis et al., 2012; Drukteinis et al., 2007), whereas it has been well established that augmented TPR constitutes the most relevant hemodynamic aberrance in chronically elevated blood pressure, at least during the chronic phase (Cain & Khalil, 2002). Regarding the present study, it may be noteworthy that the inclusion criteria of the control group also encompassed the prehypertensive range. Therefore, a question arises as to the degree to which group differences in cardiac sympathetic tone and blood pressure reactivity were driven by controls expressing signs of autonomic dysregulation in prehypertension. The blood pressure of seven (17.5%) controls fell in the prehypertensive range at the beginning of the experimental session; only three (7.5%) of them fulfilled these criteria at both the screening and experimental sessions (Chobanian et al., 2003). Considering these relatively low numbers, one may conclude that the impact of prehypertension on the group effects was rather limited and that they arose mainly due to autonomic differences between hypotensive and normotensive participants. Replicating a former finding (Duschek, Dietel, et al., 2008), BRS was markedly higher in the present hypotensive sample. In addition to the aforementioned blood pressure reduction due to overcompensation of phasic blood pressure increases (in the case of enhanced BRS), the possibility that the operating point of the baroreflex is reset to lower blood pressure values has been considered to contribute to hypotension development (Duschek, Dietel, et al., 2008). The relevance of the baroreflex system to blood pressure over the long term is emphasized by the wellknown reflex inhibition that occurs in essential hypertension (Carthy, 2013), as well as the inverse relationship between BRS and tonic blood pressure in healthy normotensive individuals (Duschek & Reyes del Paso, 2007; Hesse, Charkoudian, Liu, Joyner, & Eisenach, 2007). In a follow-up study, low BRS predicted blood pressure increases over a 5-year period (Ducher, Fauvel, & Cerrutti, 2006). Finally, experimental observations in animals indicated that prolonged stimulation of the baroreceptor afferents leads to sustained blood pressure decline (Lohmeier et al., 2004). In addition to its validity as a psychophysiological indicator of health and disease, BRS has a proven impact on various features of behavior regulation (Duschek, Werner, et al., 2013). In the present context, for example, the repeatedly documented inverse association between BRS and cognitive performance may be of interest (Duschek, Muckenthaler, & Reyes del Paso, 2009; Reyes del Paso, González, Hernández, Duschek, & Gutiérrez, 2009; Yasumasu, Reyes del Paso, Takahara, & Nakashima, 2006). This connection has been discussed in the context of the well-established central nervous inhibition that arises from baroreceptor afferents, where more pronounced Journal of Psychophysiology (2019), 33(1), 39–53

S. Duschek et al., Autonomic control in chronic hypotension

inhibition seems to occur in the case of higher BRS (Duschek, Werner, et al., 2013). This mechanism may also play a role in hypotension-related cognitive impairment (Duschek et al., 2005). Similarly, interindividual differences in cortical activation processes, as assessed by evoked EEG potentials, correlated negatively with BRS (Duschek, Wörsching, & Reyes del Paso, 2013). Baroreceptormediated inhibition may also account for this association; as such, the enhanced BRS in hypotension may be involved in the reduced cortical activation observed in this population (Duschek et al., 2006; Weisz, Schandry, Jacobs, Mialet, & Duschek, 2002). In contrast to BRS, RMSSD did not differ between the two study groups. Although the RMSSD represents overall HRV, it is most closely associated with HRV in the frequency range of respiration, and constitutes a wellestablished index of parasympathetic influences on heart rate (Berntson et al., 2016). Previous analyses of HRV in chronic hypotension yielded inconsistent results. While ECG recordings during sleep revealed higher values of RMSSD and pNN50, indicating enhanced cardiac vagal outflow to the sinus node (Covassin et al., 2012), no effects were found for nocturnal high frequency HRV (de Zambotti, Covassin, Cellini, Sarlo, Torre, et al., 2012). The present observations, of no group difference in high frequency HRV but increased BRS in hypotension, may seem counterintuitive, as these parameters are closely linked physiologically (Reyes del Paso, Hernández, & González, 2004; Reyes del Paso, Langewitz, Robles, & Perez, 1996). BRS and RSA commonly show substantial positive correlations, increasing and decreasing in parallel under conditions of mental challenge, relaxation, or pharmacological treatment (Duschek, Heiss, et al., 2009; Reyes del Paso et al., 1996, 2004). The baroreflex is believed to be one of the most important sources of HRV in the frequency range of respiration (Berntson et al., 2016; Reyes del Paso et al., 2013; van Roon, Mulder, Althaus, & Mulder, 2004). Given its rapid response characteristics, the reflex is able to phasically alter cardiac vagal outflow even at the highest frequencies. Baroreflex activity is modulated by respiratory influences of both central nervous and peripheral origin, which trigger cyclical fluctuations in the discharge rates of vagal motoneurons (Berntson et al., 1997). Despite these causal influences of the baroreflex on HRV, dissociations between BRS and high frequency HRV have been repeatedly reported. For example, it was observed that BRS and high frequency HRV correlated in opposite directions with EEG measures of cortical tone and cortical excitability (Duschek, Wörsching, et al., 2013; Duschek, Wörsching, & Reyes del Paso, 2015). Furthermore, specific associations of both parameters with cognitive performance and emotional states have been demonstrated (Delgado, Vila, & Reyes del Paso, 2014; Ó 2017 Hogrefe Publishing

S. Duschek et al., Autonomic control in chronic hypotension

Duschek, Muckenthaler, & Reyes del Paso, 2009; Yasumasu et al., 2006). It is therefore evident that high frequency HRV does not exclusively reflect the properties of the baroreflex; BRS and HRV affect psychological and physiological processes via different pathways (Duschek, Wörsching, et al., 2013; Duschek et al., 2015). Accordingly, both parameters at least partly represent different mechanisms of cardiac regulation, where baroreflex mechanisms seem to play a more prominent role in the genesis of hypotension than processes represented by HRV. In the present study various changes in cardiovascular parameters during mental stress were observed. The higher CO and lower PEP during stress versus baseline denote stress-induced sympathetic activation (Berntson et al., 2016). RMSSD did not differ between baseline and stress conditions. Though one would expect vagal withdrawal during mental effort, a similar finding was reported by de Zambotti, Covassin, Cellini, Sarlo, and Stegagno (2012). In hypotensive and normotensive women, HRV in the high frequency range did not significantly change between rest and cognitive stress induced by a 2-back bask. High frequency HRV is a well-established index of vagal influences on heart rate that correlates highly with the RMSSD (Berntson et al., 2005, 2016). While de Zambotti, Covassin, Cellini, Sarlo, and Stegagno (2012) also observed stress-induced heart rate modulation comparable to the present study, PEP was largely unaffected. BRS was lower during the stress than resting conditions; this is commonly seen under conditions of mental activity and other laboratory stressors (Brooks, Fox, Lopez, & Sleight, 1978; Duschek & Reyes del Paso, 2007; Robbe et al., 1987). BRS attrition implies a reduced buffering effect of the baroreflex, which facilitates stress-induced heart rate and blood pressure increases, and by extension enhancement of the metabolic supply of the organism, thereby also improving conditions for neural and cognitive processing (Duschek, Werner, et al., 2013). Interestingly, in both study groups, SV was lower during stress than at rest. This may be explained by the fact that heart rate acceleration is associated with a reduction in ventricular preload, which, due to the FrankStarling mechanism, leads to lower SV (Kenny et al., 1987; Levy & Pappano, 2007). The extent of the increase in systolic and diastolic blood pressure during cognitive challenge was smaller in hypotensive versus control subjects. This is consistent with previous reports of blunted cardiovascular reactivity in chronic hypotension, which has been interpreted in terms of insufficient hemodynamic adjustment to current requirements (Covassin et al., 2013; Duschek, Dietel, et al., 2008; Duschek et al., 2005). Reduced blood pressure reactivity may also be involved in deficient cerebral blood flow regulation in hypotension. Various studies using transcranial Doppler sonography have documented blunted hemodynamic Ó 2017 Hogrefe Publishing

adjustment, in terms of smaller phasic blood flow increases during cognitive and emotional processing (Cellini et al., 2015; Duschek & Schandry, 2004; Sarlo, de Zambotti, Gallicchio, Devigili, & Stegagno, 2013). Moreover, recent research indicated that the amplitudes of cerebral blood flow responses depend, to a certain degree, on the magnitude of blood pressure reactions (Duschek, Heiss, Schmidt, Werner, & Schuepbach, 2010; Duschek & Schandry, 2006; Duschek, Werner, Kapan, & Reyes del Paso, 2008). This is relevant insofar as it has repeatedly been shown that performance related to attentional and executive functions varies according to the magnitude of cerebral blood flow responses, where insufficient blood flow adjustment plays an important role in hypotension-related cognitive impairment (Duschek, Hadjamu, & Schandry, 2007a; Duschek & Schandry, 2004; Duschek, Schuepbach, & Schandry, 2008; Schuepbach, Boeker, Duschek, & Hell, 2007). The analysis of the time courses of the cardiovascular indices during mental stress revealed various patterns of changes. Heart rate and CO steeply increased during the initial phase of the arithmetic task, and thereafter decreased before reaching a plateau slightly above baseline. In contrast, RMSSD markedly decreased during the initial phase, and gradually rose during the remaining task. PEP also showed a strong initial decline but remained at a low level until the end of the period. Other variables displayed slower changes: BRS progressively decreased during the entire task, whereas TPR – as well as systolic and diastolic blood pressure – gradually increased. The smaller slope of the blood pressure rise in the hypotensive versus normotensive sample confirms the notion of blunted blood pressure reactivity in this condition (Duschek et al., 2005). The different time dynamics reflect contributions of different physiological mechanisms to the cardiovascular modulations. The rapid changes immediately after task onset observed in heart rate, SV, and PEP may have been evoked by fastacting regulatory mechanisms, especially the autonomic nervous system. Specifically, such changes may occur due to alpha- and beta-sympathetic activation, as well as vagal withdrawal in the case of heart rate. Progressive reduction of vagal inhibition after the initial task period may have contributed to heart rate recovery. This is supported by the modulation of RMSSD, characterized by a steep initial decline and progressive increase across the remaining period. The gradual increases in TPR and blood pressure during the task reflect the involvement of slower mechanisms, such as activation of adrenal medullary hormones (Berntson et al., 2016). Reduction in BRS during mental challenge is commonly ascribed to top-down modulation of cardiovascular brain stem centers by cortical units (Duschek, Werner, & Reyes del Paso, 2013; van Roon et al., 2004). Progressive BRS decrease implies a reduction in the compensatory effect of the baroreflex, which allows Journal of Psychophysiology (2019), 33(1), 39–53

