International Journal of Psychophysiology 97 (2015) 93–98
Contents lists available at ScienceDirect
International Journal of Psychophysiology journal homepage: www.elsevier.com/locate/ijpsycho
Depressive states amplify both upward and downward counterfactual thinking Xue Feng a, Ruolei Gu b,⁎, Fucheng Liang c, Lucas S. Broster d, Yunzhe Liu a, Dandan Zhang e, Yue-jia Luo e,f a
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing 100875, China Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing 100101, China Academic Administration Section, Tianjin Normal University, Tianjin 300387, China d Department of Behavioral Science, University of Kentucky College of Medicine, Lexington, KY 40506, USA e Institute of Affective and Social Neuroscience, Shenzhen University, Shenzhen, 518060, China f Sichuan Research Center of Applied Psychology, Chengdu Medical College, Chengdu 610500, China b c
a r t i c l e
i n f o
Article history: Received 12 December 2014 Received in revised form 6 April 2015 Accepted 13 April 2015 Available online 30 April 2015 Keywords: Depression Counterfactual thinking Outcome evaluation Event-related potential P3
a b s t r a c t Depression has been linked to counterfactual thinking in many behavioral studies, but the direction of this effect remains disputed. In the current study, the relationship between depression and counterfactual thinking was examined using the event-related potential (ERP) technique. In a binary choice gambling task, outcome feedback of the chosen option and that of the alternative option were both provided, so as to elicit the process of counterfactual comparison. By investigating ERP signals in response to outcome presentation, we discovered that when the fictive outcome was better or worse than the factual outcome, the amplitude of the P3 component was positively correlated with individual levels of depression, but not levels of anxiety. These results indicate that depression strengthens both upward counterfactual thinking and downward counterfactual thinking. The implication of this finding to clinical research is discussed. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Depression is a state of low mood associated with multiple psychophysiological symptoms including stress, anhedonia, and abnormalities in appetite and sleep (Krishnan and Nestler, 2008). High levels of depression give rise to biased thoughts and beliefs, such as negative self-image and self-focused attention (Beevers et al., 2011). This effect may increase the risk for the development and recurrence of depressive disorder (Gotlib and Joormann, 2010). Therefore, investigating altered belief systems in depression has important implications for pathophysiological research. The current study explores the relationship between depression and counterfactual thinking, which refers to the process of comparing reality and “what might have been” had a different decision been made (Coricelli et al., 2007). Comparisons based on alternatives that improve on reality (i.e., upward counterfactual) generate feelings of regret and disappointment (Mandel, 2003; Zeelenberg et al., 1998), while comparisons based on alternatives that worsen reality (i.e., downward counterfactual) generate feelings of rejoice and
⁎ Corresponding author. E-mail address: [email protected] (R. Gu).
http://dx.doi.org/10.1016/j.ijpsycho.2015.04.016 0167-8760/© 2015 Elsevier B.V. All rights reserved.
gratification (Galinsky et al., 2002; Medvec and Savitsky, 1997). The influence of depression on counterfactual thinking has been examined by many studies, but the precise nature of this influence is still debated. Markman and Miller (2006) reported that participants suffering from severe depressive symptoms generated counterfactuals that were less controllable (see also Eryilmaz et al., 2014; Monroe et al., 2005; Zhou and Kong, 2009). In contrast, depressed patients showed attenuated counterfactual thinking in Chase et al. (2010) (see also Quelhas et al., 2008). Finally, some studies have found no correlation between levels of depression and counterfactual thinking (e.g., Landman et al., 1995). In our opinion, these heterogeneous findings may be attributable in part to methodological issues that distinguish some or all of those studies. First, most previous studies devoted to this topic have relied on selfreport instrument rather than clandestine tools such as measures of brain activity (see Eryilmaz et al., 2014, for an exception). The limitations of these types of assessments are worth noting, including response bias and socially desirable responding (Crowley et al., 2009). Second, many studies estimate counterfactual thinking by asking retrospective questions (e.g., Markman and Miller, 2006), which may be vulnerable to the memory bias associated with depression (Bar, 2009). Finally, while most studies focus on the potential relation between depression and upward counterfactual which elicits feelings of regret, whether
94
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
depression affects downward counterfactual remains largely unanswered. Eryilmaz et al. (2014) recently reported that activities of the brain regions associated with gratification (i.e., a consequence of downward counterfactual thinking) were negatively correlated with Beck Depression Inventory (BDI) scores during resting state. This study investigated the impact of depression on counterfactual thinking in a decision-making scenario (e.g., Chase et al., 2010) with event-related potentials (ERPs) as a neural measurement. In this task, participants chose between two options and received positive or negative outcomes of the chosen option and the alternative option (Gu et al., 2011b; Yeung and Sanfey, 2004). Thus, the comparison between positive alternative outcome (AO for short) and negative chosen outcome (CO for short) elicits upward counterfactual thinking, while the comparison between negative AO and positive CO elicits downward counterfactual thinking (Giorgetta et al., 2012). For instance, when a participant discovers that the option he/she selected led to monetary loss (CO) while the other option could have resulted in monetary gain (AO), he/she may feel regret and imagine the preferable alternate selection (upward counterfactual thinking). Two ERP components, the feedback-related negativity (FRN) and the P3, were chosen as they are the major biomarkers of outcome evaluation (Gehring and Willoughby, 2002; Yeung and Sanfey, 2004). Both components have been reported to be affected by depression levels (Proudfit, 2015; Proudfit et al., in press; Roschke and Wagner, 2003). Numerous studies have confirmed that both the FRN and the P3 are sensitive to the comparison between subjective expectation and reality (Hajcak et al., 2005, 2007; Wu and Zhou, 2009). In order to clarify the discrete roles played by these components in counterfactual contexts, Osinsky et al. (2014) recently proposed that the FRN indicates a binary categorization of gain versus no-gain for both CO and AO, while the P3 indicates motivational salience derived from counterfactual comparisons of CO and AO. Consistent with this idea, Liang et al. (2015) reveal that counterfactual comparisons manifest on the P3 rather than the FRN (see also Gu et al., 2011b; Wu and Clark, 2014). Accordingly, we suggest that the P3 elicited by the comparison between CO and AO reflects the process of counterfactual thinking when other major factors that could modulate this component (e.g., event probability, stimulus novelty, and task complexity; for reviews, see Polich, 2007; Polich and Criado, 2006) are controlled. We predicted that P3 amplitude in response to the comparison between CO and AO would increase as a function of individual differences in depressive traits, indicating a stronger tendency of counterfactual thinking among depressed individuals. In contrast, such associations would be absent for FRN amplitude. 2. Methods 2.1. Participants 20 right-handed Chinese students (11 females; mean age 22.75 years, SD = 1.68) from three universities in Tianjin participated in the experiment. All the participants were included in the final sample. All had normal or corrected-to-normal vision. None of the participants had sought medication for emotional problems before. All the participants denied regular use of any substance with the potential to affect the central nervous system. None had a history of neurological disease. None of the participants reported knowledge of the Cyrillic alphabet. All participants gave their informed consent prior to the experiment. The study was approved by the ethics committee of Academy of Psychology and Behavior, Tianjin Normal University.
