Cooperative context is a determinant of the social influence on outcome evaluation: An electrophysiological study

International Journal of Psychophysiology 100 (2016) 28–35

Contents lists available at ScienceDirect

International Journal of Psychophysiology journal homepage: www.elsevier.com/locate/ijpsycho

Cooperative context is a determinant of the social influence on outcome evaluation: An electrophysiological study Kenta Kimura a,b,⁎, Jun'ichi Katayama a,c a b c

Center for Applied Psychological Science, Kwansei Gakuin University, Nishinomiya 662-8501, Japan Physiological Function Research Team, Automotive Human Factors Research Center, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba 305-8566, Japan Department of Psychological Science, Kwansei Gakuin University, Nishinomiya 662-8501, Japan

a r t i c l e

i n f o

Article history: Received 22 July 2015 Received in revised form 18 November 2015 Accepted 15 December 2015 Available online 23 December 2015 Keywords: FRN Social context Outcome evaluation Cooperative context Feedback processing

a b s t r a c t The present study examined whether or not a cooperative context is a determinant of the social influence on the evaluation of two action outcomes: a monetary outcome and a conflict of opinion with other group members. In the present study, three-person groups were randomly assigned to be either a cooperative or individual group and asked to perform a gambling task. The monetary outcomes in the cooperative group were interrelated among group members, whereas those in the individual group did not influence each other. The present results showed that monetary outcomes elicited feedback-related negativity (FRN) and a conflict of opinion with other group members elicited FRN-like negativity, which reflect an evaluation of the motivational significance of action outcomes. The FRN elicited by monetary outcomes was reduced when participants shared decisions with other group members only in the cooperative group, indicating that the cooperative context reduced the motivational significance of monetary outcomes through the diffusion of responsibility. The FRN-like negativity elicited by a conflict of opinion showed a different pattern between the cooperative and individual groups, indicating that the cooperative context can influence the evaluation of a conflict of opinion, possibly via the modulation of group cohesiveness or conflict processing. The present results suggest that a cooperative context, rather than the social setting, is a determinant of the social influence on outcome evaluation. © 2015 Elsevier B.V. All rights reserved.

1. Introduction 1.1. Outcome evaluation reflected in feedback-related negativity In everyday life, the accurate evaluation of action outcomes is fundamental for making successful and efficient behavioral adjustments. However, the evaluation of outcomes is not always accurate and can be influenced by many situational factors. Recent studies have indicated that the social context, such as the mere presence of others, and cooperating or competing with other people, dramatically changes the evaluation of action outcomes (e.g., Itagaki and Katayama, 2008; Koban et al., 2012; Leng and Zhou, 2010). Event-related brain potentials (ERP) have been used to elucidate the electrophysiological correlates of outcome evaluation in action monitoring. Several ERP studies have reported that a monetary loss or a negative performance feedback, compared to a monetary gain or a positive performance feedback, elicited a negative-going ERP component called feedback-related negativity (FRN) at around 250–300 ms after onset of the action outcome (e.g., Gehring and Willoughby, 2002; ⁎ Corresponding author at: Center for Applied Psychological Science, Kwansei Gakuin University, Nishinomiya 662-8501, Japan. E-mail address: [email protected] (K. Kimura).

http://dx.doi.org/10.1016/j.ijpsycho.2015.12.005 0167-8760/© 2015 Elsevier B.V. All rights reserved.

Miltner et al., 1997). To exclude the potential effects of other ERP components such as the P300, FRN has been typically evaluated with a difference wave approach (e.g., Holroyd and Krigolson, 2007), in which a difference wave is created by subtracting ERPs elicited by positive feedback (e.g., monetary gain) from ERPs elicited by negative feedback (e.g., monetary loss). FRN has a fronto-central distribution and likely reflects neural activity generated in the anterior cingulate cortex (ACC; Gehring and Willoughby, 2002; Miltner et al., 1997; Ridderinkhof et al., 2004). A series of studies have suggested that the processes that generate FRN are involved in quickly evaluating the motivational significance of ongoing events (e.g., Boksem et al., 2011; Gehring and Willoughby, 2002; Yeung et al., 2005). Another line of research has suggested a reinforcement learning model of the FRN, which states that the FRN reflects negative prediction error or salience prediction error signals conveyed from the mid-brain dopamine system to the ACC (e.g., Holroyd and Coles, 2002; Pfabigan et al., 2011; Talmi et al., 2013). 1.2. Outcome evaluations in a social context Several studies have reported that a cooperative context could modulate the outcome evaluation reflected in the FRN (for a review, see Koban and Pourtois, 2014). A cooperative context is defined as a