blood pressure to increase. This mechanism is illustrated by the inverse time courses of BRS and blood pressure across the task period. Some methodological limitations of the study warrant discussion. The WHO (1978) criteria, which were applied during recruitment of hypotensive individuals, that is, systolic blood pressure below 100 mmHg in women and 110 mmHg in men, are relatively lenient (Covassin et al., 2012, 2013). More recently proposed definitions include lower systolic cut-off values and also take into account the diastolic value. For example, the US National Heart, Lung, and Blood Institute (NHLBI) suggested 90 and 60 mmHg as systolic and diastolic limits (NHLBI, 2017). The application of more strict criteria would potentially reveal a different pattern of autonomic dysregulation, and the presently investigated sample is not representative of hypotensives defined according to stricter criteria, such as those of the NHLBI (2017). Another limitation pertains to the lack of assessment of subjective hypotensive complaints; therefore, it was not possible to establish the impact of autonomic alterations on symptom severity. Regarding cardiovascular recordings, it must be acknowledged that blood volume was not measured, which impeded the investigation of its possible role in mediating group differences in hemodynamic parameters. Furthermore, respiration rate was not controlled for during computation of the RMSSD. Even though recent research did not support the utility of respiration control in HRV assessment (Denver, Reed, & Porges, 2007), possible distortion of the findings, due to respiratory changes occurring during the verbal response required in the serial subtraction task, cannot be ruled out. In short, the present study identified reduced sympathetic myocardial drive as the main characteristic of the autonomic impairment associated with chronic hypotension, where diminished beta-adrenergic inotropic influence, rather than aberrant alpha-adrenergic vasomotor control, may play a dominant role. While the study groups did not differ in HRV, involvement of the parasympathetic system in this condition is suggested by increased baroreflex function. Finally, the possible implications for the pharmacological treatment of chronic hypotension should be considered; the data indicate greater suitability of beta- versus alpha sympathomimetic agents. This is supported by pilot data on the acute effects of different sympathomimetics in chronic hypotension. While beta-agonistic treatment was followed by heart rate and blood pressure elevation (Duschek, Hadjamu, & Schandry, 2007b), acute application of an alpha-agonist led to extreme baroreflex-mediated heart rate deceleration, which in turn impeded enhancement of CO (Duschek, Heiss, et al., 2009). However, as chronic hypotension does not constitute a serious medical condition, the necessity for sympathomimetic treatment is certainly restricted to a very limited number of cases. Journal of Psychophysiology (2019), 33(1), 39–53

S. Duschek et al., Autonomic control in chronic hypotension

Furthermore, empirical knowledge on potential benefits for subjective complaints, as well as possible undesirable effects, remains scant (Duschek & Schandry, 2007; Duschek, Heiss, et al., 2009); as such, further research is clearly warranted. Acknowledgments The study was supported by the Anniversary Fund of the Austrian National Bank (project 16289). We are grateful to Maximilian Stefani and Angela Bair for their help with data analysis. Ethics and Disclosure Statements All participants of the study provided written informed consent prior to the study and the study was approved by the Board for Ethical Questions in Science of the University of Innsbruck (Austria). All authors disclose no actual or potential conflicts of interest including any financial, personal, or other relationships with other people or organizations that could inappropriately influence (bias) their work.

References Berntson, G. G., Bigger, J. T. Jr., Eckberg, D. L., Grossman, P., Kaufmann, P. G., Malik, M., . . . van der Molen, M. W. (1997). Heart rate variability: Origins, methods, and interpretive caveats. Psychophysiology, 34, 623–648. Berntson, G. G., Lozano, D. L., & Chen, Y. J. (2005). Filter properties of root mean square successive difference (RMSSD) for heart rate. Psychophysiology, 42, 246–252. https://doi.org/ 10.1111/j.1469-8986.2005.00277.x Berntson, G. G., Quigley, K. S., Norman, G. J., & Lozano, D. (2016). Cardiovascular psychophysiology. In J. T. Cacioppo, L. G. Tassinary, & G. G. Berntson (Eds.), Handbook of psychophysiology (pp. 183–215). Cambridge, UK: Cambridge University Press. Bertinieri, G., di Rienzo, M., Cavallazzi, A., Ferrari, A. U., Pedotti, A., & Mancia, G. (1985). A new approach to analysis of the arterial baroreflex. Journal of Hypertension, 3, S79–S81. Bogert, L. W. J., & Lieshout, J. J. (2005). Non-invasive arterial pressure and stroke volume changes from the human finger. Experimental Physiology, 90, 437–446. https://doi.org/ 10.1113/expphysiol.2005.030262 Brooks, D., Fox, P., Lopez, R., & Sleight, P. (1978). The effect of mental arithmetic on blood pressure variability and baroreflex sensitivity in man. Journal of Physiology, 280, 75–76. Cacioppo, J. T., Berntson, G. G., Binkley, P. F., Quigley, K. S., Uchino, B. N., & Fieldstone, A. (1994). Autonomic cardiac control. II. Noninvasive indices and basal response as revealed by autonomic blockades. Psychophysiology, 31, 586–598. https://doi.org/10.1111/j.1469-8986.1994.tb02351.x Cain, A. E., & Khalil, R. A. (2002). Pathophysiology of essential hypertension: Role of the pump, the vessel, and the kidney. Seminars in Nephrology, 22, 3–16. Carthy, E. R. (2013). Autonomic dysfunction in essential hypertension: A systematic review. Annals of Medicine & Surgery, 3, 2–7. https://doi.org/10.1016/j.amsu.2013.11.002 Cellini, N., de Zambotti, M., Covassin, N., Gallicchio, G., Stegagno, L., & Sarlo, M. (2015). Reduced cerebral and cardiovascular

Ó 2017 Hogrefe Publishing

S. Duschek et al., Autonomic control in chronic hypotension

hemodynamics during sustained affective stimulation in young women with chronic low blood pressure. Physiology & Behavior, 143, 83–89. https://doi.org/10.1016/j.physbeh.2015. 02.047 Chobanian, A. V., Bakris, G. L., Black, H. R., Cushman, W. C., Green, L. A., Izzo, J. L., . . . Roccella, E. J. (2003). Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension, 42, 1206–1252. https://doi.org/10.1161/01.HYP.0000107251. 49515.c2 Covassin, N., de Zambotti, M., Cellini, N., Sarlo, M., & Stegagno, L. (2012). Nocturnal cardiovascular activity in essential hypotension: Evidence of differential autonomic regulation. Psychosomatic Medicine, 74, 952–960. https://doi.org/10.1097/PSY. 0b013e318272db69 Covassin, N., de Zambotti, M., Cellini, N., Sarlo, M., & Stegagno, L. (2013). Cardiovascular down-regulation in essential hypotension: Relationships with autonomic control and sleep. Psychophysiology, 50, 767–776. https://doi.org/10.1111/ psyp.12055 Davis, J. T., Rao, F., Naqshbandi, D., Fung, M. M., Zhang, K., Schork, A. J., . . . O’Connor, D. T. (2012). Autonomic and hemodynamic origins of pre-hypertension: Central role of heredity. Journal of the American College of Cardiology, 59, 2206–2216. https://doi.org/10.1016/j.jacc.2012.02.040 De Buyzere, M., Clement, D. L., & Duprez, D. (1998). Chronic low blood pressure: A review. Cardiovascular Drugs and Therapy, 12, 29–35. https://doi.org/10.1023/A:1007729229483 Delgado, L. C., Vila, J., & Reyes del Paso, G. A. (2014). Proneness to worry is negatively associated with blood pressure and baroreflex sensitivity: Further evidence of the blood pressure emotional dampening hypothesis. Biological Psychology, 96, 20–27. https://doi.org/10.1016/j.biopsycho.2013.11.005 Denver, J. W., Reed, S. F., & Porges, S. W. (2007). Methodological issues in the quantification of respiratory sinus arrhythmia. Biological Psychology, 74, 286–294. https://doi.org/10.1016/ j.biopsycho.2005.09.005 de Zambotti, M., Covassin, N., Cellini, N., Sarlo, M., & Stegagno, L. (2012). Cardiac autonomic profile during rest and working memory load in essential hypotensive women. International Journal of Psychophysiology, 85, 200–205. https://doi.org/ 10.1016/j.ijpsycho.2012.05.002 de Zambotti, M., Covassin, N., Cellini, N., Sarlo, M., Torre, J., & Stegagno, L. (2012). Hemodynamic and autonomic modifications during sleep stages in young hypotensive women. Biological Psychology, 91, 22–27. https://doi.org/10.1016/ j.biopsycho.2012.05.009 Drukteinis, J. S., Roman, M. J., Fabsitz, R. R., Lee, E. T., Best, L. G., Russell, M., & Devereux, R. B. (2007). Cardiac and systemic hemodynamic characteristics of hypertension and prehypertension in adolescents and young adults: The Strong Heart Study. Circulation, 115, 221–227. https://doi.org/ 10.1161/CIRCULATIONAHA.106.668921 Ducher, M., Fauvel, J. P., & Cerrutti, C. (2006). Risk profile in hypertension genesis: A five-year follow-up study. American Journal of Hypertension, 19, 775–781. https://doi.org/10.1016/ j.amjhyper.2005.07.019 Duschek, S., Dietel, A., Schandry, R., & Reyes del Paso, G. A. (2008). Increased baroreflex sensitivity and reduced cardiovascular reactivity in chronic low blood pressure. Hypertension Research, 31, 1873–1878. https://doi.org/10.1291/hypres.31.1873 Duschek, S., Hadjamu, M., & Schandry, R. (2007a). Enhancement of cerebral blood flow and cognitive performance due to pharmacological blood pressure elevation in chronic hypotension. Psychophysiology, 44, 145–153. https://doi.org/10.1111/ j.1469-8986.2006.00472.x