electrically shielded room. Each trial began with the presentation of two options represented as Cyrillic letters (“д” and “ю” respectively), each of which was presented inside a white rectangle (2.51° × 2.51° of visual angle) appearing on either side of a fixation point. The position of these letters was counterbalanced across trials. Participants selected one option by pressing the “F” or “J” keys on the keyboard with left or right index finger, respectively. The selected rectangle was outlined in red for 500 ms. After that, the AO was presented in the unselected rectangle for 1000 ms simultaneously with the disappearance of the two letters. Then the AO faded away, leaving the rectangles and the fixation point on screen for 500 ms. Finally, the CO was presented in the chosen rectangle for 1000 ms. The formal task consisted of two blocks of 100 trials each. Stimulus display and behavioral data acquisition were conducted using E-Prime software 1.1 (PST, Inc., Pittsburgh, PA). There were two kinds of outcome valence: positive (“+”) and negative (“−”). A positive and a negative CO indicated that participants gained or lost one point in the current trial, respectively (Fig. 1). Unbeknownst to the participant, no matter which option was chosen, the probability of receiving a positive or negative CO or AO in each trial was 50%. Moreover, the valences of AO and CO were independent of one another. Thus, we suggest that event probability, stimulus novelty, and task complexity have been controlled across conditions. After participants finished the task, he or she was instructed to fill the Chinese version of Zung's self-rating depression scale (SDS) and the trait form of Spielberger's State-Trait anxiety inventory (STAI-T), both of which have demonstrated good internal consistency reliability, convergent validity, and discriminate validity (STAI-T: Shek, 1993; Spielberger et al., 1983. SDS: Shu, 1993; Zung et al., 1965). The STAI-T was included in this study in appreciation of the strong relationship between anxiety and depression (Stavrakaki and Vargo, 1986). During debriefing, participants were informed that the outcomes were predetermined and there was no optimal strategy for the task. Accordingly, each participant was paid 30 Chinese Yuan (approximately 5 US dollars) for participation. 2.3. Electrophysiological recording and preprocessing The EEG activity was recorded from 64 scalp sites using tin electrodes mounted in an elastic cap (NeuroScan, Inc., Herndon, VA), with an online reference to the left mastoid and off-line algebraic rereference to the average of the left and right mastoids. Horizontal electrooculogram (EOG) was recorded from electrodes placed at the outer canthi of both eyes. Vertical EOG was recorded from electrodes placed above and below the left eye. All interelectrode impedance was maintained at b5 k Ω. EEG and EOG signals were amplified with a 0.05–100 Hz bandpass filter and continuously sampled at 500 Hz/channel. During the off-line analysis, ocular artifacts were removed from the EEG signal using a regression procedure implemented in the Neuroscan software (Semlitsch et al., 1986). The EEG was digitally filtered through a zero phase shift, with 0.05–30 Hz band-pass filtering for the FRN and 2 Hz low-pass filtering for the P3 component (Gu et al., 2011a; Yeung and Sanfey, 2004). The data were then averaged into 1000 ms epochs (200 ms baseline) time-locked to outcome presentation. Any trials where EEG voltages exceeded a threshold of ± 100 μV during the recording epoch were excluded from the analysis. After the artifact rejection processing, the numbers of trials survived in each condition were as follows: AO (+) CO (+): 45.63 of the 50 trials (91.26%); AO (−) CO (+): 44.84 (89.68%); AO (+) CO (−): 45.21 (90.42%); and AO (−) CO (−): 46.05 (92.10%).
2.2. Procedure 2.4. Data analysis The task procedure replicated that of Gu et al. (2011b). Before the experiment, participants were informed about the rules and framework of the task and encouraged to respond so as to maximize the total score. Participants sat approximately 100 cm from a computer screen in an
The time windows for peak identification were determined through visual inspection of grand-averaged waveforms. The electrode at which the ERP components reached their maximum was detected along the
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
95
Fig. 1. The schematic diagrams of two example trials. RT = response time. (a) A negative alternative outcome (AO) and a positive chosen outcome (CO) produce a downward counterfactual scenario. (b) A positive AO and a negative CO produce an upward counterfactual scenario.
midline (Fz, FCz, Cz, CPz, Pz, POz and Oz); thereafter, the arithmetic mean of peak amplitudes of this electrode and 8 adjacent electrodes was calculated for further analysis (Gu et al., 2011a). For all the analyses listed below, the significance level was set at 0.05. Post-hoc testing of significant main effects was conducted using the LSD method. Significant interactions were analyzed using simpleeffects models. Statistical analysis was performed using SPSS 19.0 (IBM, Armonk, NY).
CPz (M = 11.14 μV, SD = 1.08 μV). Accordingly, the means of this electrode and 8 adjacent electrodes (C1, Cz, C2, CP1, CP2, P1, Pz, and P2) were entered into further analysis. The main effect of AO valence was not significant (t(19) = 0.998, p = .331) (Fig. 2c). No significant correlation was detected between the P3 amplitude and personality scores (ps N .05).
3. Results 3.1. Emotional scales SDS score was 40.61 (SD = 7.70, range = 30–58) and STAI-T score was 40.50 (SD = 9.28, range = 24–62), respectively. The correlation between SDS and STAT-T was significant (r = .742, p b .001). 3.2. Behavioral results Participants selected “д” and “ю” 50.07% (SD = 12.44) and 49.45% (SD = 12.42) of the time, respectively. The averaged reaction time for decision-making was 0.90 s (SD = 0.94). None of these behavioral measures were correlated with SDS or STAI-T scores (ps N .05). Seeing that there was no optimal strategy in the task, the behavioral data were not analyzed further. 3.3. ERP results 3.3.1. Alternative outcome (AO, presented antecedently) The ERP responses elicited by AO were entered into pairedsamples t tests using AO valence (positive/negative) as the withinsubject factor. 3.3.1.1. The FRN component. The FRN was measured as the most negative peak within the 200–400 ms time window. It was largest at electrode FCz (M = 2.80 μV, SD = 0.47 μV). Accordingly, the means of this electrode and 8 adjacent electrodes (F1, Fz, F2, FC1, FC2, C1, Cz and C2) were entered into further analysis. The main effect of AO valence approached significance (t(19) = 1.856, p = .079), with the FRN showed a tendency to be greater in response to positive AO than to negative AO (3.24 μV vs. 4.58 μV) (Fig. 2a). No significant correlation was detected between the FRN amplitude and personality scores in each condition (ps N .05). 3.3.1.2. The P3 component. The P3 was measured as the most positive peak within the 300–500 ms time window. It was largest at electrode
Fig. 2. (a) Grand-average ERPs elicited by AO with a 0.05–30 Hz band-pass filter at the FCz recording site, where the FRN reached its maximum. (b) Grand-average ERPs elicited by CO at the CPz site, where the P3 reached its maximum, with a 0.05–30 Hz band-pass filter and (c) with a 2 Hz low-pass filter. The time point “0” indicates the AO onset. The time windows for peak detection of the FRN (a) and P3 (b, c) were indicated as light gray shaded areas. The tic markers indicate the timepoints in which the P3 peaks were detected.
96
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
3.3.2. Chosen outcome (presented subsequently) The ERP responses elicited by CO were entered into 2 (CO valence: positive/negative) × 2 (AO valence: positive/negative) ANOVAs. 3.3.2.1. The FRN component. The FRN was measured as the most negative peak within the 200–400 ms time window. It was largest at electrode FCz (M = 4.95 μV, SD = 0.75 μV). Accordingly, the means of this electrode and 8 adjacent electrodes (F1, Fz, F2, FC1, FC2, C1, Cz, and C2) were entered into further analysis. The main effect of CO valence was significant (F(1, 19) = 29.705, p b .001); the FRN was greater in response to negative CO than positive CO (3.35 μV vs. 7.38 μV). The main effect of AO valence was also significant (F(1, 19) = 14.893, p b .005); the FRN was greater when following negative AO than positive AO (4.39 μV vs. 6.34 μV). The AO × CO interaction was insignificant (F(1, 19) = 0.893, p = .356) (Fig. 3a). No significant correlation was detected between the FRN amplitude and personality scores (ps N .05). 3.3.2.2. The P3 component. The P3 was measured as the most positive peak within the 300–500 ms time window. It was largest at electrode CPz (M = 17.13 μV, SD = 1.37 μV). Accordingly, the mean amplitudes of this electrode and 8 adjacent electrodes (C1, Cz, C2, CP1, CP2, P1, Pz, and P2) were entered into ANOVA analysis. The main effect of CO valence approached significance (F(1, 19) = 3.795, p =
.066); positive CO showed a tendency to elicit a larger P3 compared to negative CO (11.76 μV vs. 10.92 μV). The main effect of AO valence was significant (F(1, 19) = 23.429, p b .01); the CO elicited a larger P3 when it was preceded by positive AO than negative AO (12.44 μV vs. 10.24 μV). The AO × CO interaction was also significant (F(1, 19) = 21.561, p b .01); the influence of AO on P3 amplitude was significant (F(1, 19) = 37.711, p b 0.001) when CO was positive, but was insignificant when CO was negative (F(1, 19) = 0.536, p = .473) (Fig. 3c). The means and standard deviations of the P3 amplitude and latency in each condition are presented in Table 1. Pearson correlation analysis revealed significant correlations between SDS scores and P3 amplitude when positive CO followed negative AO (r = 0.457, p b .05) and when negative CO followed positive AO (r = 0.462, p b .05). In brief, participant's level of depression was correlated with the P3 component when the valences of AO and CO were different (Fig. 4). No significant correlation was detected between the P3 amplitude and STAI-T scores. To assess the unique contributions of depression, a linear regression was performed (see Foti and Hajcak, 2009, for similar methods). When entering both SDS and STAI-T scores to predict the P3 amplitude elicited by positive CO following negative AO, the regression coefficient for depression was significant (β = 0.377, p b .05) but anxiety was not (β = −0.168, p N .05); the same was true in the condition of negative CO following positive AO, such that depression (β = 0.462, p b .05), but not anxiety (β = 0.290, p N .05), was a significant predictor. 4. Discussion Previous studies have argued about whether and how depression is associated with counterfactual thinking (e.g., Chase et al., 2010; Markman and Miller, 2006). Using the ERP technique, the current study investigated this issue in a trial-by-trial decision-making task. The ERP results reveal that individual depressive level was positively correlated with the amplitude of the P3 component elicited by chosen outcome presentation in two conditions. Specifically, the correlations were significant when a positive CO appeared after a negative AO (i.e., the real outcome was better than the alternative) and when a negative CO appeared after a positive AO (i.e., the real outcome was worse than the alternative). As indicated in the Section 1, these two conditions represent situations that motivate downward and upward counterfactual thinking, respectively. According to Osinsky et al. (2014), the P3 but not the FRN is involved in counterfactual comparisons (see also Liang et al., 2015). This idea is confirmed in the current study, such that the P3 rather than the FRN showed a significant interaction between AO and CO. We suggest that the current ERP findings indicate that depressive participants were more likely to be engaged in counterfactual thinking than their non-depressive counterparts. Presumably, this phenomenon is linked to the observation that depressed individuals are more susceptible to irrelevant thoughts and have more difficulty “getting over it” and refocusing on upcoming challenges (Lepore et al., 1996; Monroe et al., 2005). In addition, extravagant counterfactual thinking accounts for overgeneral memory bias and imagination (Markman et al., 1995; Miller and Taylor, 1995), which contributes to severe depressive illness (Meyers et al., 2006; Ridout
Table 1 The P3 amplitude and latency in each condition. The values in this table are arithmetic means of 9 electrodes (C1, Cz, C2, CP1, CPz, CP2, P1, Pz, and P2).