K. Kimura, J. Katayama / International Journal of Psychophysiology 100 (2016) 28–35

situation where a pair or a group of participants is rewarded based on their joint performance. Li et al. (2010) asked participants to perform a gambling task either individually or in a three-person group. The FRN elicited by monetary outcomes was smaller when participants performed the task in a three-person group. This suggests that sharing actions with other group members can reduce the motivational significance of action outcomes. Their results can be explained by the diffusion of responsibility, which is a well-known social phenomenon whereby a person is less likely to feel personal responsibility for action when he/she works with others (e.g., Latané and Darley, 1968). Accordingly, in this paper, we define the diffusion of responsibility as a reduced sense of personal responsibility through the sharing of actions or decisions with others. Our previous study examined outcome evaluation in a group decision-making task in which a three-person group chose one of two cards based on majority rule and received monetary gains and losses according to this group decision (Kimura and Katayama, 2013). This study revealed that the FRN elicited by monetary outcomes was reduced when group members chose the same card, indicating that making the same decision as others reduced the personal responsibility for the outcomes associated with the group decision. These previous findings suggest that the outcome evaluation reflected in the FRN is sensitive to the modulation of personal responsibility. In social situations, a conflict of opinion with others can elicit FRNlike negativity, which has been shown to have a morphology and latency similar to those for FRN elicited by monetary outcomes and performance feedback (e.g., Kimura and Katayama, 2013). The presence of FRN-like negativity has been reported when the opinions of participants differed from those of a normative group (Shestakova et al., 2013) and from those of the other group members (Chen et al., 2012; Kimura and Katayama, 2013; Kimura et al., 2013). These results suggest that a conflict of opinion is an important action outcome and is detected by the generic action-monitoring system. Since FRN-like negativity increased as the number of group members who made different decisions increased (Chen et al., 2012), this negativity, like the FRN elicited by monetary outcomes, is thought to reflect the evaluation of motivational significance of a conflict of opinion.

29

and the participants' motor actions were exactly the same as those in the cooperative group, but the participant's monetary outcomes were not interrelated. In this gambling task, a participant's choice was classified into one of three trial types: unanimous, majority, or minority. In unanimous trials, the three group members choose the same card; in majority trials, a participant made the same choice as another group member; and in minority trials, the participant's choice was different from those of the other group members. If a cooperative context is of crucial importance for the reduced motivational significance of monetary outcomes due to a diffusion of responsibility, the FRN elicited by monetary outcomes in the unanimous and majority trials should be reduced in the cooperative group, but not in the individual group. Further, if a cooperative context influences the evaluation of a conflict of opinion with other group members, the FRN-like negativity elicited by a conflict of opinion should show different patterns between the cooperative and individual groups. 2. Method 2.1. Participants Participants (N = 42) were recruited from Kwansei Gakuin University and divided into 14 gender-matched three-person groups. The groups were randomly assigned to cooperative (1 male and 6 female groups; range = 18–23 years, M = 19.71 years) or individual (1 male and 6 female groups; range = 18–23 years, M = 19.29 years) conditions. All participants were right handed, had normal or corrected to normal vision, and did not have a history of neurological or mental diseases. The Kwansei Gakuin University (KGU) Research Ethics Review Board approved the study under the KGU Regulations for Research with Human Participants. Written informed consent was obtained from all the participants. Participants were paid 1000 Japanese yen (approximately US $9) and a bonus based on their performance. Data from two participants in both groups had to be excluded from the ERP analysis because of excessive movement artifacts. Thus, the electrophysiological data are reported for the remaining 19 participants in each group.

1.3. Present study 2.2. Experimental task and procedure The present study focused on the effect of a cooperative context on the FRN elicited by the monetary outcome and the FRN-like negativity elicited by the presence of a conflict of opinion. An important question is whether or not a cooperative context is a determinant of the effect of a diffusion of responsibility on outcome evaluation. Previous studies have shown that performing a gambling task in a three-person group (Li et al., 2010) and having the same decision as others (Kimura and Katayama, 2013) decreased the sense of personal responsibility for the action outcome and hence the FRN. However, these studies did not distinguish between the effects of a cooperative context and those of the social setting; i.e., the mere presence of other group members, joint actions, concurrent engagement in tasks, and so forth. Therefore, it remained unclear whether the effects of a cooperative context on the FRN are due to the cooperative context itself or are merely due to the social setting. The purpose of the present study was to examine whether or not a cooperative context is a determinant of the social influence on the outcome evaluation. To this end, we needed to compare a cooperative context condition with a valid control condition in which the group members' action outcomes would not influence each other, while the physical situation and the participants' motor actions were the same. In the present study, three-person groups performed a gambling task in which the participants were presented with the decisions and outcomes of other group members. Participants were randomly assigned to a cooperative or an individual group. In the cooperative group, participants were rewarded based on their joint performance in the gambling task. In the individual group, the physical conditions