Ó 2017 Hogrefe Publishing

Duschek, S., Hadjamu, M., & Schandry, R. (2007b). Dissociation between cortical activation and cognitive performance in the pharmacological treatment of chronic hypotension. Biological Psychology, 75, 277–285. https://doi.org/10.1016/j.biopsycho. 2007.03.007 Duschek, S., Heiss, H., Buechner, B., Werner, N. S., Schandry, R., & Reyes del Paso, G. A. (2009). Hemodynamic determinants of chronic hypotension and their modification through vasopressor application. Journal of Physiological Sciences, 59, 105–112. https://doi.org/10.1007/s12576-008-0015-5 Duschek, S., Heiss, H., Schmidt, F. H., Werner, N., & Schuepbach, D. (2010). Interactions between systemic hemodynamics and cerebral blood flow during attentional processing. Psychophysiology, 47, 1159–1166. https://doi.org/10.1111/j.1469-8986. 2010.01020.x Duschek, S., Hoffmann, A., & Reyes del Paso, G. A. (2017). Affective impairment in chronically low blood pressure. Journal of Psychosomatic Research, 93, 33–40. https://doi.org/ 10.1016/j.jpsychores.2016.12.008 Duschek, S., Hoffmann, A., Reyes del Paso, G. A., & Ettinger, U. (2017). Autonomic cardiovascular control and executive function in chronically low blood pressure. Annals of Behavioral Medicine, 51, 442–453. Duschek, S., Matthias, E., & Schandry, R. (2005). Essential hypotension is accompanied by deficits in attention and working memory. Behavioral Medicine, 30, 149–158. https:// doi.org/10.3200/BMED.30.4.149-160 Duschek, S., Meinhardt, J., & Schandry, R. (2006). Reduced cortical activity due to chronic low blood pressure: An EEG study. Biological Psychology, 72, 241–250. https://doi.org/ 10.1016/j.biopsycho.2005.06.011 Duschek, S., Muckenthaler, M., & Reyes del Paso, G. A. (2009). Relationships between features of cardiovascular control and cognitive performance. Biological Psychology, 81, 110–117. https://doi.org/10.1016/j.biopsycho.2009.03.003 Duschek, S., & Reyes del Paso, G. A. (2007). Quantification of cardiac baroreflex function at rest and during autonomic stimulation. Journal of Physiological Sciences, 57, 259–268. https://doi.org/10.2170/physiolsci.RP008807 Duschek, S., & Schandry, R. (2004). Cognitive performance and cerebral blood flow in essential hypotension. Psychophysiology, 41, 905–913. https://doi.org/10.1111/j.1469-8986.2004. 00249.x Duschek, S., & Schandry, R. (2006). Deficient adjustment of cerebral blood flow to cognitive activity due to chronically low blood pressure. Biological Psychology, 72, 311–317. https://doi. org/10.1016/j.biopsycho.2005.12.003 Duschek, S., & Schandry, R. (2007). Reduced brain perfusion and cognitive performance due to essential hypotension. Clinical Autonomic Research, 17, 69–76. https://doi.org/10.1007/ s10286-006-0379-7 Duschek, S., Schuepbach, D., & Schandry, R. (2008). Time-locked association between rapid cerebral blood flow modulation and attentional performance. Clinical Neurophysiology, 119, 1292–1299. https://doi.org/10.1016/j.clinph.2008.01.102 Duschek, S., Weisz, N., & Schandry, R. (2003). Reduced cognitive performance and prolonged reaction time accompany moderate hypotension. Clinical Autonomic Research, 13, 427–432. https://doi.org/10.1007/s10286-003-0124-4 Duschek, S., Werner, N., Kapan, N., & Reyes del Paso, G. A. (2008). Patterns of cerebral blood flow and systemic hemodynamics during arithmetic processing. Journal of Psychophysiology, 22, 81–90. https://doi.org/10.1027/0269-8803.22.2.81 Duschek, S., Werner, N. S., & Reyes del Paso, G. A. (2013). The behavioral impact of baroreflex function: A review. Psychophysiology, 50, 1183–1193. https://doi.org/10.1111/psyp.12136

Journal of Psychophysiology (2019), 33(1), 39–53

Duschek, S., Wörsching, J., & Reyes del Paso, G. A. (2013). Interactions between autonomic cardiovascular regulation and cortical activity: a CNV study. Psychophysiology, 50, 388–397. https://doi.org/10.1111/psyp.12026 Duschek, S., Wörsching, J., & Reyes del Paso, G. A. (2015). Autonomic cardiovascular regulation and cortical tone. Clinical Physiology and Functional Imaging, 35, 383–392. https://doi. org/10.1111/cpf.12174 Freeman, R., Wieling, W., Axelrod, F. B., Benditt, D. G., Benarroch, E., Biaggioni, I., . . . van Dijk, J. G. (2011). Consensus statement on the definition of orthostatic hypotension, neurally mediated syncope and the postural tachycardia syndrome. Clinical Autonomic Research, 21, 69–72. Hassan, S., & Turner, P. (1983). Systolic time intervals: A review of the method in the non-invasive investigation of cardiac function in health, disease and clinical pharmacology. Postgraduate Medical Journal, 59, 423–434. Heijer, T., Skoog, I., Oudkerk, M., de Leeuw, F. E., de Groot, J. C., Hofman, A., & Breteler, M. M. (2003). Association between blood pressure levels over time and brain atrophy in the elderly. Neurobiology of Aging, 24, 307–313. https://doi.org/10.1016/ S0197-4580(02)00088-X Hesse, C., Charkoudian, N., Liu, Z., Joyner, M. J., & Eisenach, J. H. (2007). Baroreflex sensitivity inversely correlates with ambulatory blood pressure in healthy normotensive humans. Hypertension, 50, 41–46. https://doi.org/10.1161/HYPERTENSIONAHA.107. 090308 Jochemsen, H. M., Muller, M., Visseren, F. L., Scheltens, P., Vincken, K. L., Mali, W. P., . . . Geerlings, M. I. (2013). Blood pressure and progression of brain atrophy: The SMART-MR Study. JAMA Neurology, 70, 1046–1053. https://doi.org/ 10.1001/jamaneurol.2013.217 Kenny, J., Plappert, T., Doubilet, P., Salzman, D., & Sutton, M. (1987). Effects of heart rate on ventricular size, stroke volume, and output in the normal human fetus: A prospective Doppler echocardiographic study. Circulation, 76, 52–58. https://doi. org/10.1161/01.CIR.76.1.52 Levy, M. N., & Pappano, A. J. (2007). Cardiovascular physiology. Philadelphia, PA: Mosby Elsevier. Lohmeier, T. E., Irwin, E. D., Rossing, M. A., Serdar, D. J., & Kieval, R. S. (2004). Prolonged activation of the baroreflex produces sustained hypotension. Hypertension, 43, 306–311. https://doi. org/10.1161/01.HYP.0000111837.73693.9b Lovallo, W. R., & al’Absi, M. (1998). Hemodynamics during rest and behavioral stress in normotensive men at high risk for hypertension. Psychophysiology, 35, 47–53. Lund-Johansen, P. (1991). Twenty-year follow-up of hemodynamics in essential hypertension during rest and exercise. Hypertension, 18, SIII54–SIII61. Moshkovitz, Y., Kaluski, E., Milo, O., Vered, Z., & Cotter, G. (2004). Recent developments in cardiac output determination by bioimpedance: Comparison with invasive cardiac output and potential cardiovascular applications. Current Opinion in Cardiology, 19, 229–237. https://doi.org/10.1097/00001573200405000-00008 Muller, M., Sigurdsson, S., Kjartansson, O., Aspelund, T., Lopez, O. L., Jonnson, P. V., . . . Launer, L. J. (2014). Joint effect of midand late-life blood pressure on the brain: The AGES-Reykjavik study. Neurology, 82, 2187–2195. https://doi.org/10.1212/ WNL.0000000000000517 National Heart, Lung, and Blood Institute. (2017). What is hypotension? Retrieved from https://www.nhlbi.nih.gov/ health/health-topics/topics/hyp Pemberton, J. (1989). Does constitutional hypotension exist? British Medical Journal, 298, 660–662.