Fig. 3. (a) Grand-average ERPs elicited by CO with a 0.05–30 Hz band-pass filter at the FCz recording site, where the FRN reached its maximum. (b) Grand-average ERPs elicited by CO at the CPz site, where the P3 reached its maximum, with a 0.05–30 Hz band-pass filter and (c) with a 2 Hz low-pass filter. The time point “0” indicates the CO onset. The time windows for peak detection of the FRN (a) and P3 (b, c) were indicated as light gray shaded areas. The tic markers indicate the timepoints in which the P3 peaks were detected.
Experimental conditions
P3 amplitude mean (SD)
P3 latency mean (SD)
AO (+) AO (−) AO (+) CO (+) AO (−) CO (+) AO (+) CO (−) AO (−) CO (−)
7.56 μV (3.01) 7.18 μV (3.56) 14.20 μV (5.13) 9.32 μV (3.85) 10.67 μV (4.02) 11.17 μV (4.88)
398.31 ms (49.88) 392.29 ms (67.53) 400.29 ms (41.53) 353.74 ms (58.18) 383.92 ms (49.43) 438.48 ms (49.86)
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
97
press; but see Mueller et al., in press), which contradicts our current findings. We suspect that this inconsistency occurred because in those studies, participants were only aware of the outcome associated with their decisions; thus, they did not engage in counterfactual thinking (Proudfit, 2015). As an alternative explanation, previous studies focusing on outcome evaluation have suggested that the FRN and P3 represent quick detection and deliberate evaluation of the current event, respectively (Gu et al., 2011a; Philiastides et al., 2010; Wu and Zhou, 2009). Therefore, the preeminence of the P3 in the current study may have resulted from the outcome information being relatively complicated and requiring more deliberate thinking, while the FRN effect would manifest where the outcomes were simple and easy to capture. Finally, the relationship between depressive symptoms and ERP signals with regard to outcome evaluation may be unstable for non-clinical samples. In any case, further investigation is necessary to clarify this issue. Acknowledgments
Fig. 4. Scatter plot of the correlations between individual SDS scores and the P3 amplitude in response to positive CO preceded by negative AO (“AO (−) CO (+)”, represented by red dots) and negative CO preceded by positive AO (“AO (+) CO (−)”, represented by blue dots). The regression lines for both conditions are provided. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
et al., 2003). Thus, we suggest that the current finding may be relevant to the psychological mechanisms of depressive disorders. One might suggest that our major findings could be explained in terms of the occurrence of an unexpected outcome rather than counterfactual thinking. We disagree with this opinion for two reasons. First, the suggestion indicates that depressed participants were more surprised when the valences of AO and CO were different. However, this was unlikely because AO and CO valences were independent of one another during the task (see Section 2). Second, numerous studies suggest that the FRN labels the violation of outcome expectancy (e.g., Bellebaum and Daum, 2008; Nieuwenhuis et al., 2004; Pfabigan et al., 2011), but its amplitude was insensitive to either the AO × CO interaction or levels of depression in this study. Instead, we interpret the P3 findings in light of Osinsky et al. (2014) and suggest that this component was involved in counterfactual comparisons between chosen and unchosen outcomes (see also Liang et al., 2015). Most previous studies have concentrated on the impact of depression on upward counterfactual thinking, inferring that depressive symptoms might be associated with stronger feelings of regret (Monroe et al., 2005). However, the current study indicates that levels of depression are related to both upward and downward counterfactual thinking. That is, people who suffer from depressive symptoms are more likely to think about “what would have happened” regardless of the valence of current outcomes. In our opinion, this phenomenon may have its root in abnormalities of neural systems. Coricelli et al. (2005) have linked the process of counterfactual comparison to a number of brain regions, including the medial temporal cortex (see also Coricelli et al., 2007). It is possible that a hyperactive medial temporal cortex could result in depressive individuals' tendency to counterfactual thinking. The current study has not included EEG source localization owing to its unreliable spatial accuracy (Zhukov et al., 2000). However, many brain-imaging studies have found stronger medial temporal activation in major depressive disorder (Beauregard et al., 1998; Kumari et al., 2003), which may support our idea. Finally, the inconsistency between the current findings and previous studies needs to be addressed. A series of studies conducted by Proudfit and colleagues have revealed that the FRN, rather than the P3, was associated with depression, such that the FRN amplitude was inversely related to depression scores (e.g., Bress et al., 2013; Foti and Hajcak, 2009; Foti et al., 2011; for reviews, see Proudfit, 2015; Proudfit et al., in
This work was supported by the National Natural Science Foundation of China (31300847, 31300867, 30930031, 81471376, 91132704), the National Basic Research Program of China (973 Program, 2014CB744600), the foundation of the National Key Laboratory of Human Factors Engineering (HF2012-K-03), and Global Research Initiative Program, National Institutes of Health, United States of America (1R01TW007897). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors thank Yi Luo for help with manuscript preparation. References Bar, M., 2009. A cognitive neuroscience hypothesis of mood and depression. Trends Cogn. Sci. 13, 456–463. Beauregard, M., Leroux, J.M., Bergman, S., Arzoumanian, Y., Beaudoin, G., Bourgouin, P., Stip, E., 1998. The functional neuroanatomy of major depression: an fMRI study using an emotional activation paradigm. Neuroreport 9, 3253–3258. Beevers, C.G., Disner, S.G., Haigh, E.A.P., Beck, A.T., 2011. Neural mechanisms of the cognitive model of depression. Nat. Rev. Neurosci. 12, 467–477. Bellebaum, C., Daum, I., 2008. Learning-related changes in reward expectancy are reflected in the feedback-related negativity. Eur. J. Neurosci. 27, 1823–1835. Bress, J.N., Foti, D., Kotov, R., Klein, D.N., Hajcak, G., 2013. Blunted neural response to rewards prospectively predicts depression in adolescent girls. Psychophysiology 50, 74–81. Chase, H.W., Camille, N., Michael, A., Bullmore, E.T., Robbins, T.W., Sahakian, B.J., 2010. Regret and the negative evaluation of decision outcomes in major depression. Cogn. Affect. Behav. Neurosci. 10, 406–413. Coricelli, G., Critchley, H.D., Joffily, M., O'Doherty, J.P., Sirigu, A., Dolan, R.J., 2005. Regret and its avoidance: a neuroimaging study of choice behavior. Nat. Neurosci. 8, 1255–1262. Coricelli, G., Dolan, R.J., Sirigu, A., 2007. Brain, emotion and decision making: the paradigmatic example of regret. Trends Cogn. Sci. 11, 258–265. Crowley, M.J., Wu, J., Crutcher, C., Bailey, C.A., Lejuez, C.W., Mayes, L.C., 2009. Risk-taking and the feedback negativity response to loss among at-risk adolescents. Dev. Neurosci. Basel 31, 137–148. Eryilmaz, H., Van De Ville, D., Schwartz, S., Vuilleumier, P., 2014. Lasting impact of regret and gratification on resting brain activity and its relation to depressive traits. J. Neurosci. 34, 7825–7835. Foti, D., Hajcak, G., 2009. Depression and reduced sensitivity to non-rewards versus rewards: evidence from event-related potentials. Biol. Psychol. 81, 1–8. Foti, D., Kotov, R., Klein, D.N., Hajcak, G., 2011. Abnormal neural sensitivity to monetary gains versus losses among adolescents at risk for depression. J. Abnorm. Child Psychol. 39, 913–924. Galinsky, A.D., Seiden, V.L., Kim, P.H., Medvec, V.H., 2002. The dissatisfaction of having your first offer accepted: the role of counterfactual thinking in negotiations. Personal. Soc. Psychol. Bull. 28, 271–283. Gehring, W.J., Willoughby, A.R., 2002. The medial frontal cortex and the rapid processing of monetary gains and losses. Science 295, 2279–2282. Giorgetta, C., Grecucci, A., Bonini, N., Coricelli, G., Demarchi, G., Braun, C., Sanfey, A.G., 2012. Waves of regret: a meg study of emotion and decision-making. Neuropsychologia 51, 38–51. Gotlib, I.H., Joormann, J., 2010. Cognition and depression: current status and future directions. Annu. Rev. Clin. Psychol. 6, 285–312. Gu, R., Lei, Z., Broster, L., Wu, T., Jiang, Y., Luo, Y.J., 2011a. Beyond valence and magnitude: a flexible evaluative coding system in the brain. Neuropsychologia 49, 3891–3897. Gu, R., Wu, T., Jiang, Y., Luo, Y.J., 2011b. Woulda, coulda, shoulda: the evaluation and the impact of the alternative outcome. Psychophysiology 48, 1354–1360.