The participants engaged in a gambling task in which participants in a three-person group individually chose one of two cards and received a monetary gain or loss in each trial. The presentation of visual stimuli and recording of the participants' responses were controlled with Presentation software (Neurobehavioral Systems). All visual stimuli were presented via a projector (Sight3D, Solidray) onto the center of a screen (2.4 m × 1.8 m; Kikuchi Science Laboratory) located approximately 2 m in front of the participants. Fig. 1a shows a schematic diagram of the present task. In the beginning of each trial, the names of the three participants and two white rectangles (approximately 25 cm × 46 cm) with a thin black border were displayed. Participants were instructed that these two rectangles were cards indicating small monetary gains and losses. After 500– 1500 ms, the fixation cross turned red and the instruction, “Choice!!”, was presented above the cards. During this period, each participant was asked to choose the left or right card by pressing a left button with the left thumb or a right button with the right thumb, respectively. Once all of the participants had chosen a card, the color of the fixation cross returned to black and the instruction above the cards disappeared. After 750 ms, black circles (approximately 8 cm) superimposed on the corresponding card adjacent to the participants' names indicated the choice of each participant (others' choice onset). After 1000 ms, the black circles disappeared, and the two cards and the fixation cross remained for 1000 ms. Both cards were then covered with translucent red or blue rectangles and the text “+ 10 yen” or “− 10 yen” was shown above them (monetary outcome onset). Thus, both the color

30

K. Kimura, J. Katayama / International Journal of Psychophysiology 100 (2016) 28–35

Fig. 1. (a) Timeline of trials in the gambling task. Individuals in three-person groups were asked to choose either card at Card Choice. At Others' Choice, their choices were made public. Individual monetary outcomes were then presented. ERPs were calculated from the onsets of Others' Choice and Monetary Outcome. (b) The three trial types in the present gambling task. Examples of each trial type for a participant framed by a gray box are illustrated.

and the text indicated whether each participant gained or lost 10 Japanese yen (about $0.09) in that trial. The color of the monetary gain and loss was counterbalanced across groups. After 1000 ms, only the fixation cross was presented for 500–1500 ms (inter-trial interval). Before the experiment, the three participants were individually invited to the laboratory. The participants in each group were strangers to each other. After the experimenter briefly introduced the three participants to each other and gave them a brief description of the experiment, they were asked to sit next to each other in front of the screen, on chairs placed 1 m apart. Electrodes were then attached and the participants were given detailed task instructions. The instructions for the gambling task differed between the cooperative and individual groups. In the cooperative condition, participants were told that the monetary outcomes for the three participants would be combined and the sum of the accumulated amount would be divided equally and paid to each participant at the end of the experiment. In the individual condition, participants were told that the accumulated money for each participant would be added to, or subtracted from, their individual bonus awards. They were also instructed not to communicate with each other during the experimental task. After the instructions, participants were given a practice block of 10 trials to familiarize themselves with the task. The experimental task consisted of 10 blocks of 30 trials (300 total trials). At the end of each block, the accumulated sum of money earned in that block was displayed on the screen. 2.3. Data acquisition EEG data were simultaneously recorded from the three participants in each group by using three BrainAmp EEG systems (Brain Products, BrainAmp). An Easycap with Ag–AgCl electrodes was used to record from five electrode locations: Frontal (Fz), Fronto-central (FCz), Central (Cz), Centro-parietal (CPz), and Parietal (Pz), according to the extended 10–20 system. Although the number of electrodes was minimal, previous studies have indicated that midline electrodes are sufficient to identify the FRN component (Hajcak et al., 2006; Kimura and Katayama, 2013). These midline electrodes and two electrodes placed on both earlobes were referenced to the nose tip, and electrode impedances were kept below 5 kΩ. Blinks and horizontal eye movements were monitored by electrooculograms (EOG) recorded from two electrodes placed approximately 1 cm above and below the right eye and two electrodes