Journal of Psychophysiology (2019), 33(1), 39–53

S. Duschek et al., Autonomic control in chronic hypotension

Perakakis, P., Joffily, M., Taylor, M., Guerra, P., & Vila, J. (2010). KARDIA: A Matlab software for the analysis of cardiac interbeat intervals. Computer Methods and Programs in Biomedicine, 98, 83–89. https://doi.org/10.1016/j.cmpb.2009.10.002 Pilgrim, J. A. (1994). Psychological aspects of high and low blood pressure. Psychological Medicine, 24, 9–14. https://doi.org/ 10.1017/S0033291700026787 Prospective Studies Collaboration. (2002). Age-specific relevance of usual blood pressure to vascular mortality: A meta-analysis of individual data for one million adults in 61 prospective studies. The Lancet, 360, 1903–1913. https://doi.org/10.1016/ S0140-6736(02)11911-8 Qiu, C., Winblad, B., & Fratiglioni, L. (2005). The age-dependent relation of blood pressure to cognitive function and dementia. Lancet Neurology, 4, 487–499. https://doi.org/10.1016/S14744422(05)70141-1 Qureshi, A. I., Suri, M. F., Kirmani, J. F., & Divani, A. A. (2005). Prevalence and trends of pre-hypertension and hypertension in United States: National Health and Nutrition Examination Surveys 1976 to 2000. Medical Science Monitor, 11, CR403–CR409. Raaijmakers, E., Faes, T. J., Scholten, R. J., Goovaerts, H. G., & Heethaar, R. M. (1999). A meta-analysis of published studies concerning the validity of thoracic impedance cardiography. Annals of the New York Academy of Sciences, 873, 121–127. https://doi.org/10.1111/j.1749-6632.1999.tb09458.x Reyes del Paso, G. A. (1994). A program to assess baroreceptor cardiac reflex function. Behaviour Research Methods Instruments & Computers, 26, 62–64. https://doi.org/10.3758/ BF03204567 Reyes del Paso, G. A., González, M. I., & Hernández, J. A. (2010). Comparison of baroreceptor cardiac reflex sensitivity estimates from inter-systolic and ECG R-R intervals. Psychophysiology, 47, 1102–1108. https://doi.org/10.1111/j.1469-8986. 2010.01018.x Reyes del Paso, G. A., González, M. I., Hernández, J. A., Duschek, S., & Gutiérrez, N. (2009). Tonic blood pressure modulated the relationship between baroreceptor cardiac reflex sensitivity and cognitive performance. Psychophysiology, 46, 932–938. https://doi.org/10.1111/j.1469-8986.2009.00832.x Reyes del Paso, G. A., Hernández, J. A., & González, M. I. (2004). Differential analysis in the time domain of the baroreceptor cardiac reflex sensitivity as a function of sequence length. Psychophysiology, 41, 483–488. https://doi.org/10.1111/ j.1469-8986.2004.00178.x Reyes del Paso, G. A., Langewitz, W., Robles, H., & Perez, N. (1996). A between-subjects comparison of respiratory sinus arrythmia and baroreceptor cardiac reflex sensitivity as noninvasive measures of tonic parasympathetic cardiac control. International Journal of Psychophysiology, 22, 163–171. https:// doi.org/10.1016/0167-8760(96)00020-7 Reyes del Paso, G. A., Langewitz, W., Mulder, L. J. M., van Roon, A., & Duschek, S. (2013). The utility of low frequency heart rate variability as an index of sympathetic cardiac tone: A review with emphasis on a re-analysis of previous studies. Psychophysiology, 50, 477–487. https://doi.org/10.1111/psyp. 12027 Robbe, H., Mulder, L., Ruddel, H., Langewitz, W., Veldman, J., & Mulder, G. (1987). Assessment of BRS by means of spectral analysis. Hypertension, 10, 538–543. Rosengren, A., Tibblin, G., & Wilhelmsen, L. (1993). Low systolic blood pressure and self-perceived wellbeing in middle aged men. British Medical Journal, 306, 243–246. https://doi.org/ 10.1136/bmj.306.6872.243 Sarlo, M., de Zambotti, M., Gallicchio, G., Devigili, A., & Stegagno, L. (2013). Impaired cerebral and systemic hemodynamics under

Ó 2017 Hogrefe Publishing

S. Duschek et al., Autonomic control in chronic hypotension

cognitive load in young hypotensives: A transcranial Doppler study. Journal of Behavioral Medicine, 36, 134–142. https://doi. org/10.1007/s10865-012-9410-8 Schuepbach, D., Boeker, H., Duschek, S., & Hell, D. (2007). Rapid cerebral hemodynamic modulation during mental planning and movement execution: Evidence of time-locked relationship with complex behavior. Clinical Neurophysiology, 118, 2254–2262. https://doi.org/10.1016/j.clinph.2007.07.013 Steptoe, A., & Vögele, C. (1990). Cardiac baroreceptor reflex function during postural change assessed using noninvasive spontaneous sequence analysis in young men. Cardiovascular Research, 24, 627–632. https://doi.org/ 10.1093/cvr/24.8.627 van Roon, A. M., Mulder, L. J. M., Althaus, M., & Mulder, G. (2004). Introducing a baroreflex model for studying cardiovascular effects of mental workload. Psychophysiology, 41, 961–981. https://doi.org/10.1111/j.0048-5772.2004.00251.x Vasan, R. S., Larson, M. G., Leip, E. P., Kannel, W. B., & Levy, D. (2001). Assessment of frequency of progression to hypertension in non-hypertensive participants in the Framingham Heart Study: A cohort study. The Lancet, 358, 1682–1686. https://doi. org/10.1016/S0140-6736(01)06710-1 Weisz, N., Schandry, R., Jacobs, A., Mialet, J., & Duschek, S. (2002). Early contingent negative variation of the EEG and attentional flexibility are reduced in hypotension. International Journal of Psychophysiology, 45, 253–260. https://doi.org/ 10.1016/S0167-8760(02)00032-6 Wessely, S., Nickson, J., & Cox, B. (1990). Symptoms of low blood pressure: A population study. British Medical Journal, 301, 362–365. https://doi.org/10.1136/bmj.301.6748.362

Ó 2017 Hogrefe Publishing

Wesseling, K. H., Jansen, J. R., Settels, J. J., & Schreuder, J. J. (1993). Computation of aortic flow from pressure in humans using a nonlinear, three-element model. Journal of Applied Physiology, 74, 2566–2573. Wesseling, K. H., De Wit, B., Van der Hoeven, G. M. A., Van Goudoever, J., & Settels, J. J. (1995). Physiocal, calibrating finger vascular physiology for Finapres. Homeostasis, 36, 67–82. WHO. (1978). Arterial hypertension. In Technical Report Series No. 628. Geneva, Switzerland: World Health Organisation. Yasumasu, T., Reyes del Paso, G. A., Takahara, K., & Nakashima, Y. (2006). Reduced baroreflex cardiac sensitivity predicts increased cognitive performance. Psychophysiology, 43, 41–45. https://doi.org/10.1111/j.1469-8986.2006.00377.x Received November 29, 2016 Revision received April 13, 2017 Accepted April 13, 2017 Published online November 21, 2017

Stefan Duschek UMIT – University for Health Sciences, Medical Informatics and Technology Institute of Psychology Eduard-Wallnöfer-Zentrum 1 6060 Hall in Tirol Austria stefan.duschek@umit.at

Journal of Psychophysiology (2019), 33(1), 39–53

Article

Neural Correlates of Empathy for Physical and Psychological Pain Vera Flasbeck and Martin Brüne LWL University Hospital, Department of Psychiatry, Psychotherapy and Preventive Medicine, Division of Cognitive Neuropsychiatry and Psychiatric Preventive Medicine, Ruhr-University Bochum, Germany

Abstract: Empathy is known as the ability to share and understand someone else’s feelings. Previous research has either addressed the neural correlates of empathy for pain or social exclusion, but no study has examined empathy for physical and psychological (social) pain simultaneously. Forty-seven participants completed our novel “Social Interaction Empathy Task” during electroencephalogram (EEG) recording. Participants had to observe and rate the intensity of physical and psychological pain in social interactions from a first- and thirdperson perspective. At the behavioral level, subjects did not differentiate between the perspectives and rated physically painful scenarios as more painful than psychologically painful and neutral interactions. Psychologically painful pictures were also rated as more painful than neutral pictures. Analysis of event-related potentials (ERPs) revealed an early and a late response with a higher ERP response to physical and psychological pain compared to neutral interactions. Moreover, a significant difference emerged between the two dimensions of painful interactions. Furthermore, we found that the activity over frontal regions for discrimination of painful interactions was lateralized to the right hemisphere. Moreover, we detected significant correlations with the self-rated perspective taking ability. This suggests the psychological and physical pain qualities are processed differently but both are related to empathic traits. We further suggest that the right hemisphere may be specifically involved in the processing of empathy-related tasks. Keywords: empathy, ERP, physical pain, social exclusion, perspective taking