98
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
Hajcak, G., Holroyd, C.B., Moser, J.S., Simons, R.F., 2005. Brain potentials associated with expected and unexpected good and bad outcomes. Psychophysiology 42, 161–170. Hajcak, G., Moser, J.S., Holroyd, C.B., Simons, R.F., 2007. It's worse than you thought: the feedback negativity and violations of reward prediction in gambling tasks. Psychophysiology 44, 905–912. Krishnan, V., Nestler, E.J., 2008. The molecular neurobiology of depression. Nature 455, 894–902. Kumari, V., Mitterschiffthaler, M.T., Teasdale, J.D., Malhi, G.S., Brown, R.G., Giampietro, V., Brammer, M.J., Poon, L., Simmons, A., Williams, S.C., Checkley, S.A., Sharma, T., 2003. Neural abnormalities during cognitive generation of affect in treatment-resistant depression. Biol. Psychiatry 54, 777–791. Landman, J., Vandewater, E.A., Stewart, A.J., Malley, J.E., 1995. Missed opportunities: psychological ramifications of counterfactual thought in midlife women. J. Adult Dev. 2, 87–97. Lepore, S.J., Silver, R.C., Wortman, C.B., Wayment, H.A., 1996. Social constraints, intrusive thoughts, and depressive symptoms among bereaved mothers. J. Pers. Soc. Psychol. 70, 271–282. Liang, J., Lin, H., Xiang, J., Wu, H., Li, X., Liang, H., Zheng, X., 2015. Counterfactual comparison modulates fairness consideration in the mini-ultimatum game: an event-related potentials study. Scand. J. Psychol. 56, 124–131. Mandel, D.R., 2003. Counterfactuals, emotions, and context. Cogn. Emot. 17, 139–159. Markman, K.D., Miller, A.K., 2006. Depression, control, and counterfactual thinking: functional for whom? J. Soc. Clin. Psychol. 25, 210–227. Markman, K.D., Gavanski, I., Sherman, S.J., McMullen, M.N., 1995. The impact of perceived control on the imagination of better and worse possible worlds. Personal. Soc. Psychol. Bull. 21, 588–595. Medvec, V.H., Savitsky, K., 1997. When doing better means feeling worse: the effects of categorical cutoff points on counterfactual thinking and satisfaction. J. Pers. Soc. Psychol. 72, 1284–1296. Meyers, B.S., English, J., Gabriele, M., Peasley-Miklus, C., Heo, M., Flint, A.J., Mulsant, B.H., Rothschild, A.J., 2006. A delusion assessment scale for psychotic major depression: reliability, validity, and utility. Biol. Psychiatry 60, 1336–1342. Miller, D.T., Taylor, B.R., 1995. Counterfactual thought, regret, and superstition: how to avoid kicking yourself. In: Roese, N.J., Olson, J.M. (Eds.), What Might Have Been: The Social Psychology of Counterfactual Thinking. Erlbaum, Mahwah, NJ, pp. 305–332. Monroe, M.R., Skowronski, J.J., MacDonald, W., Wood, S.E., 2005. The mildly depressed experience more post-decisional regret than the non-depressed. J. Soc. Clin. Psychol. 24, 665–690. Mueller, E.M., Pechtel, P., Cohen, A.L., Douglas, S.R., Pizzagalli, D.A., 2015. Potentiated processing of negative feedback in depression is attenuated by anhedonia. Depress. Anxiety 32, 296–305. Nieuwenhuis, S., Holroyd, C.B., Mol, N., Coles, M.G.H., 2004. Reinforcement-related brain potentials from medial frontal cortex: origins and functional significance. Neurosci. Biobehav. Rev. 28, 441–448. Osinsky, R., Walter, H., Hewig, J., 2014. What is and what could have been: an ERP study on counterfactual comparisons. Psychophysiology 51, 773–781. Pfabigan, D.M., Alexopoulos, J., Bauer, H., Sailer, U., 2011. Manipulation of feedback expectancy and valence induces negative and positive reward prediction error signals manifest in event-related brain potentials. Psychophysiology 48, 656–664.
Philiastides, M.G., Biele, G., Vavatzanidis, N., Kazzer, P., Heekeren, H.R., 2010. Temporal dynamics of prediction error processing during reward-based decision making. NeuroImage 53, 221–232. Polich, J., 2007. Updating P300: an integrative theory of P3a and P3b. Clin. Neurophysiol. 118, 2128–2148. Polich, J., Criado, J.R., 2006. Neuropsychology and neuropharmacology of P3a and P3b. Int. J. Psychophysiol. 60, 172–185. Proudfit, G.H., 2015. The reward positivity: from basic research on reward to a biomarker for depression. Psychophysiology 52, 449–459. Proudfit, G.H., Bress, J.N., Foti, D., Kujawa, A., Klein, D.N., 2014. Depression and event-related potentials: emotional disengagement and reward insensitivity. Curr. Opin. Psychol. http://dx.doi.org/10.1016/j.copsyc.2014.12.018 (in press). Quelhas, A.C., Power, M.J., Juhos, C., Senos, J., 2008. Counterfactual thinking and functional differences in depression. Clin. Psychol. Psychother. 15, 352–365. Ridout, N., Astell, A., Reid, I., Glen, T., O'Carroll, R., 2003. Memory bias for emotional facial expressions in major depression. Cogn. Emot. 17, 101–122. Roschke, J., Wagner, P., 2003. A confirmatory study on the mechanisms behind reduced P300 waves in depression. Neuropsychopharmacol.: Off. Publ. Am. Coll. Neuropsychopharmacol. 28 (Suppl. 1), S9–S12. Semlitsch, H.V., Anderer, P., Schuster, P., Presslich, O., 1986. A solution for reliable and valid reduction of ocular artifacts, applied to the P300 ERP. Psychophysiology 23, 695–703. Shek, D.T., 1993. The Chinese version of the state-trait anxiety inventory: its relationship to different measures of psychological well-being. J. Clin. Psychol. 49, 349–358. Shu, L., 1993. Self-rating depression scale and depression status inventory. Chin. J. Ment. Health 7, 160–162 (Supplement). Spielberger, C.D., Gorsuch, R.L., Lushene, R., Vagg, P.R., Jacobs, G.A., 1983. Manual for the State-Trait Anxiety Inventory. Consulting Psychologist Press, Palo Alto, CA. Stavrakaki, C., Vargo, B., 1986. The relationship of anxiety and depression — a review of the literature. Br. J. Psychiatr. 149, 7–16. Wu, Y., Clark, L., 2014. Disappointment and regret enhance corrugator reactivity in a gambling task. Psychophysiology 52, 518–523. Wu, Y., Zhou, X.L., 2009. The P300 and reward valence, magnitude, and expectancy in outcome evaluation. Brain Res. 1286, 114–122. Yeung, N., Sanfey, A.G., 2004. Independent coding of reward magnitude and valence in the human brain. J. Neurosci. 24, 6258–6264. Zeelenberg, M., van Dijk, W.W., van der Pligt, J., Manstead, A.S.R., van Empelen, P., Reinderman, D., 1998. Emotional reactions to the outcomes of decisions: the role of counterfactual thought in the experience of regret and disappointment. Organ. Behav. Hum. Decis. Process. 75, 117–141. Zhou, L.P., Kong, L.M., 2009. Experimental study of characteristics of counterfactual thinking in depressive individuals. Health Med. Res. Pract. 6, 19–24. Zhukov, L., Weinstein, D., Johnson, C., 2000. Independent component analysis for EEG source localization. IEEE Eng. Med. Biol. Mag. 19, 87–96. Zung, W.W.K., Richards, C.B., Short, M.J., 1965. Self-rating depression scale in an outpatient clinic — further validation of SDS. Arch. Gen. Psychiatry 13, 508–515.