placed in the outer canthi of both eyes. The Afz electrode served as a ground site. The EEG and EOG signals were filtered through a bandpass of 0.05–100 Hz and digitized at a sampling rate of 1000 Hz. 2.4. Data analysis Brain Vision Analyzer software (Brain Products) was used to analyze the stored EEG data. First, the data were digitally filtered offline through a 20-Hz phase-shift-free, Butterworth low-pass filter. The method developed by Gratton et al. (1983) was applied to correct for EOG artifacts. The data were then re-referenced to the average activity of the earlobe electrodes. In the subsequent EEG analysis, the EEGLAB toolbox (Delorme and Makeig, 2004; available at http://sccn.ucsd.edu/eeglab/) was used in Matlab. The EEG data were segmented into epochs of 1200 ms (including a 200 ms pre-stimulus baseline) for others' choice onset and monetary outcomes onset. Epochs in which signals exceeded an absolute value of 75 μV on any of the electrodes were excluded from averaging; this resulted in the exclusion of about 4% of trials. The present study classified a participant's choice in each trial into one of the three trial types: unanimous, majority, or minority (see Fig. 1b). In unanimous trials, the three group members choose the same card; in majority trials, a participant made the same choice as another group member; and in minority trials, a participant's choice was different from those of the other group members. Accordingly, the EEG epochs for others' choice onset were averaged for the six conditions defined by two groups (Cooperative and Individual) and three trial types (Unanimous, Majority, and Minority). The EEG epochs for monetary outcomes were averaged for the 12 conditions defined by two groups (Cooperative and Individual), three trial types (Unanimous, Majority, and Minority), and two monetary outcomes (Gain and Loss). To quantify the FRN-like negativity, the mean amplitude was calculated for a time window between 250–300 ms following the presentation of others' choices. This time window was chosen because the FRN-like negativity observed in the grand-averaged ERPs showed a peak at around 280 ms. Since the FRN-like negativity, consistent with our previous study (Kimura and Katayama, 2013), showed the most negative amplitude at the Fz electrode site for all three trial types in both groups, we focused our statistical analysis on the amplitude at the Fz electrode. In terms of the FRN associated with monetary

K. Kimura, J. Katayama / International Journal of Psychophysiology 100 (2016) 28–35

31

outcomes, the mean amplitude of the gain and loss ERPs was calculated for a time window between 250–300 ms. In addition, to exclude the potential effects of other ERP components such as P300, we used the difference wave approach to isolate the FRN (e.g., Holroyd and Krigolson, 2007). We created difference waves by subtracting gain ERPs from loss ERPs for each group and each trial type. The mean amplitudes of each difference wave were calculated for a time window between 250–300 ms. This time window was chosen because the negative deflection observed in the difference wave showed a peak at around 280 ms. The mean amplitudes in the grand-averaged difference wave were of comparable magnitude at front-central electrode sites (Fz, FCz, and Cz) for all three trial types in both groups. For reasons of simplicity and consistency with our previous study (Kimura and Katayama, 2013), our statistical analysis was focused on the amplitude at the Fz electrode site, where it was maximal. The mean amplitudes of ERPs following the presentation of others' choices and the loss-minus-gain difference waves were subjected to a mixed-model ANOVA with the 2 groups (Cooperative, Individual) as between-participant factors and the 3 trial types (Unanimous, Majority, and Minority) as within-participant factors. The mean amplitudes of gain ERPs and loss ERPs were subjected to a mixed-model analysis of variance (ANOVA) with the 2 groups (Cooperative, Individual) as between-participant factors, the 3 trial types (Unanimous, Majority, and Minority) as within-participant factors, and the 2 monetary outcomes (Gain, Loss) as within-participant factors. When appropriate, corrected p values are presented in the results, along with the uncorrected degrees of freedom and the Greenhouse–Geisser epsilon (ε). Post hoc comparisons were made by using Shaffer's modified sequentially rejective multiple test procedure, which extends Bonferroni t tests in a stepwise fashion. The significance level was set at .05 for all statistical tests. 3. Results 3.1. Behavioral data Table 1 shows the mean percentages of monetary gain and loss for the three trial types in the cooperative and individual groups. The participants in the cooperative group gained 18.10 yen on average, whereas those in the individual group gained 8.57 yen. The best and worst cumulative monetary outcomes ranged from + 260 to − 280 yen in the cooperative group and from + 240 to − 360 yen in the individual group. There was no difference in the cumulative monetary gain of each participant between the cooperative and individual groups, t (40) = 0.19, p N .1. In addition, there was no difference in the percentages of the trial types between groups, χ2 (5) = 1.19, p N .1.

Fig. 2. (a) Grand-averaged ERPs elicited by Others' Choice for each condition in cooperative (left) and individual (right) groups at Fz, Cz, and Pz. (b) Grand-average mean amplitudes of the FRN (250–300 ms at Fz) for the others' choice of the three trial types in the two groups. Error bars indicate standard errors of the mean.