Empathy is a complex mental state which enables us to share and to understand someone else’s emotions (Gonzalez-Liencres, Shamay-Tsoory, & Brüne, 2013; Shamay-Tsoory, 2011; Shamay-Tsoory, Aharon-Peretz, & Perry, 2009). It requires an affective component to match the emotional state of someone else and a cognitive component involving the capacity to differentiate self from others (for reviews, see Gonzalez-Liencres et al., 2013; Singer, 2006). Empathy for another’s pain has become a common tool for the investigation of empathic abilities in human beings. Neuroimaging research has revealed overlapping brain areas that are activated during empathy for another’s pain and the first-hand experience of pain, whereby these affective parts of the “pain matrix” comprise the bilateral anterior insular cortex and medial/ anterior cingulate cortex (ACC; Botvinick et al., 2005; Jackson, Brunet, Meltzoff, & Decety, 2006; Lamm, Decety, & Singer, 2011; Singer et al., 2004). Moreover, the activity of the affective brain regions correlates with the judgment of pain intensity and with self-rated empathic abilities (Jackson, Meltzoff, & Decety, 2005; Saarela et al., 2007). Several studies have tried to differentiate between a firstperson perspective and a third-person perspective during the empathy task. That is, individuals were asked to Journal of Psychophysiology (2019), 33(1), 54–63 https://doi.org/10.1027/0269-8803/a000205

indicate how they would feel in the observed situation (i.e., first-person perspective, FPP) and how the observed character would feel (i.e., third-person perspective, TPP). While some studies reported differences between the perspectives with a higher pain rating and faster reactions in the FPP (Li & Han, 2010; van der Heiden, Scherpiet, Konicar, Birbaumer, & Veit, 2013), other studies did not find any difference. For example, Jackson et al. (2006) reported that pain empathy evoked overlapping brain activity in both perspectives, but also a higher involvement of the secondary somatosensory cortex, ACC, and insula in the FPP and higher recruitment of the right temporoparietal junction in the TPP condition (Jackson et al., 2006). In addition, Abu-Akel, Palgi, Klein, Decety, and Shamay-Tsoory (2015) examined empathy for pain in both FPP and TPP under oxytocin treatment and could not detect differences between the perspectives in the placebo group. However, after administration of oxytocin participants rated the pain as more intense in the TPP compared to the FPP suggesting a modulatory effect of oxytocin on empathy (Abu-Akel et al., 2015). Another approach to investigate empathy for pain and its temporal dynamics entailed the analysis of event-related brain potentials (ERPs). Fan and Han were the first to Ó 2017 Hogrefe Publishing

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

examine empathy for pain by presenting images of hands in neutral or physically painful situations. The participants’ attention was required for some of the task conditions and in some conditions withdrawn from the pain cues (Fan & Han, 2008). Painful stimuli elicited a positive shift of the ERP compared to neutral stimuli at short-latencies over frontal and central areas and at long-latencies over central-parietal regions. Moreover, the empathic responses could be modulated by attention and stimulus reality (cartoon or pictures) of the pain cues. The ERP pattern and the positive shifts of ERP components for painful stimuli relative to neutral stimuli could be replicated in further studies (Decety, Yang, & Cheng, 2010; Han, Fan, & Mao, 2008; Li & Han, 2010). Another dimension of pain, namely psychological or social pain, has been investigated in neuroimaging studies, demonstrating that social exclusion during a virtual balltossing game activates parts of the pain matrix, including the ACC, insula, and the right ventromedial prefrontal cortex (Bolling et al., 2011; Eisenberger, Lieberman, & Williams, 2003; Kross, Berman, Mischel, Smith, & Wager, 2011; Masten et al., 2009), whereby the intensity of activity in these regions was correlated with the magnitude of subjective distress. A few event-related studies utilizing the Cyberball game (Williams, Cheung, & Choi, 2000) showed that social exclusion led to an increase in ERP compared to neutral trials at N2 and P3b (Crowley et al., 2009; Weschke & Niedeggen, 2013). Hence, physical and social pain seem to share some elements concerning their neural representation, especially in affective areas that probably reflect the unpleasantness of the stimuli (for review, see Eisenberger, 2012). To our knowledge, no study exists that has compared the ERPs of both empathy for psychological or social pain and physical pain in the same experiment to further investigate the commonalities and differences between the empathic responses to the two pain types. Moreover, in light of critique of the Cyberball game suggesting that it reflects expectancy violation and conflict-based neural alarm activation rather than social pain per se (Themanson, Khatcherian, Ball, & Rosen, 2013; Weschke & Niedeggen, 2015), we designed a novel “Social Interaction Empathy Task,” where healthy participants had to rate the pain intensity of images depicting social interactions containing psychological (social) or physical pain from a FPP and a TPP during electroencephalogram (EEG) recording. This allowed us to directly compare the processing of the empathic responses to different pain qualities which could add important information to the discussion of shared brain areas involved in empathy for physical and social pain. We further aimed to investigate whether empathic abilities affect processing of the different painful social interactions. Ó 2017 Hogrefe Publishing

We anticipated differences between ERPs of neutral and physical and psychologically painful interactions. Furthermore, we expected associations with self-rated empathic abilities. In addition, we investigated whether the neural correlates of empathy for pain in social interactions showed any lateralization and perspective effects. We hypothesized higher association of right frontal brain activity with empathy for pain than with the left hemisphere, based on previous work suggesting that greater right frontal activity is associated with negative affect and withdrawal, whereas the left frontal brain is more recruited by positive affect and approach motivation (Harmon-Jones, 2003).

Material and Methods Participants Forty-seven female healthy participants were recruited for the study. The participants’ age ranged from 18 to 50 years and the mean age was 26.47 (SD = 6.44). The mean IQ was 117 (SD = 16.69) as assessed by the verbal intelligence test MWT-A (Mehrfachwahl-Wortschatz-Intelligenztest; Lehrl, Merz, Burkard, & Fischer, 1991). Exclusion criteria were an IQ below 90, neurological and psychiatric diseases, addiction disorders, severe somatic disorders, and pregnancy. The study was approved by the Ethics Committee of the Medical Faculty of the Ruhr-University Bochum. All subjects gave full informed consent in writing.

Social Interaction Empathy Task The paradigm used in this study was designed to address the question how participants distinguish between physical and psychological pain in social interactions and how the ERPs differ during the processing of the different pain qualities. The paradigm used in this study consists of images depicting situations in which a woman and a man interacted that were presented to the subjects. The pictures showed either physically or psychologically painful situations or neutral situations (i.e., no pain). For each condition there was a set of six different photos. Prior to the experiment the painfulness depicted in the images was rated by 15 healthy volunteers (pilot rating procedure is described in Flasbeck, Enzi, & Brüne, 2017). One neutral picture was subsequently excluded from the study due to its ambiguity. All pictures were taken in the same room with the same persons with the aim to keep the complexity of the pictures constant across conditions. Physically painful photos showed a scene in which a man accidentally hit a woman’s finger with a hammer, another showed the man kicking the woman’s leg instead of a ball, and another Journal of Psychophysiology (2019), 33(1), 54–63

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

Figure 1. Description of a scenario of the “Social Interaction Empathy Task.” Every trial started with a fixation cross, followed by a picture showing either a psychologically or a physically painful or a neutral social interaction. After another fixation cross phase the pain rating scale appeared whereby the participants should type their pain rating on a range from 1 (= “not painful at all”) until 9 (= “very painful”) from the first-person perspective or in the other part from the third-person perspective.

photo depicted the man pinching the hand of a woman in a door. Further scenes showed a man cutting a woman’s hand with a bread knife and with a scissor and a man scalding the woman with coffee. All images were prepared in ways strongly suggesting that the pain was caused by accident, not intentionally. Psychologically painful pictures showed the woman being abandoned by her partner and being stood up. Other pictures showed the man laughing at the woman, shouting at her, and sending her away, respectively. Another image showed the man disliking/ refusing the woman’s present. Neutral situations showed the two persons taking a meal, inspecting a shirt, and talking. Other pictures showed the man writing while the woman is having a drink, and the man making a phone call while the woman is reading a book. An example for each condition (physical and psychological pain and neutral) is given in Figure 1. The random presentation of the selected photos was carried out using Presentation® software (Neurobehavioral Systems, Inc. Version 14.9, Albany, CA). After each picture, participants had to rate the pain intensity on a scale from 1 (= “not painful at all”) to 9 (= “very painful”). In detail, every trial started with a fixation cross for 500–1,000 ms, followed by a photo showing one picture out of the three conditions for 3,000 ms, after which another fixation cross Journal of Psychophysiology (2019), 33(1), 54–63

appeared for 800–1,600 ms. Finally, the pain rating scale was shown for 3,000 ms whereby participants had to push a button indicating the subjective pain rating on the abovementioned scale (Figure 1). If they did not respond within 3,000 ms the experiments continued and the next trial started. Responses were only included in the analysis if they were given spontaneously or during the 3,000 ms period. Late responses or missed trials were not counted. Reaction time was only included in the analysis, if the judgment occurred within the 3,000 ms period. In total, there were 48 trials for each condition (8 repetitions of 6 different physical painful pictures, 8 repetitions of 6 different psychological painful, 10 (two pictures 9) repetitions of 5 different neutral pictures). The duration of the experiment was approximately 19 min in total. The experiment consisted of two parts composed of the above-mentioned trials, which differ in the perspective the participants were asked to adopt. In the first part the participants had to rate the pain intensity according to their own feelings when asked to imagine themselves being in the presented situation. That is, they had to rate from the firstperson or self-perspective. In the second part the participants were asked to rate the pain intensity the woman shown in the same pictures feels, that is, from a thirdperson or other-perspective. In both parts the woman was Ó 2017 Hogrefe Publishing

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

the person experiencing the somatic or psychological pain and the participants were asked to put themselves in her shoes. The order of the two conditions was randomized and separated by a short break. Participants were doing the task while EEG was recorded. For the analysis of pain empathy, we included the pain ratings, the reaction time (RT) and the EEG data.