Contents lists available at ScienceDirect
International Journal of Psychophysiology journal homepage: www.elsevier.com/locate/ijpsycho
Depressive states amplify both upward and downward counterfactual thinking Xue Feng a, Ruolei Gu b,⁎, Fucheng Liang c, Lucas S. Broster d, Yunzhe Liu a, Dandan Zhang e, Yue-jia Luo e,f a
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing 100875, China Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing 100101, China Academic Administration Section, Tianjin Normal University, Tianjin 300387, China d Department of Behavioral Science, University of Kentucky College of Medicine, Lexington, KY 40506, USA e Institute of Affective and Social Neuroscience, Shenzhen University, Shenzhen, 518060, China f Sichuan Research Center of Applied Psychology, Chengdu Medical College, Chengdu 610500, China b c
a r t i c l e
i n f o
Article history: Received 12 December 2014 Received in revised form 6 April 2015 Accepted 13 April 2015 Available online 30 April 2015 Keywords: Depression Counterfactual thinking Outcome evaluation Event-related potential P3
a b s t r a c t Depression has been linked to counterfactual thinking in many behavioral studies, but the direction of this effect remains disputed. In the current study, the relationship between depression and counterfactual thinking was examined using the event-related potential (ERP) technique. In a binary choice gambling task, outcome feedback of the chosen option and that of the alternative option were both provided, so as to elicit the process of counterfactual comparison. By investigating ERP signals in response to outcome presentation, we discovered that when the fictive outcome was better or worse than the factual outcome, the amplitude of the P3 component was positively correlated with individual levels of depression, but not levels of anxiety. These results indicate that depression strengthens both upward counterfactual thinking and downward counterfactual thinking. The implication of this finding to clinical research is discussed. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Depression is a state of low mood associated with multiple psychophysiological symptoms including stress, anhedonia, and abnormalities in appetite and sleep (Krishnan and Nestler, 2008). High levels of depression give rise to biased thoughts and beliefs, such as negative self-image and self-focused attention (Beevers et al., 2011). This effect may increase the risk for the development and recurrence of depressive disorder (Gotlib and Joormann, 2010). Therefore, investigating altered belief systems in depression has important implications for pathophysiological research. The current study explores the relationship between depression and counterfactual thinking, which refers to the process of comparing reality and “what might have been” had a different decision been made (Coricelli et al., 2007). Comparisons based on alternatives that improve on reality (i.e., upward counterfactual) generate feelings of regret and disappointment (Mandel, 2003; Zeelenberg et al., 1998), while comparisons based on alternatives that worsen reality (i.e., downward counterfactual) generate feelings of rejoice and
⁎ Corresponding author. E-mail address: [email protected] (R. Gu).
http://dx.doi.org/10.1016/j.ijpsycho.2015.04.016 0167-8760/© 2015 Elsevier B.V. All rights reserved.
gratification (Galinsky et al., 2002; Medvec and Savitsky, 1997). The influence of depression on counterfactual thinking has been examined by many studies, but the precise nature of this influence is still debated. Markman and Miller (2006) reported that participants suffering from severe depressive symptoms generated counterfactuals that were less controllable (see also Eryilmaz et al., 2014; Monroe et al., 2005; Zhou and Kong, 2009). In contrast, depressed patients showed attenuated counterfactual thinking in Chase et al. (2010) (see also Quelhas et al., 2008). Finally, some studies have found no correlation between levels of depression and counterfactual thinking (e.g., Landman et al., 1995). In our opinion, these heterogeneous findings may be attributable in part to methodological issues that distinguish some or all of those studies. First, most previous studies devoted to this topic have relied on selfreport instrument rather than clandestine tools such as measures of brain activity (see Eryilmaz et al., 2014, for an exception). The limitations of these types of assessments are worth noting, including response bias and socially desirable responding (Crowley et al., 2009). Second, many studies estimate counterfactual thinking by asking retrospective questions (e.g., Markman and Miller, 2006), which may be vulnerable to the memory bias associated with depression (Bar, 2009). Finally, while most studies focus on the potential relation between depression and upward counterfactual which elicits feelings of regret, whether
94
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
depression affects downward counterfactual remains largely unanswered. Eryilmaz et al. (2014) recently reported that activities of the brain regions associated with gratification (i.e., a consequence of downward counterfactual thinking) were negatively correlated with Beck Depression Inventory (BDI) scores during resting state. This study investigated the impact of depression on counterfactual thinking in a decision-making scenario (e.g., Chase et al., 2010) with event-related potentials (ERPs) as a neural measurement. In this task, participants chose between two options and received positive or negative outcomes of the chosen option and the alternative option (Gu et al., 2011b; Yeung and Sanfey, 2004). Thus, the comparison between positive alternative outcome (AO for short) and negative chosen outcome (CO for short) elicits upward counterfactual thinking, while the comparison between negative AO and positive CO elicits downward counterfactual thinking (Giorgetta et al., 2012). For instance, when a participant discovers that the option he/she selected led to monetary loss (CO) while the other option could have resulted in monetary gain (AO), he/she may feel regret and imagine the preferable alternate selection (upward counterfactual thinking). Two ERP components, the feedback-related negativity (FRN) and the P3, were chosen as they are the major biomarkers of outcome evaluation (Gehring and Willoughby, 2002; Yeung and Sanfey, 2004). Both components have been reported to be affected by depression levels (Proudfit, 2015; Proudfit et al., in press; Roschke and Wagner, 2003). Numerous studies have confirmed that both the FRN and the P3 are sensitive to the comparison between subjective expectation and reality (Hajcak et al., 2005, 2007; Wu and Zhou, 2009). In order to clarify the discrete roles played by these components in counterfactual contexts, Osinsky et al. (2014) recently proposed that the FRN indicates a binary categorization of gain versus no-gain for both CO and AO, while the P3 indicates motivational salience derived from counterfactual comparisons of CO and AO. Consistent with this idea, Liang et al. (2015) reveal that counterfactual comparisons manifest on the P3 rather than the FRN (see also Gu et al., 2011b; Wu and Clark, 2014). Accordingly, we suggest that the P3 elicited by the comparison between CO and AO reflects the process of counterfactual thinking when other major factors that could modulate this component (e.g., event probability, stimulus novelty, and task complexity; for reviews, see Polich, 2007; Polich and Criado, 2006) are controlled. We predicted that P3 amplitude in response to the comparison between CO and AO would increase as a function of individual differences in depressive traits, indicating a stronger tendency of counterfactual thinking among depressed individuals. In contrast, such associations would be absent for FRN amplitude. 2. Methods 2.1. Participants 20 right-handed Chinese students (11 females; mean age 22.75 years, SD = 1.68) from three universities in Tianjin participated in the experiment. All the participants were included in the final sample. All had normal or corrected-to-normal vision. None of the participants had sought medication for emotional problems before. All the participants denied regular use of any substance with the potential to affect the central nervous system. None had a history of neurological disease. None of the participants reported knowledge of the Cyrillic alphabet. All participants gave their informed consent prior to the experiment. The study was approved by the ethics committee of Academy of Psychology and Behavior, Tianjin Normal University.
electrically shielded room. Each trial began with the presentation of two options represented as Cyrillic letters (“д” and “ю” respectively), each of which was presented inside a white rectangle (2.51° × 2.51° of visual angle) appearing on either side of a fixation point. The position of these letters was counterbalanced across trials. Participants selected one option by pressing the “F” or “J” keys on the keyboard with left or right index finger, respectively. The selected rectangle was outlined in red for 500 ms. After that, the AO was presented in the unselected rectangle for 1000 ms simultaneously with the disappearance of the two letters. Then the AO faded away, leaving the rectangles and the fixation point on screen for 500 ms. Finally, the CO was presented in the chosen rectangle for 1000 ms. The formal task consisted of two blocks of 100 trials each. Stimulus display and behavioral data acquisition were conducted using E-Prime software 1.1 (PST, Inc., Pittsburgh, PA). There were two kinds of outcome valence: positive (“+”) and negative (“−”). A positive and a negative CO indicated that participants gained or lost one point in the current trial, respectively (Fig. 1). Unbeknownst to the participant, no matter which option was chosen, the probability of receiving a positive or negative CO or AO in each trial was 50%. Moreover, the valences of AO and CO were independent of one another. Thus, we suggest that event probability, stimulus novelty, and task complexity have been controlled across conditions. After participants finished the task, he or she was instructed to fill the Chinese version of Zung's self-rating depression scale (SDS) and the trait form of Spielberger's State-Trait anxiety inventory (STAI-T), both of which have demonstrated good internal consistency reliability, convergent validity, and discriminate validity (STAI-T: Shek, 1993; Spielberger et al., 1983. SDS: Shu, 1993; Zung et al., 1965). The STAI-T was included in this study in appreciation of the strong relationship between anxiety and depression (Stavrakaki and Vargo, 1986). During debriefing, participants were informed that the outcomes were predetermined and there was no optimal strategy for the task. Accordingly, each participant was paid 30 Chinese Yuan (approximately 5 US dollars) for participation. 2.3. Electrophysiological recording and preprocessing The EEG activity was recorded from 64 scalp sites using tin electrodes mounted in an elastic cap (NeuroScan, Inc., Herndon, VA), with an online reference to the left mastoid and off-line algebraic rereference to the average of the left and right mastoids. Horizontal electrooculogram (EOG) was recorded from electrodes placed at the outer canthi of both eyes. Vertical EOG was recorded from electrodes placed above and below the left eye. All interelectrode impedance was maintained at b5 k Ω. EEG and EOG signals were amplified with a 0.05–100 Hz bandpass filter and continuously sampled at 500 Hz/channel. During the off-line analysis, ocular artifacts were removed from the EEG signal using a regression procedure implemented in the Neuroscan software (Semlitsch et al., 1986). The EEG was digitally filtered through a zero phase shift, with 0.05–30 Hz band-pass filtering for the FRN and 2 Hz low-pass filtering for the P3 component (Gu et al., 2011a; Yeung and Sanfey, 2004). The data were then averaged into 1000 ms epochs (200 ms baseline) time-locked to outcome presentation. Any trials where EEG voltages exceeded a threshold of ± 100 μV during the recording epoch were excluded from the analysis. After the artifact rejection processing, the numbers of trials survived in each condition were as follows: AO (+) CO (+): 45.63 of the 50 trials (91.26%); AO (−) CO (+): 44.84 (89.68%); AO (+) CO (−): 45.21 (90.42%); and AO (−) CO (−): 46.05 (92.10%).