3.2. Electrophysiological data 3.2.1. ERPs associated with others' choices Fig. 2a shows the grand-averaged ERPs elicited by the presentation of others' choices for the three trial types in the cooperative and individual groups at the Fz, Cz, and Pz electrode sites. Similar to previous studies (Chen et al., 2012; Kimura and Katayama, 2013), the waveforms showed FRN-like negativity and a subsequent P300 component. Fig. 2b shows the mean amplitudes of the FRN-like negativity for each trial type in the two groups at the Fz electrode site. Table 1 Mean percentage (and SD) of each trial type in cooperative and individual groups.

Unanimous Majority Minority

Cooperative group

Individual group

Gain

Loss

Gain

Loss

15.02 (2.28) 23.66 (2.43) 11.61 (1.50)

14.65 (2.48) 23.22 (2.05) 11.83 (2.11)

14.71 (3.10) 23.52 (2.93) 11.90 (2.58)

14.29 (2.05) 23.81 (3.10) 11.76 (2.16)

A mixed-model ANOVA (2 groups × 3 trial types) on the mean amplitudes at the Fz electrode site revealed a significant interaction between group and trial type, F (2, 72) = 3.78, p b .05, MSE = 2.81, ε = .98. Post hoc tests revealed that the mean amplitudes in the individual group were significantly more negative in majority and minority trials than in unanimous trials, t (18) = 2.66, p b .05, Cohen's d = .63 and t (18) = 3.36, p b .05, Cohen's d = .79, respectively. The mean amplitudes of the two trial types were not different. The cooperative group showed a different pattern; the mean amplitude in majority trials was more negative than those in unanimous and minority trials, t (18) = 3.16, p b .05, Cohen's d = .75, and t (18) = 3.32, p b .05, Cohen's d = .78, respectively, whereas the amplitudes did not differ between unanimous and minority trials. 3.2.2. ERPs associated with monetary outcomes Figs. 3 and 4 show the grand-averaged ERPs elicited by monetary outcomes for the three trial types in cooperative and individual groups

32

K. Kimura, J. Katayama / International Journal of Psychophysiology 100 (2016) 28–35

Fig. 3. Grand-averaged ERPs elicited by monetary outcomes and difference waves (Loss-minus-Gain) for each condition in the cooperative group at Fz, Cz, and Pz.

at the Fz, Cz, and Pz electrode sites. These figures also show the difference waves that were created by subtracting gain ERPs from loss ERPs for each trial type. As shown in the difference waves, monetary loss

was associated with a negative-going deflection that peaked at around 280 ms. Fig. 5 presents the mean amplitudes of the gain and loss ERPs for each trial type in the two groups at the Fz electrode site. Fig. 6

Fig. 4. Grand-averaged ERPs elicited by monetary outcomes and difference waves (Loss-minus-Gain) for each condition in the individual group at Fz, Cz, and Pz.

K. Kimura, J. Katayama / International Journal of Psychophysiology 100 (2016) 28–35

33

amplitude in unanimous trials showed a tendency for a group difference, in that the individual group showed a more negative amplitude than the cooperative group (p = .06). There were no significant group differences in majority or minority trials. 4. Discussion

Fig. 5. Grand-average mean amplitudes of the gain and loss ERPs (250–300 ms at Fz) for the three trial types in the two groups. Error bars indicate standard errors of the mean.

illustrates the mean amplitudes of the difference waves for each trial type in the two groups at the Fz electrode site. We conducted a mixed-model ANOVA (2 groups × 3 trial types × 2 monetary outcomes) on the mean amplitudes of the gain and loss ERPs within the 250–300 ms time window. This revealed a significant threeway interaction, F (2, 72) = 3.40, p b .05, MSE = 5.17, ε = .87. A subsequent two-way ANOVA (3 trial types × 2 monetary outcomes) was conducted in both groups separately. In the cooperative group, a two-way ANOVA revealed a significant interaction, F (2, 36) = 5.29, p b .05, MSE = 5.70, ε = .82. Post hoc tests showed that the mean amplitudes of the gain ERPs differed among the trial types; the mean amplitude of the gain ERPs in minority trials was more positive than those in unanimous and majority trials, t (18) = 3.95, p b .01, Cohen's d = .93 and t (18) = 5.18, p b .01, Cohen's d = 1.22, respectively. The mean amplitude of the loss ERPs did not differ among the trial types. In the individual group, a two-way ANOVA revealed significant main effects of trial type, F (1, 36) = 23.44, p b .01, MSE = 7.55, and monetary outcome, F (1, 36) = 24.41, p b .01, MSE = 28.49. Post hoc tests revealed that the mean amplitudes of the gain and loss ERPs were more positive in minority trials than in unanimous and majority trials, t (18) = 4.92, p b .01, Cohen's d = 1.16 and t (18) = 6.82, p b .01, Cohen's d = 1.61, respectively. A mixed-model ANOVA (2 groups × 3 trial types) on the mean amplitude of the loss-minus-gain difference waves at the Fz electrode showed a significant interaction, F (2, 72) = 3.40, p b .05, MSE = 10.34, ε = .87. Post hoc tests revealed that the mean amplitude in the cooperative group differed among the trial types; the mean amplitude in minority trials was more negative than those in unanimous and majority trials, t (18) = 2.62, p b .05, Cohen's d = .62 and t (18) = 2.42, p b .01, Cohen's d = .57, respectively. In the individual group, the mean amplitudes did not differ among the three trial types. The mean