EEG Recording and Analysis The EEG was recorded from 32 scalp electrodes mounted on an elastic cap in accordance with the 10–20 system by the BrainVision Recorder software (Version 1.20.001; Brain Products, Munich, Germany) and was referenced using the electrodes on the mastoids. Eye blinks and movements were controlled with an electrode located 2 cm diagonally lateral below the left eye. During recording the impedance was kept at 5 kΩ and the data acquisition was conducted with a sampling rate of 250 Hz. ERP analysis was carried out using BrainVision Analyzer 2.0 (Version 2.01.3931; Brain Products). For all conditions high and low band-pass filter were applied (0.1 and 100 Hz), and eye and muscle movements removed using Independent Component Analysis (ICA), implementing 512 ICA steps. Two hundred milliseconds before stimulus onset was used as a baseline for the calculation of ERPs which were conducted for the different conditions separately. The extracted ERPs lasted for 1,000 ms and were computed separately for Fz, F3, F4, Cz, C3, C4, Pz, P3, and P4 electrodes. Artifact rejection above 100 μV and below 100 μV was performed and ERPs were averaged over trials for each condition, perspective, and participant. Due to visual inspection of grand-averaged ERPs we decided to focus on the time frame between 330 and 450 ms after stimulus onset and the time frame of 500–700 ms reflecting a late positive potential (LPP). Mean voltage amplitudes were extracted and processed with statistical software.

Examination of Self-Reported Empathy The German version of the Interpersonal Reactivity Index (IRI; Davis, 1983), called “Saarbrücker PersönlichkeitsFragebogen” (Paulus, 2006), was used for assessment of self-reported empathic abilities. The Questionnaire consists of two cognitive subscales, namely “perspective taking” (PT) and “Fantasy” (FS), and two affective subscales: “empathic concern” (EC) and “personal distress” (PD).

Statistical Analysis Statistical analysis was conducted using SPSS Statistics version 24 (IBM Corp., Armonk, NY). Prior to statistical Ó 2017 Hogrefe Publishing

Figure 2. Pain rating results of the social interaction empathy task. Participants discriminated between painful conditions and the neutral situation and between physical and psychological pain, whereas physically painful pictures were rated as the most painful situations; ***p < .001.

calculations EEG data outliers were removed, outliers defined as the deviation of more than three standard deviations. For analysis of pain rating and reaction time 3 2 repeated-measures analyses of variance (ANOVAs) were carried out for within-subject factors “condition” (physically painful, psychologically painful, neutral) and “perspective” (first-person perspective, third-person perspective). For analysis of EEG data 3 3 2 2 3 repeatedmeasures ANOVA was conducted with the within-subject factors “lateral position” (left lateral F3, C3, P3; medial Fz, Cz, Pz; right lateral F4, C4, P4), anterior-posterior level (frontal F3, Fz, F4; central C3, Cz, C4; parietal P3, Pz, P4), time (early 330–450 ms after onset; late 500–700 ms after onset), perspective (first-person perspective, third-person perspective), and condition (physically painful, psychologically painful, neutral). For investigation of frontal lateralization effects, repeated-measures ANOVAs were used with the factors “condition” (physically painful, psychologically painful, neutral), “perspective” (first-person perspective, thirdperson perspective), and lateralization (left hemisphere F3, right hemisphere F4). The ANOVA results reported were Greenhouse-Geisser corrected. Dependent two-tailed t-tests were used for post hoc comparisons for significant effects and interactions. Correlations between the ERP results and IRI scores were calculated using Pearson’s correlation coefficient. To investigate whether the ERPs reflect pain quantities of the different conditions or different pain processing we calculated correlations between pain rating (physically painful, psychologically painful, and neutral) and ERPs (ERPs of physically painful, psychologically painful, and neutral pictures). To avoid errors due to multiple comparisons correlations were Bonferroni-Holm corrected for each electrode. For all other tests a significance level of p < .05 was chosen. Journal of Psychophysiology (2019), 33(1), 54–63

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

Figure 3. Grand-averaged ERPs of physically painful, psychologically painful, and neutral pictures over the frontal F3, F4, and Fz electrodes, the central C3, C4, and Cz electrodes, and the parietal P3, P4, and Pz electrodes. ERPs were recorded during the first-person perspective part. The voltage topographies (right) illustrate the amplitude distribution for 360 ms and 600 ms after pictures onset.

Results Social Interaction Empathy Task: Pain Ratings Repeated-measures ANOVA with the factors “condition” (physically painful, psychologically painful, neutral), and “perspective” (FPP, TPP) revealed the main effect of condition, F(1.88, 86.44) = 279.69, p < .001, indicating different ratings of pain intensity between the conditions. As there was no effect of perspective further analysis was computed with perspectives pooled for each condition. Post hoc comparisons showed that neutral pictures (M = 1.30, SD = 0.70) were rated as less painful than physically (M = 6.73, SD = 1.41) and psychologically (M = 4.48, SD = 1.61) painful pictures (neutral vs. physically painful: t46 = 23.64, p < .001; neutral vs. psychologically painful: t46 = 15.58, p < .001). In addition, physically painful

Journal of Psychophysiology (2019), 33(1), 54–63

situations were judged as more painful compared to psychologically painful pictures (t46 = 8.80, p < .001; Figure 2).

Social Interaction Empathy Task: Reaction Time Repeated-measures ANOVA revealed again a main effect of condition, F(1.75, 80.46) = 8.09, p = .001, but no effect of perspective. Further comparisons of condition with the perspectives pooled indicated that participants responded more rapidly to neutral pictures (RT neutral pictures M = 0.67, SD = 0.14 s) compared to both painful conditions (RT physically painful pictures M = 0.72, SD = 0.15 s, RT psychologically painful pictures M = 0.74, SD = 0.17 s; RT neutral vs. RT physical pain: t46 = 2.49, p = .017; RT neutral vs. RT psychological pain: t46 = 3.57, p = .001).

Ó 2017 Hogrefe Publishing

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

Figure 4. Grand-averaged ERPs and topographic maps recorded during the third-person perspective part for physically painful, psychologically painful, and neutral pictures over frontal F3, F4, and Fz electrodes, central C3, C4, and Cz electrodes, and parietal P3, P4, and Pz electrodes.

Electrophysiological Data Figures 3 and 4 show the grand-averaged ERPs to the pictures of the three conditions at the lateral and medial central, frontal, and parietal electrodes, respectively. All pictures evoked a negative component between 90 and 150 ms after picture onset (N120), followed by a peak at 170 ms (between 150 and 200 ms). Negative components followed, peaking at 270 ms (N270) and at 360 ms after stimulus onset. These ERP components were most prominent over the frontal and central electrodes. Furthermore, a late positive potential (LPP) between approximately 500–700 ms could be detected over all electrodes, reaching the maximum over the parietal electrodes. For further analysis, we focused on the early anterior component at 330–450 ms after stimulus onset and on the LPP. Repeated-measures ANOVA revealed main effects of the anterior posterior level, F(1.14, 46.82) = 75.58, p < .001, the time, F(1, 41) = 15.35, p < .001, and the condition, F(1.99, 81.65) = 75.66, p < .001. No main effects were found Ó 2017 Hogrefe Publishing

for the lateral position and the perspective. Post hoc comparisons regarding the anterior posterior level effect showed that frontal electrodes recorded more negative potentials when compared to central and parietal electrodes (frontal vs. central electrodes: t42 = 4.06, p < .001; frontal vs. parietal electrodes: t41 = 8.61, p < .001). Central electrodes recorded also more negative potentials than parietal electrodes which recorded positive potentials (central vs. parietal electrodes: t42 = 12.23, p < .001; see Figure 3). The main effect of time indicates more positive potentials at LPP compared to the earlier phase independent of the electrodes, which is also visible in Figures 3 and 4. Further analysis of the main effect condition showed that painful conditions induced a positive shift compared to the neutral condition. More precisely, physically painful pictures lead to significantly more positive potentials compared to neutral pictures (physical pain vs. neutral: t42 = 12.18, p < .001) and psychologically painful pictures (physical pain vs. psychological pain: t41 = 5.66, p < .001). Moreover, psychologically painful pictures evoked also higher Journal of Psychophysiology (2019), 33(1), 54–63

potentials than neutral pictures (psychological pain vs. neutral: t42 < 7.01, p = .001). In addition, we found a Lateral Position Anterior Posterior Level interaction, F(2.61, 106.93) = 4.38, p = .008, an Anterior Posterior Level Time interaction, F(1.15, 46.93) = 7.82, p = .006, a Lateral Position Condition interaction, F(3.65, 149.83) = 8.22, p < .001, and an Anterior Posterior Level Condition interaction, F(2.46, 100.90) = 8.9, p < .001. We also found three-way interactions but decided to spare them due to the complexity of the interpretation. The Anterior Posterior Level Time interaction finding demonstrated that the earlier component reached the peak over the frontal electrodes (frontal electrodes vs. central electrodes: t43 = 3.09, p = .004; frontal electrodes vs. parietal electrodes: t42 = 8.14, p < .001) and LPP reached the maximum over the parietal electrodes (parietal electrodes vs. frontal electrodes: t43 = 8.73, p < .001; parietal electrodes vs. central: t42 = 11.05, p < .001; central electrodes differed also significantly from the remaining electrodes: central vs. parietal electrode at early phase: t44 = 12.29, p < .001; 4.72, central vs. frontal electrode at LPP: t43 = p < .001). In addition, the earlier components were more negative over frontal and central electrodes when compared to the late component (frontal early vs. LPP: t42 = 3.06, p = .004; central early vs. LPP: t44 = 5.91, p < .001: parietal ns). Post hoc comparisons of the Lateral Position Condition interaction revealed a difference for neutral pictures when comparing left and medial electrodes (t43 = 2.73, p = .009). Besides, neutral pictures evoked more negative potentials compared to physically painful and psychologically painful pictures over medial and right electrodes (medial electrodes: physical pain vs. neutral: t45 = 2.67, p < .001; psychological pain vs. neutral: t44 = 1.55, p < .001; right electrodes: physical pain vs. neutral: t45 = 2.22, p < .001; psychological pain vs. neutral: t44 = 1.79, p < .001). Over right and medial electrodes, the potentials evoked by psychologically painful pictures also differed from activity when seeing physical pain (central: physical pain vs. psychological pain: t45 = 5.60, p < .001; right: physical pain vs. psychological pain: t45 = 2.47, p = .017). When looking at the left electrode, only a difference between the pain types and physical pain and neutral pictures emerged (physical pain vs. psychological pain: t42 = 5.10, p < .001; physical pain vs. neutral: t42 = 6.51, p < .001). Because of the above-mentioned findings and the topographic map (Figures 3 and 4) we suggested to find a frontal lateralization and conducted repeated-measures ANOVA with the factors “condition” (physically painful, psychologically painful, neutral), “perspective” (first-person perspective, third-person perspective), and lateralization (left hemisphere F3, right hemisphere F4) for the time