2.2. Procedure 2.4. Data analysis The task procedure replicated that of Gu et al. (2011b). Before the experiment, participants were informed about the rules and framework of the task and encouraged to respond so as to maximize the total score. Participants sat approximately 100 cm from a computer screen in an
The time windows for peak identification were determined through visual inspection of grand-averaged waveforms. The electrode at which the ERP components reached their maximum was detected along the
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
95
Fig. 1. The schematic diagrams of two example trials. RT = response time. (a) A negative alternative outcome (AO) and a positive chosen outcome (CO) produce a downward counterfactual scenario. (b) A positive AO and a negative CO produce an upward counterfactual scenario.
midline (Fz, FCz, Cz, CPz, Pz, POz and Oz); thereafter, the arithmetic mean of peak amplitudes of this electrode and 8 adjacent electrodes was calculated for further analysis (Gu et al., 2011a). For all the analyses listed below, the significance level was set at 0.05. Post-hoc testing of significant main effects was conducted using the LSD method. Significant interactions were analyzed using simpleeffects models. Statistical analysis was performed using SPSS 19.0 (IBM, Armonk, NY).
CPz (M = 11.14 μV, SD = 1.08 μV). Accordingly, the means of this electrode and 8 adjacent electrodes (C1, Cz, C2, CP1, CP2, P1, Pz, and P2) were entered into further analysis. The main effect of AO valence was not significant (t(19) = 0.998, p = .331) (Fig. 2c). No significant correlation was detected between the P3 amplitude and personality scores (ps N .05).
3. Results 3.1. Emotional scales SDS score was 40.61 (SD = 7.70, range = 30–58) and STAI-T score was 40.50 (SD = 9.28, range = 24–62), respectively. The correlation between SDS and STAT-T was significant (r = .742, p b .001). 3.2. Behavioral results Participants selected “д” and “ю” 50.07% (SD = 12.44) and 49.45% (SD = 12.42) of the time, respectively. The averaged reaction time for decision-making was 0.90 s (SD = 0.94). None of these behavioral measures were correlated with SDS or STAI-T scores (ps N .05). Seeing that there was no optimal strategy in the task, the behavioral data were not analyzed further. 3.3. ERP results 3.3.1. Alternative outcome (AO, presented antecedently) The ERP responses elicited by AO were entered into pairedsamples t tests using AO valence (positive/negative) as the withinsubject factor. 3.3.1.1. The FRN component. The FRN was measured as the most negative peak within the 200–400 ms time window. It was largest at electrode FCz (M = 2.80 μV, SD = 0.47 μV). Accordingly, the means of this electrode and 8 adjacent electrodes (F1, Fz, F2, FC1, FC2, C1, Cz and C2) were entered into further analysis. The main effect of AO valence approached significance (t(19) = 1.856, p = .079), with the FRN showed a tendency to be greater in response to positive AO than to negative AO (3.24 μV vs. 4.58 μV) (Fig. 2a). No significant correlation was detected between the FRN amplitude and personality scores in each condition (ps N .05). 3.3.1.2. The P3 component. The P3 was measured as the most positive peak within the 300–500 ms time window. It was largest at electrode
Fig. 2. (a) Grand-average ERPs elicited by AO with a 0.05–30 Hz band-pass filter at the FCz recording site, where the FRN reached its maximum. (b) Grand-average ERPs elicited by CO at the CPz site, where the P3 reached its maximum, with a 0.05–30 Hz band-pass filter and (c) with a 2 Hz low-pass filter. The time point “0” indicates the AO onset. The time windows for peak detection of the FRN (a) and P3 (b, c) were indicated as light gray shaded areas. The tic markers indicate the timepoints in which the P3 peaks were detected.
96
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
3.3.2. Chosen outcome (presented subsequently) The ERP responses elicited by CO were entered into 2 (CO valence: positive/negative) × 2 (AO valence: positive/negative) ANOVAs. 3.3.2.1. The FRN component. The FRN was measured as the most negative peak within the 200–400 ms time window. It was largest at electrode FCz (M = 4.95 μV, SD = 0.75 μV). Accordingly, the means of this electrode and 8 adjacent electrodes (F1, Fz, F2, FC1, FC2, C1, Cz, and C2) were entered into further analysis. The main effect of CO valence was significant (F(1, 19) = 29.705, p b .001); the FRN was greater in response to negative CO than positive CO (3.35 μV vs. 7.38 μV). The main effect of AO valence was also significant (F(1, 19) = 14.893, p b .005); the FRN was greater when following negative AO than positive AO (4.39 μV vs. 6.34 μV). The AO × CO interaction was insignificant (F(1, 19) = 0.893, p = .356) (Fig. 3a). No significant correlation was detected between the FRN amplitude and personality scores (ps N .05). 3.3.2.2. The P3 component. The P3 was measured as the most positive peak within the 300–500 ms time window. It was largest at electrode CPz (M = 17.13 μV, SD = 1.37 μV). Accordingly, the mean amplitudes of this electrode and 8 adjacent electrodes (C1, Cz, C2, CP1, CP2, P1, Pz, and P2) were entered into ANOVA analysis. The main effect of CO valence approached significance (F(1, 19) = 3.795, p =
.066); positive CO showed a tendency to elicit a larger P3 compared to negative CO (11.76 μV vs. 10.92 μV). The main effect of AO valence was significant (F(1, 19) = 23.429, p b .01); the CO elicited a larger P3 when it was preceded by positive AO than negative AO (12.44 μV vs. 10.24 μV). The AO × CO interaction was also significant (F(1, 19) = 21.561, p b .01); the influence of AO on P3 amplitude was significant (F(1, 19) = 37.711, p b 0.001) when CO was positive, but was insignificant when CO was negative (F(1, 19) = 0.536, p = .473) (Fig. 3c). The means and standard deviations of the P3 amplitude and latency in each condition are presented in Table 1. Pearson correlation analysis revealed significant correlations between SDS scores and P3 amplitude when positive CO followed negative AO (r = 0.457, p b .05) and when negative CO followed positive AO (r = 0.462, p b .05). In brief, participant's level of depression was correlated with the P3 component when the valences of AO and CO were different (Fig. 4). No significant correlation was detected between the P3 amplitude and STAI-T scores. To assess the unique contributions of depression, a linear regression was performed (see Foti and Hajcak, 2009, for similar methods). When entering both SDS and STAI-T scores to predict the P3 amplitude elicited by positive CO following negative AO, the regression coefficient for depression was significant (β = 0.377, p b .05) but anxiety was not (β = −0.168, p N .05); the same was true in the condition of negative CO following positive AO, such that depression (β = 0.462, p b .05), but not anxiety (β = 0.290, p N .05), was a significant predictor. 4. Discussion Previous studies have argued about whether and how depression is associated with counterfactual thinking (e.g., Chase et al., 2010; Markman and Miller, 2006). Using the ERP technique, the current study investigated this issue in a trial-by-trial decision-making task. The ERP results reveal that individual depressive level was positively correlated with the amplitude of the P3 component elicited by chosen outcome presentation in two conditions. Specifically, the correlations were significant when a positive CO appeared after a negative AO (i.e., the real outcome was better than the alternative) and when a negative CO appeared after a positive AO (i.e., the real outcome was worse than the alternative). As indicated in the Section 1, these two conditions represent situations that motivate downward and upward counterfactual thinking, respectively. According to Osinsky et al. (2014), the P3 but not the FRN is involved in counterfactual comparisons (see also Liang et al., 2015). This idea is confirmed in the current study, such that the P3 rather than the FRN showed a significant interaction between AO and CO. We suggest that the current ERP findings indicate that depressive participants were more likely to be engaged in counterfactual thinking than their non-depressive counterparts. Presumably, this phenomenon is linked to the observation that depressed individuals are more susceptible to irrelevant thoughts and have more difficulty “getting over it” and refocusing on upcoming challenges (Lepore et al., 1996; Monroe et al., 2005). In addition, extravagant counterfactual thinking accounts for overgeneral memory bias and imagination (Markman et al., 1995; Miller and Taylor, 1995), which contributes to severe depressive illness (Meyers et al., 2006; Ridout
Table 1 The P3 amplitude and latency in each condition. The values in this table are arithmetic means of 9 electrodes (C1, Cz, C2, CP1, CPz, CP2, P1, Pz, and P2).