Fig. 6. Grand-average mean amplitudes of the difference waves (250–300 ms at Fz) for the three trial types in the two groups. Error bars indicate standard errors of the mean.

The purpose of the present study was to examine whether or not the cooperative context is a determinant of the social influence on the outcome evaluation. In the cooperative group, the FRN-like negativity elicited by the others' choices in majority trials was larger than those in unanimous and minority trials, whereas in the individual group, the FRN-like negativities in majority and minority trials were larger than those in unanimous trials. In terms of the FRN associated with monetary outcomes, the gain ERPs in the cooperative group showed a more positive amplitude in minority trials than in unanimous and majority trials, whereas there was no difference in the amplitudes of loss ERPs among trial types. In the individual group, the mean amplitudes of the gain and loss ERPs were more positive in minority trials than in unanimous and majority trials. The results regarding the loss-minus-gain difference waves confirmed that the FRN elicited by monetary outcomes was more negative in minority trials than in unanimous and majority trials in the cooperative group, whereas there was no significant difference among trial types in the individual group. 4.1. Effects of the cooperative context on outcome evaluation The present study revealed that, in the cooperative context, the FRN elicited by monetary outcomes was reduced when the participants made the same decisions as other group members, suggesting that the motivational significance of the monetary outcome might be decreased by a diffusion of responsibility. In contrast, monetary outcomes were not interrelated in the individual group, and there was no change in the FRN, indicating that making the same decision as another, by itself, did not influence the FRN. The two groups performed the gambling task under identical physical conditions. The only difference between the groups was whether or not monetary outcomes were interrelated. Therefore, the present results can be interpreted as showing that the cooperative context, rather than the social setting (i.e., the mere presence of other group members, joint actions, concurrent engagement in tasks, and so forth), is a determinant of the effects of a diffusion of responsibility on outcome evaluation. A reinforcement learning model of the FRN suggests that the FRN reflects negative or salience prediction error signals conveyed from the mid-brain dopamine system to the ACC (e.g., Holroyd and Coles, 2002; Pfabigan et al., 2011; Talmi et al., 2013). According to this model, the present results suggest that the cooperative context may modulate the detection of a prediction error. In the present experimental setting, it is possible that the presentation of others' choices influenced the participants' predictions about monetary outcomes. For instance, participants might predict monetary gains when they chose the same card as the other group members (i.e., unanimous trials) and their prediction errors in that trial might be enhanced by the presentation of monetary losses. However, if this was the case, the FRN would differ among trial types in both groups because the decisions of other group members were presented in both groups. Therefore, the most natural interpretation of the present results would be that the cooperative context influences the motivational significance of monetary outcomes. The FRN did not differ between unanimous and majority trials in the cooperative group. This suggests that the number of group members who share the same decisions does not linearly reduce their motivational significance. According to the social impact theory (e.g., Latané, 1981), one of the determinants of social impact is the number of people who are influencing an individual; the impact of others on the individual increases as the number of others increases, although the rate of increase in the impact decreases as new individuals are added. This