Journal of Psychophysiology (2019), 33(1), 54–63

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

frame of 330–450 ms after picture onset. Again, we found a reliable main effect of condition, F(1.95, 83.87) = 11.43, p < .001, and a Lateralization Condition interaction, F(1.89, 81.33) = 4.23, p = .017. The interaction displays a difference between neutral and painful (painful conditions pooled) ERPs only at the right hemisphere (F4 electrode), but not over the left F3 (F4: pain vs. neutral: t45 = 4.78, p < .001; F3: pain vs. neutral: ns). A closer look at the painful conditions indicated that the ERPs of physically and psychologically painful pictures showed a positive shift compared to neutral pictures over F4 (physical pain vs. neutral: t46 = 5.11, p < .001; psychological pain vs. neutral: t46 = 3.26, p = .002). In addition, the ERPs of physical pain differed from psychologically pain (physical pain vs. psychological pain: t46 = 2.09, p = .042). Over F3, we found only difference between neutral and physically painful pictures (t45 = 2.59, p = .013). Analysis of the Anterior Posterior Level Condition interaction revealed that all conditions differ from all other conditions at central and parietal levels, but not at the frontal level (central: physical vs. psychological pain: t44 = 7.40, p < .001; physical pain vs. neutral: t44 = 13.55, p < .001; psychological pain vs. neutral: t45 = 6.30, p < .001, parietal: physical vs. psychological pain: t44 = 7.62, p < .001; physical pain vs. neutral: t45 = 14.74, p < .001; psychological pain vs. neutral: t44 = 7.82, p < .001). Over the frontal electrodes, ERPs of physical and psychological pain did not differ (physical vs. psychological pain: ns; physical pain vs. neutral: t43 = 4.62, p < .001; psychological pain vs. neutral: t43 = 3.86, p < .001). One reason might be that the differentiation of mental and physical pain takes place at the right hemisphere which was covered by the other electrodes in this comparison.

Correlation Analysis We detected significant correlations between the ERP results and self-rated empathic abilities assessed by IRI questionnaires. ERPs over the left central and medial parietal electrodes of painful conditions correlated significantly with the perspective taking scale of the IRI (C3 psychological pain – PT: r = .37, p = .011; Pz physical pain: r = .32, p = .039), suggesting a modulation effect of perspective taking abilities on processing of painful social interactions. If the differences in ERPs reflect different pain intensities instead of processing of the different social interactions, we expected to find correlations between pain rating and ERPs. We calculated correlations between pain ratings of physically painful, psychologically painful, and neutral interactions (perspectives pooled) and ERPs of physically painful, psychologically painful, and neutral pictures and could not find any significant association.

Ó 2017 Hogrefe Publishing

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

Discussion The aim of the present study was to investigate and characterize the neural correlates of empathy for physical and psychological pain. Previous studies of empathy for pain investigated either empathy for physical pain or social pain, which was mainly assessed via social exclusion, respectively. Therefore, we developed the “Social Interaction Empathy Task” which enabled us to investigate ERPs to both pain dimensions and neutral interactions at the same time. Due to the different ratings of pain intensity, we anticipated differences in ERP between physical and psychological pain and neutral interactions. Because participants judged physically painful pictures as more painful than psychologically painful and neutral images we assumed that ERPs to physical pain would differ most strongly from neutral pictures. In the present study, the higher pain ratings were accompanied by longer reaction times needed for the evaluation of painful pictures, suggesting that painful interactions captured the participants’ attention more than neutral scenarios. Consistent with our hypotheses concerning ERP responses, we found a main effect of condition indicating that physical pain evoked a positive shift relative to neutral and psychological pain. Moreover, psychological pain also differed significantly from neutral conditions. In accordance with other studies, we found an early component over frontal-central areas and a late component reaching the maximum over the parietal electrodes (Fan & Han, 2008). In our study we did not find any effect of perspective on pain rating, reaction time, or ERP responses. The study of Li and Han, using the painful and neutral pictures of hands that should be rated from the first- and third-person perspective, reported interactions of pain with perspective in reaction time, accuracy, and ERP at 370–420 ms over the central-parietal area, whereas there was no main effect of perspective (Li & Han, 2010). Another study could not find significant differences between the first- and third-person perspective in a behavior task testing empathy for pain (Abu-Akel et al., 2015). In contrast, a functional magnetic resonance imaging (fMRI) study could demonstrate that pain rating from both the FPP and TPP activated the neural network of pain processing. In addition, the FPP yielded a higher involvement of the secondary somatosensory cortex, ACC, and insula, whereas the right temporo-parietal junction was more activated during the TPP (Jackson et al., 2006). Taken together, these results suggest that perspective effects critically depend on the methodological approach and task sensitivity. In our study, the analysis of hemispheric differences revealed an asymmetrical frontal involvement in pain empathy. We found differences in ERPs between neutral Ó 2017 Hogrefe Publishing

and painful conditions only at the right hemisphere, but not at the left frontal hemisphere. The difference in ERP responses to the observation of physical and psychological painful conditions showed the same pattern as observed for the other electrodes. These findings are inconsistent with the results of Fan and Han (2008) reporting a larger effect of pain empathy over the left hemisphere. Here, it is important to note that the “classical” empathy for pain tasks is difficult to compare to the social interaction empathy task because the social complexity of the task might lead to different processing of pain. Thus, the higher involvement of the right hemisphere could be due to the social component of the task which in turn leads to higher activity of the right dorsolateral prefrontal cortex (DLPFC). There are different theories existing describing the role of the DLPFC in overriding self-interest motives or the involvement of the DLPFC in regulating emotional responses. Supporting the first theory, Knoch et al. demonstrated that repetitive transcranial magnetic stimulation (rTMS) of the right but not left DLPFC reduced the rejections of unfair co-players in the Ultimatum Game, indicating that subjects’ impact of fairness and self-interest goals (here the economic profit) were shifted (Knoch, Pascual-Leone, Meyer, Treyer, & Fehr, 2006). It was also shown with fMRI that unfair offers evoked activity in the DLPFC and the insula (Sanfey, Rilling, Aronson, Nystrom, & Cohen, 2003). Another study reported that rTMS over the right DLPFC increased costly punishment in the Dictator Game, whereas empathy moderated this effect (Brüne et al., 2012). Regarding the second theory, it was reported that the DLPFC and the ACC were involved in down- and upregulation of negative emotions (Ochsner et al., 2004). In addition, Harmon-Jones concluded that the asymmetrical frontal cortical activity is not solely related due to experience and expression of emotions with a higher involvement of the left hemisphere in positive and the right hemisphere in negative emotions, but also caused by motivation (Harmon-Jones, 2003). Aside from these theories, De Greck reported stronger hemodynamic responses in Chinese participants in the left DLPFC during empathy with anger compared to German participants (de Greck et al., 2012). Consequently, further studies are necessary to investigate the functional role of the frontal cortical areas. Our correlation analysis showed that perspective taking abilities were related to the ERPs of painful pictures. Importantly, we did not find correlations between pain rating and ERPs, suggesting that ERPs do not reflect the pain intensity but pain quality, that is, the differential processing of physical and psychological pain. To conclude, we found an early and a late empathic response with a higher ERP response to physical than to psychological pain. Moreover, we found painful situations Journal of Psychophysiology (2019), 33(1), 54–63

being discernible at right frontal cortical areas, and correlations of lateral central and parietal ERPs emerged with empathic abilities. However, our findings imply a limitation insofar as our results are not generalizable for both sexes because we recruited exclusively female participants. Future research should also include males and focus on the neuronal basis of differences between psychological and physical pain and disentangle the influence of the perspective that participants adopt regarding empathy and pain processing Ethical Standards The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Conflict of Interest None. Acknowledgment We thank Elke Köhler for supporting the EEG recordings and good cooperation.