Fig. 3. (a) Grand-average ERPs elicited by CO with a 0.05–30 Hz band-pass filter at the FCz recording site, where the FRN reached its maximum. (b) Grand-average ERPs elicited by CO at the CPz site, where the P3 reached its maximum, with a 0.05–30 Hz band-pass filter and (c) with a 2 Hz low-pass filter. The time point “0” indicates the CO onset. The time windows for peak detection of the FRN (a) and P3 (b, c) were indicated as light gray shaded areas. The tic markers indicate the timepoints in which the P3 peaks were detected.
Experimental conditions
P3 amplitude mean (SD)
P3 latency mean (SD)
AO (+) AO (−) AO (+) CO (+) AO (−) CO (+) AO (+) CO (−) AO (−) CO (−)
7.56 μV (3.01) 7.18 μV (3.56) 14.20 μV (5.13) 9.32 μV (3.85) 10.67 μV (4.02) 11.17 μV (4.88)
398.31 ms (49.88) 392.29 ms (67.53) 400.29 ms (41.53) 353.74 ms (58.18) 383.92 ms (49.43) 438.48 ms (49.86)
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
97
press; but see Mueller et al., in press), which contradicts our current findings. We suspect that this inconsistency occurred because in those studies, participants were only aware of the outcome associated with their decisions; thus, they did not engage in counterfactual thinking (Proudfit, 2015). As an alternative explanation, previous studies focusing on outcome evaluation have suggested that the FRN and P3 represent quick detection and deliberate evaluation of the current event, respectively (Gu et al., 2011a; Philiastides et al., 2010; Wu and Zhou, 2009). Therefore, the preeminence of the P3 in the current study may have resulted from the outcome information being relatively complicated and requiring more deliberate thinking, while the FRN effect would manifest where the outcomes were simple and easy to capture. Finally, the relationship between depressive symptoms and ERP signals with regard to outcome evaluation may be unstable for non-clinical samples. In any case, further investigation is necessary to clarify this issue. Acknowledgments
Fig. 4. Scatter plot of the correlations between individual SDS scores and the P3 amplitude in response to positive CO preceded by negative AO (“AO (−) CO (+)”, represented by red dots) and negative CO preceded by positive AO (“AO (+) CO (−)”, represented by blue dots). The regression lines for both conditions are provided. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
et al., 2003). Thus, we suggest that the current finding may be relevant to the psychological mechanisms of depressive disorders. One might suggest that our major findings could be explained in terms of the occurrence of an unexpected outcome rather than counterfactual thinking. We disagree with this opinion for two reasons. First, the suggestion indicates that depressed participants were more surprised when the valences of AO and CO were different. However, this was unlikely because AO and CO valences were independent of one another during the task (see Section 2). Second, numerous studies suggest that the FRN labels the violation of outcome expectancy (e.g., Bellebaum and Daum, 2008; Nieuwenhuis et al., 2004; Pfabigan et al., 2011), but its amplitude was insensitive to either the AO × CO interaction or levels of depression in this study. Instead, we interpret the P3 findings in light of Osinsky et al. (2014) and suggest that this component was involved in counterfactual comparisons between chosen and unchosen outcomes (see also Liang et al., 2015). Most previous studies have concentrated on the impact of depression on upward counterfactual thinking, inferring that depressive symptoms might be associated with stronger feelings of regret (Monroe et al., 2005). However, the current study indicates that levels of depression are related to both upward and downward counterfactual thinking. That is, people who suffer from depressive symptoms are more likely to think about “what would have happened” regardless of the valence of current outcomes. In our opinion, this phenomenon may have its root in abnormalities of neural systems. Coricelli et al. (2005) have linked the process of counterfactual comparison to a number of brain regions, including the medial temporal cortex (see also Coricelli et al., 2007). It is possible that a hyperactive medial temporal cortex could result in depressive individuals' tendency to counterfactual thinking. The current study has not included EEG source localization owing to its unreliable spatial accuracy (Zhukov et al., 2000). However, many brain-imaging studies have found stronger medial temporal activation in major depressive disorder (Beauregard et al., 1998; Kumari et al., 2003), which may support our idea. Finally, the inconsistency between the current findings and previous studies needs to be addressed. A series of studies conducted by Proudfit and colleagues have revealed that the FRN, rather than the P3, was associated with depression, such that the FRN amplitude was inversely related to depression scores (e.g., Bress et al., 2013; Foti and Hajcak, 2009; Foti et al., 2011; for reviews, see Proudfit, 2015; Proudfit et al., in
This work was supported by the National Natural Science Foundation of China (31300847, 31300867, 30930031, 81471376, 91132704), the National Basic Research Program of China (973 Program, 2014CB744600), the foundation of the National Key Laboratory of Human Factors Engineering (HF2012-K-03), and Global Research Initiative Program, National Institutes of Health, United States of America (1R01TW007897). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors thank Yi Luo for help with manuscript preparation. References Bar, M., 2009. A cognitive neuroscience hypothesis of mood and depression. Trends Cogn. Sci. 13, 456–463. Beauregard, M., Leroux, J.M., Bergman, S., Arzoumanian, Y., Beaudoin, G., Bourgouin, P., Stip, E., 1998. The functional neuroanatomy of major depression: an fMRI study using an emotional activation paradigm. Neuroreport 9, 3253–3258. Beevers, C.G., Disner, S.G., Haigh, E.A.P., Beck, A.T., 2011. Neural mechanisms of the cognitive model of depression. Nat. Rev. Neurosci. 12, 467–477. Bellebaum, C., Daum, I., 2008. Learning-related changes in reward expectancy are reflected in the feedback-related negativity. Eur. J. Neurosci. 27, 1823–1835. Bress, J.N., Foti, D., Kotov, R., Klein, D.N., Hajcak, G., 2013. Blunted neural response to rewards prospectively predicts depression in adolescent girls. Psychophysiology 50, 74–81. Chase, H.W., Camille, N., Michael, A., Bullmore, E.T., Robbins, T.W., Sahakian, B.J., 2010. Regret and the negative evaluation of decision outcomes in major depression. Cogn. Affect. Behav. Neurosci. 10, 406–413. Coricelli, G., Critchley, H.D., Joffily, M., O'Doherty, J.P., Sirigu, A., Dolan, R.J., 2005. Regret and its avoidance: a neuroimaging study of choice behavior. Nat. Neurosci. 8, 1255–1262. Coricelli, G., Dolan, R.J., Sirigu, A., 2007. Brain, emotion and decision making: the paradigmatic example of regret. Trends Cogn. Sci. 11, 258–265. Crowley, M.J., Wu, J., Crutcher, C., Bailey, C.A., Lejuez, C.W., Mayes, L.C., 2009. Risk-taking and the feedback negativity response to loss among at-risk adolescents. Dev. Neurosci. Basel 31, 137–148. Eryilmaz, H., Van De Ville, D., Schwartz, S., Vuilleumier, P., 2014. Lasting impact of regret and gratification on resting brain activity and its relation to depressive traits. J. Neurosci. 34, 7825–7835. Foti, D., Hajcak, G., 2009. Depression and reduced sensitivity to non-rewards versus rewards: evidence from event-related potentials. Biol. Psychol. 81, 1–8. Foti, D., Kotov, R., Klein, D.N., Hajcak, G., 2011. Abnormal neural sensitivity to monetary gains versus losses among adolescents at risk for depression. J. Abnorm. Child Psychol. 39, 913–924. Galinsky, A.D., Seiden, V.L., Kim, P.H., Medvec, V.H., 2002. The dissatisfaction of having your first offer accepted: the role of counterfactual thinking in negotiations. Personal. Soc. Psychol. Bull. 28, 271–283. Gehring, W.J., Willoughby, A.R., 2002. The medial frontal cortex and the rapid processing of monetary gains and losses. Science 295, 2279–2282. Giorgetta, C., Grecucci, A., Bonini, N., Coricelli, G., Demarchi, G., Braun, C., Sanfey, A.G., 2012. Waves of regret: a meg study of emotion and decision-making. Neuropsychologia 51, 38–51. Gotlib, I.H., Joormann, J., 2010. Cognition and depression: current status and future directions. Annu. Rev. Clin. Psychol. 6, 285–312. Gu, R., Lei, Z., Broster, L., Wu, T., Jiang, Y., Luo, Y.J., 2011a. Beyond valence and magnitude: a flexible evaluative coding system in the brain. Neuropsychologia 49, 3891–3897. Gu, R., Wu, T., Jiang, Y., Luo, Y.J., 2011b. Woulda, coulda, shoulda: the evaluation and the impact of the alternative outcome. Psychophysiology 48, 1354–1360.