34

K. Kimura, J. Katayama / International Journal of Psychophysiology 100 (2016) 28–35

predicts that a diffusion of responsibility does not necessarily reduce the motivational significance of action outcomes in proportion to the number of people who share the same decisions. The present results are consistent with this view. It is possible that the extent to which a diffusion of responsibility reduces the motivational significance of action outcomes can be predicted by the social impact theory. The results of this study can be interpreted as suggesting that the outcome evaluation reflected in the FRN mirrors an individual's success and failure even when the individual performs a task in a cooperative context. In the cooperative group, an individual's monetary loss in minority trials would result in a monetary gain for the group, since the other two group members would receive monetary gains. Nevertheless, the ERPs elicited by monetary losses in minority trials showed typical loss ERPs comparable to those observed in previous studies (e.g., Gehring and Willoughby, 2002). Similarly, even though monetary gains in minority trials resulted in monetary losses for the group, typical gain ERPs were observed. These results indicate that, even when monetary outcomes are interrelated, the evaluation of action outcomes is based on an individual's success or failure (e.g., Itagaki and Katayama, 2008) rather than on the monetary outcomes for the group. The present results in the individual group were clearly inconsistent with the findings reported by Boksem et al. (2011). They reported that the amplitude of the FRN following performance feedback in a time estimation task was greater when the other player received correct feedback, compared to when they received incorrect feedback. In their experiment, correct and incorrect feedback were independent of each other, and therefore, they concluded that social comparison plays an important role in modulating the evaluation of action outcomes. One possible explanation for this discrepancy is the nature of the experimental task. In the time estimation task, incorrect feedback results from poor performance by a participant. On the other hand, the monetary outcomes in the gambling task were beyond an individual's control, since the outcomes were delivered stochastically and independently of individual ability. It is possible that social comparison may have a greater influence under experimental conditions in which action outcomes are indicative of one's ability. Future studies will be needed to examine the relationship between task type and social factors and their effects on outcome evaluation. The present results showed that the effect of the cooperative context on the FRN was caused by the modulation of gain ERPs but not loss ERPs. Traditionally, the FRN has been thought to reflect the activity of the general error-monitoring system (e.g., Miltner et al., 1997). However, recent studies have suggested that the FRN actually reflects the neural activity related to reward-processing rather than error-monitoring processes (e.g., Foti et al., 2011, 2014; Holroyd et al., 2008). This perspective can be supported by accumulated evidence on reward positivity (for a review, see Proudfit, 2015). It has been argued that the negative deflection elicited by negative feedback can be thought of as an N200 that is generally elicited by task-relevant stimuli, whereas the positive deflection elicited by positive feedback (i.e., reward positivity), which reflects reward processing, is superimposed on this negativity (e.g., Holroyd et al., 2008). From this perspective, the results of the present study suggest that sharing one's decision with group members in a cooperative context may specifically reduce the reward value of monetary gains. 4.2. Effects of the cooperative context on the evaluation of a conflict of opinions Consistent with previous studies, we observed FRN-like negativity following the presentation of others' choices (Chen et al., 2012; Kimura and Katayama, 2013; Kimura et al., 2013). The pattern of amplitudes of the FRN-like negativity was different between the cooperative and individual groups; the FRN-like negativity in the individual group was larger in majority and minority trials than in unanimous trials, whereas the FRN-like negativity in the cooperative group was larger

in majority trials than in the other two trial types. The results in the individual group are consistent with those in previous studies (Chen et al., 2012; Kimura and Katayama, 2013). This suggests that decision conflicts with other group members are detected automatically in a social situation, even when monetary outcomes are not interrelated. One possible explanation for the FRN-like negativity in the cooperative group is the development of group cohesion. It is widely known that cohesiveness among group members develops from cooperative work, in which group members try to accomplish shared goals (Lott and Lott, 1965). Since group members in cohesive groups experience less anxiety and tension (Shaw and Shaw, 1962), the cooperative context might reduce the affective impact of a conflict of opinion in minority trials. Another possible explanation for the results is the interrelatedness of monetary outcomes. In the cooperative context, potential monetary outcomes in minority trials could be evaluated as they had little influence on the group monetary outcomes, because whether the group won or lost in the current trial was dependent on the monetary outcomes for the other two members. Therefore, it is possible that the motivational significance of being a minority, which has been thought to be associated with a greater amplitude of the FRN-like negativity (Chen et al., 2012; Kimura and Katayama, 2013), can be reduced in a cooperative group. This suggests that the evaluation of a conflict of opinion can be flexibly modulated by the social context. In addition, another interpretation can be possible. As mentioned above, a recent theory of FRN suggests that the negative deflection elicited by action outcomes can be thought of as an N200 that is generally elicited by task-relevant stimuli (e.g., Holroyd et al., 2008). It has been reported that N200 could be sensitive not only to task-relevant stimuli but also to the conflict between perceptual representations held in working memory (e.g., Jia et al., 2007). From this standpoint, the present results can be interpreted as indicating that the FRN-like negativity detected different sources of decision conflict in the two group; the participants in the individual group might be sensitive to conflicts between one's own and others' decisions, whereas the participants in the cooperative group might be sensitive to decision conflicts between other group members. This suggests that the source of a decision conflict detected by mechanisms underlying the FRN-like negativity may be flexibly modulated based on the current social context. Our previous study observed another pattern of modulation of the FRN-like negativity under a social context in which a three-person group chose one of two cards based on majority rule (Kimura & Katayama, 2013), which might support this possibility. 4.3. Future study Further investigations will be needed to elucidate the effects of the cooperative context on outcome evaluation. The mechanisms that underlie the modulation of outcome evaluation by a diffusion of responsibility will need to be elucidated. Recently, Koban and Pourtois (2014) suggested that the anterior insula integrates the rapid evaluation of action outcomes, which is processed in the dorsal ACC, with information on the social context. According to this suggestion, the present results can be interpreted as indicating that the information regarding the cooperative context processed in the anterior insula is sensitive to the interrelatedness of action outcomes in a group, but not to the social setting, such as the mere presence of and concurrent engagement with others. However, the detailed mechanisms that underlie such integration remain unclear. Future studies using electrophysiological and neuroimaging methods are necessary to investigate the timing of, as well as the brain regions involved in, such integration. Further studies that control for the frequency of events are also necessary. The present study did not control for the frequency of trial types or monetary outcomes, to prevent participants from speculating about the experimental manipulations. However, Walsh and Anderson (2012) suggested that the FRN was greater for improbable outcomes than for probable outcomes. Although there were no significant