References Abu-Akel, A., Palgi, S., Klein, E., Decety, J., & Shamay-Tsoory, S. (2015). Oxytocin increases empathy to pain when adopting the other- but not the self-perspective. Social Neuroscience, 10, 7–15. doi: 10.1080/17470919.2014.948637 Bolling, D. Z., Pitskel, N. B., Deen, B., Crowley, M. J., McPartland, J. C., Mayes, L. C., & Pelphrey, K. A. (2011). Dissociable brain mechanisms for processing social exclusion and rule violation. NeuroImage, 54, 2462–2471. doi: 10.1016/j.neuroimage.2010. 10.049 Botvinick, M., Jha, A. P., Bylsma, L. M., Fabian, S. A., Solomon, P. E., & Prkachin, K. M. (2005). Viewing facial expressions of pain engages cortical areas involved in the direct experience of pain. NeuroImage, 25, 312–319. doi: 10.1016/j.neuroimage. 2004.11.043 Brüne, M., Scheele, D., Heinisch, C., Tas, C., Wischniewski, J., & Güntürkün, O. (2012). Empathy moderates the effect of repetitive transcranial magnetic stimulation of the right dorsolateral prefrontal cortex on costly punishment. PLoS One, 7, e44747. doi: 10.1371/journal.pone.0044747 Crowley, M. J., Wu, J., McCarty, E. R., David, D. H., Bailey, C. A., & Mayes, L. C. (2009). Exclusion and micro-rejection: Event-related potential response predicts mitigated distress. Neuroreport, 20, 1518–1522. doi: 10.1097/WNR.0b013e328330377a Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach. Journal of Personality and Social Psychology, 44, 113–126. de Greck, M., Shi, Z., Wang, G., Zuo, X., Yang, X., Wang, X., . . . Han, S. (2012). Culture modulates brain activity during empathy with anger. NeuroImage, 59, 2871–2882. doi: 10.1016/j. neuroimage.2011.09.052

Journal of Psychophysiology (2019), 33(1), 54–63

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

Decety, J., Yang, C. Y., & Cheng, Y. (2010). Physicians downregulate their pain empathy response: An event-related brain potential study. NeuroImage, 50, 1676–1682. doi: 10.1016/ j.neuroimage.2010.01.025 Eisenberger, N. I. (2012). The neural bases of social pain: Evidence for shared representations with physical pain. Psychosomatic Medicine, 74, 126–135. doi: 10.1097/PSY.0b013e3182464dd1 Eisenberger, N. I., Lieberman, M. D., & Williams, K. D. (2003). Does rejection hurt? An FMRI study of social exclusion. Science, 302, 290–292. doi: 10.1126/science.1089134 Fan, Y., & Han, S. (2008). Temporal dynamic of neural mechanisms involved in empathy for pain: An event-related brain potential study. Neuropsychologia, 46, 160–173. doi: 10.1016/ j.neuropsychologia.2007.07.023 Flasbeck, V., Enzi, B., & Brüne, M. (2017). Altered empathy for psychological and physical pain in borderline personality disorder. Journal of Personality Disorders, 31, 689–708. doi: 10.1521/pedi_2017_31_276 Gonzalez-Liencres, C., Shamay-Tsoory, S. G., & Brüne, M. (2013). Towards a neuroscience of empathy: Ontogeny, phylogeny, brain mechanisms, context and psychopathology. Neuroscience and Biobehavioral Reviews, 37, 1537–1548. doi: 10.1016/ j.neubiorev.2013.05.001 Han, S., Fan, Y., & Mao, L. (2008). Gender difference in empathy for pain: An electrophysiological investigation. Brain Research, 1196, 85–93. doi: 10.1016/j.brainres.2007.12.062 Harmon-Jones, E. (2003). Clarifying the emotive functions of asymmetrical frontal cortical activity. Psychophysiology, 40, 838–848. doi: 10.1111/1469-8986.00121 Jackson, P. L., Brunet, E., Meltzoff, A. N., & Decety, J. (2006). Empathy examined through the neural mechanisms involved in imagining how I feel versus how you feel pain. Neuropsychologia, 44, 752–761. doi: 10.1016/j.neuropsychologia.2005.07.015 Jackson, P. L., Meltzoff, A. N., & Decety, J. (2005). How do we perceive the pain of others? A window into the neural processes involved in empathy. NeuroImage, 24, 771–779. doi: 10.1016/ j.neuroimage.2004.09.006 Knoch, D., Pascual-Leone, A., Meyer, K., Treyer, V., & Fehr, E. (2006). Diminishing reciprocal fairness by disrupting the right prefrontal cortex. Science, 314, 829–832. doi: 10.1126/ science.1129156 Kross, E., Berman, M. G., Mischel, W., Smith, E. E., & Wager, T. D. (2011). Social rejection shares somatosensory representations with physical pain. Proceedings of the National Academy of Sciences of the United States of America, 108, 6270–6275. doi: 10.1073/pnas.1102693108 Lamm, C., Decety, J., & Singer, T. (2011). Meta-analytic evidence for common and distinct neural networks associated with directly experienced pain and empathy for pain. NeuroImage, 54, 2492–2502. doi: 10.1016/j.neuroimage.2010.10.014 Lehrl, S., Merz, J., Burkard, G., & Fischer, B. (1991). MWT-A: Mehrfachwahl-Wortschatz-Intelligenztest: (Parallelform zum MWT-B) [Multiple Choice Vocabulary Test]. Erlangen, Germany: Perimed-Fachbuch-Verlag-Gesellschaft. Li, W., & Han, S. (2010). Perspective taking modulates eventrelated potentials to perceived pain. Neuroscience Letters, 469, 328–332. doi: 10.1016/j.neulet.2009.12.021 Masten, C. L., Eisenberger, N. I., Borofsky, L. A., Pfeifer, J. H., McNealy, K., Mazziotta, J. C., & Dapretto, M. (2009). Neural correlates of social exclusion during adolescence: Understanding the distress of peer rejection. Social Cognitive and Affective Neuroscience, 4, 143–157. doi: 10.1093/scan/nsp007 Ochsner, K. N., Ray, R. D., Cooper, J. C., Robertson, E. R., Chopra, S., Gabrieli, J. D., & Gross, J. J. (2004). For better or for worse: Neural systems supporting the cognitive down-and

Ó 2017 Hogrefe Publishing

V. Flasbeck & M. Brüne, Neural Correlates of Empathy

up-regulation of negative emotion. NeuroImage, 23, 483–499. doi: 10.1016/j.neuroimage.2004.06.030 Paulus, C. (2006). Saarbrücker Persönlichkeits-Fragebogen (SPF) [Saarbrücken personality questionnaire]. Based on the interpersonal reactivity index (IRI) including new items (V3. 0). Saarbrücken, Germany: Universität des Saarlandes. Saarela, M. V., Hlushchuk, Y., Williams, A. C., Schürmann, M., Kalso, E., & Hari, R. (2007). The compassionate brain: Humans detect intensity of pain from another’s face. Cerebral Cortex, 17, 230–237. doi: 10.1093/cercor/bhj141 Sanfey, A. G., Rilling, J. K., Aronson, J. A., Nystrom, L. E., & Cohen, J. D. (2003). The neural basis of economic decisionmaking in the ultimatum game. Science, 300, 1755–1758. doi: 10.1126/science.1082976 Shamay-Tsoory, S. G. (2011). The neural bases for empathy. Neuroscientist, 17, 18–24. doi: 10.1177/1073858410379268 Shamay-Tsoory, S. G., Aharon-Peretz, J., & Perry, D. (2009). Two systems for empathy: A double dissociation between emotional and cognitive empathy in inferior frontal gyrus versus ventromedial prefrontal lesions. Brain, 132, 617–627. doi: 10.1093/ brain/awn279 Singer, T. (2006). The neuronal basis and ontogeny of empathy and mind reading: Review of literature and implications for future research. Neuroscience and Biobehavioral Reviews, 30, 855– 863. doi: 10.1016/j.neubiorev.2006.06.011 Singer, T., Seymour, B., O’Doherty, J., Kaube, H., Dolan, R. J., & Frith, C. D. (2004). Empathy for pain involves the affective but not sensory components of pain. Science, 303, 1157–1162. doi: 10.1126/science.1093535 Themanson, J. R., Khatcherian, S. M., Ball, A. B., & Rosen, P. J. (2013). An event-related examination of neural activity during social interactions. Social Cognitive and Affective Neuroscience, 8, 727–733. doi: 10.1093/scan/nss058

Ó 2017 Hogrefe Publishing

van der Heiden, L., Scherpiet, S., Konicar, L., Birbaumer, N., & Veit, R. (2013). Inter-individual differences in successful perspective taking during pain perception mediates emotional responsiveness in self and others: An fMRI study. NeuroImage, 65, 387–394. doi: 10.1016/j.neuroimage.2012.10.003 Weschke, S., & Niedeggen, M. (2013). The effect of the physical presence of co-players on perceived ostracism and eventrelated brain potentials in the cyberball paradigm. PLoS One, 8, e71928z. doi: 10.1371/journal.pone.0044747 Weschke, S., & Niedeggen, M. (2015). ERP effects and perceived exclusion in the Cyberball paradigm: Correlates of expectancy violation? Brain Research, 1624, 265–274. doi: 10.1016/ j.brainres.2015.07.038 Williams, K. D., Cheung, C. K., & Choi, W. (2000). Cyberostracism: Effects of being ignored over the Internet. Journal of Personality and Social Psychology, 79, 748. doi: 10.1037/0022-3514.79. 5.748 Received March 18, 2016 Accepted January 7, 2017 Published online November 21, 2017 Martin Brüne LWL University Hospital Department of Psychiatry, Psychotherapy and Preventive Medicine Division of Cognitive Neuropsychiatry and Psychiatric Preventive Medicine Ruhr-University Bochum Alexandrinenstr. 1 44791 Bochum Germany martin.bruene@rub.de

Journal of Psychophysiology (2019), 33(1), 54–63