98
X. Feng et al. / International Journal of Psychophysiology 97 (2015) 93–98
Hajcak, G., Holroyd, C.B., Moser, J.S., Simons, R.F., 2005. Brain potentials associated with expected and unexpected good and bad outcomes. Psychophysiology 42, 161–170. Hajcak, G., Moser, J.S., Holroyd, C.B., Simons, R.F., 2007. It's worse than you thought: the feedback negativity and violations of reward prediction in gambling tasks. Psychophysiology 44, 905–912. Krishnan, V., Nestler, E.J., 2008. The molecular neurobiology of depression. Nature 455, 894–902. Kumari, V., Mitterschiffthaler, M.T., Teasdale, J.D., Malhi, G.S., Brown, R.G., Giampietro, V., Brammer, M.J., Poon, L., Simmons, A., Williams, S.C., Checkley, S.A., Sharma, T., 2003. Neural abnormalities during cognitive generation of affect in treatment-resistant depression. Biol. Psychiatry 54, 777–791. Landman, J., Vandewater, E.A., Stewart, A.J., Malley, J.E., 1995. Missed opportunities: psychological ramifications of counterfactual thought in midlife women. J. Adult Dev. 2, 87–97. Lepore, S.J., Silver, R.C., Wortman, C.B., Wayment, H.A., 1996. Social constraints, intrusive thoughts, and depressive symptoms among bereaved mothers. J. Pers. Soc. Psychol. 70, 271–282. Liang, J., Lin, H., Xiang, J., Wu, H., Li, X., Liang, H., Zheng, X., 2015. Counterfactual comparison modulates fairness consideration in the mini-ultimatum game: an event-related potentials study. Scand. J. Psychol. 56, 124–131. Mandel, D.R., 2003. Counterfactuals, emotions, and context. Cogn. Emot. 17, 139–159. Markman, K.D., Miller, A.K., 2006. Depression, control, and counterfactual thinking: functional for whom? J. Soc. Clin. Psychol. 25, 210–227. Markman, K.D., Gavanski, I., Sherman, S.J., McMullen, M.N., 1995. The impact of perceived control on the imagination of better and worse possible worlds. Personal. Soc. Psychol. Bull. 21, 588–595. Medvec, V.H., Savitsky, K., 1997. When doing better means feeling worse: the effects of categorical cutoff points on counterfactual thinking and satisfaction. J. Pers. Soc. Psychol. 72, 1284–1296. Meyers, B.S., English, J., Gabriele, M., Peasley-Miklus, C., Heo, M., Flint, A.J., Mulsant, B.H., Rothschild, A.J., 2006. A delusion assessment scale for psychotic major depression: reliability, validity, and utility. Biol. Psychiatry 60, 1336–1342. Miller, D.T., Taylor, B.R., 1995. Counterfactual thought, regret, and superstition: how to avoid kicking yourself. In: Roese, N.J., Olson, J.M. (Eds.), What Might Have Been: The Social Psychology of Counterfactual Thinking. Erlbaum, Mahwah, NJ, pp. 305–332. Monroe, M.R., Skowronski, J.J., MacDonald, W., Wood, S.E., 2005. The mildly depressed experience more post-decisional regret than the non-depressed. J. Soc. Clin. Psychol. 24, 665–690. Mueller, E.M., Pechtel, P., Cohen, A.L., Douglas, S.R., Pizzagalli, D.A., 2015. Potentiated processing of negative feedback in depression is attenuated by anhedonia. Depress. Anxiety 32, 296–305. Nieuwenhuis, S., Holroyd, C.B., Mol, N., Coles, M.G.H., 2004. Reinforcement-related brain potentials from medial frontal cortex: origins and functional significance. Neurosci. Biobehav. Rev. 28, 441–448. Osinsky, R., Walter, H., Hewig, J., 2014. What is and what could have been: an ERP study on counterfactual comparisons. Psychophysiology 51, 773–781. Pfabigan, D.M., Alexopoulos, J., Bauer, H., Sailer, U., 2011. Manipulation of feedback expectancy and valence induces negative and positive reward prediction error signals manifest in event-related brain potentials. Psychophysiology 48, 656–664.
Philiastides, M.G., Biele, G., Vavatzanidis, N., Kazzer, P., Heekeren, H.R., 2010. Temporal dynamics of prediction error processing during reward-based decision making. NeuroImage 53, 221–232. Polich, J., 2007. Updating P300: an integrative theory of P3a and P3b. Clin. Neurophysiol. 118, 2128–2148. Polich, J., Criado, J.R., 2006. Neuropsychology and neuropharmacology of P3a and P3b. Int. J. Psychophysiol. 60, 172–185. Proudfit, G.H., 2015. The reward positivity: from basic research on reward to a biomarker for depression. Psychophysiology 52, 449–459. Proudfit, G.H., Bress, J.N., Foti, D., Kujawa, A., Klein, D.N., 2014. Depression and event-related potentials: emotional disengagement and reward insensitivity. Curr. Opin. Psychol. http://dx.doi.org/10.1016/j.copsyc.2014.12.018 (in press). Quelhas, A.C., Power, M.J., Juhos, C., Senos, J., 2008. Counterfactual thinking and functional differences in depression. Clin. Psychol. Psychother. 15, 352–365. Ridout, N., Astell, A., Reid, I., Glen, T., O'Carroll, R., 2003. Memory bias for emotional facial expressions in major depression. Cogn. Emot. 17, 101–122. Roschke, J., Wagner, P., 2003. A confirmatory study on the mechanisms behind reduced P300 waves in depression. Neuropsychopharmacol.: Off. Publ. Am. Coll. Neuropsychopharmacol. 28 (Suppl. 1), S9–S12. Semlitsch, H.V., Anderer, P., Schuster, P., Presslich, O., 1986. A solution for reliable and valid reduction of ocular artifacts, applied to the P300 ERP. Psychophysiology 23, 695–703. Shek, D.T., 1993. The Chinese version of the state-trait anxiety inventory: its relationship to different measures of psychological well-being. J. Clin. Psychol. 49, 349–358. Shu, L., 1993. Self-rating depression scale and depression status inventory. Chin. J. Ment. Health 7, 160–162 (Supplement). Spielberger, C.D., Gorsuch, R.L., Lushene, R., Vagg, P.R., Jacobs, G.A., 1983. Manual for the State-Trait Anxiety Inventory. Consulting Psychologist Press, Palo Alto, CA. Stavrakaki, C., Vargo, B., 1986. The relationship of anxiety and depression — a review of the literature. Br. J. Psychiatr. 149, 7–16. Wu, Y., Clark, L., 2014. Disappointment and regret enhance corrugator reactivity in a gambling task. Psychophysiology 52, 518–523. Wu, Y., Zhou, X.L., 2009. The P300 and reward valence, magnitude, and expectancy in outcome evaluation. Brain Res. 1286, 114–122. Yeung, N., Sanfey, A.G., 2004. Independent coding of reward magnitude and valence in the human brain. J. Neurosci. 24, 6258–6264. Zeelenberg, M., van Dijk, W.W., van der Pligt, J., Manstead, A.S.R., van Empelen, P., Reinderman, D., 1998. Emotional reactions to the outcomes of decisions: the role of counterfactual thought in the experience of regret and disappointment. Organ. Behav. Hum. Decis. Process. 75, 117–141. Zhou, L.P., Kong, L.M., 2009. Experimental study of characteristics of counterfactual thinking in depressive individuals. Health Med. Res. Pract. 6, 19–24. Zhukov, L., Weinstein, D., Johnson, C., 2000. Independent component analysis for EEG source localization. IEEE Eng. Med. Biol. Mag. 19, 87–96. Zung, W.W.K., Richards, C.B., Short, M.J., 1965. Self-rating depression scale in an outpatient clinic — further validation of SDS. Arch. Gen. Psychiatry 13, 508–515.