K. Kimura, J. Katayama / International Journal of Psychophysiology 100 (2016) 28–35

differences in percentages of trial types and monetary outcomes between the two groups in the present study, the cumulative monetary gain in the cooperative group tended to be higher than that in the individual group. Therefore, future studies are needed to isolate the effect of the cooperative context from the frequency effects of trial types and monetary outcomes. 5. Conclusion In summary, the present study examined whether or not the cooperative context, rather than the social setting, is a critical factor in modulating the evaluation of two action outcomes: monetary outcomes and conflict of opinion. We used a gambling task and recorded ERPs in three-person groups in cooperative or individual contexts. The present study revealed that the FRN associated with monetary outcomes was reduced when participants shared the monetary outcomes with other group members in a cooperative context. In contrast, monetary outcomes were not shared in the individual context, and there was no change in the FRN. Further, this modulation of the FRN in the cooperative group was caused by the modulation of gain ERPs but not by the modulation of loss ERPs. The FRN-like negativity elicited by a conflict of opinion showed different patterns between the cooperative and individual contexts, indicating that the cooperative context can influence the motivational significance of a conflict of opinion, possibly via the modulation of group cohesiveness or the interrelatedness of action outcomes. The present results suggest that the cooperative context, rather than the social setting, is a determinant of the social influence on outcome evaluation. Acknowledgments This study was supported by a Grant-in-Aid for Scientific Research (B) (25285206) from the Japan Society for the Promotion of Science (JSPS) awarded to Jun'ichi Katayama. This study was conducted as part of a project supported by MEXT (The Ministry of Education, Culture, Sports, Science and Technology, JAPAN) — Supported Program for the Strategic Research Foundation at Private Universities (2010–2014; Project number S1001050) to the Center for Applied Psychological Science (CAPS), Kwansei Gakuin University. References Boksem, M.A., Kostermans, E., De Cremer, D., 2011. Failing where others have succeeded: medial frontal negativity tracks failure in a social context. Psychophysiology 48, 973–979. http://dx.doi.org/10.1111/j.1469-8986.2010.01163.x. Chen, J., Wu, Y., Tong, G., Guan, X., Zhou, X., 2012. ERP correlates of social conformity in a line judgment task. BMC Neurosci. 13, 43. http://dx.doi.org/10.1186/1471-2202-13-43. Delorme, A., Makeig, S., 2004. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics. J. Neurosci. Methods 134, 9–21. http://dx.doi.org/10.1016/j.jneumeth. 2003.10.009. Foti, D., Weinberg, A., Dien, J., Hajcak, G., 2011. Event-related potential activity in the basal ganglia differentiates rewards from nonrewards: temporospatial principal components analysis and source localization of the feedback negativity. Hum. Brain Mapp. 32, 2207–2216. http://dx.doi.org/10.1002/hbm.21182. Foti, D., Weinberg, A., Bernat, E.M., Proudfit, G.H., 2014. Anterior cingulate activity to monetary loss and basal ganglia activity to monetary gain uniquely contribute to the feedback negativity. Clin. Neurophysiol. 2457, 514–518. http://dx.doi.org/10. 1016/j.clinph.2014.08.025. Gehring, W.J., Willoughby, A.R., 2002. The medial frontal cortex and the rapid processing of monetary gains and losses. Science 295, 2279–2282. http://dx.doi.org/10.1126/ science.1066893. Gratton, G., Coles, M.G., Donchin, E., 1983. A new method for off-line removal of ocular artifact. Electroencephalogr. Clin. Neurophysiol. 55, 468–484. http://dx.doi.org/10. 1016/0013-4694(83)90135-9.

35

Hajcak, G., Moser, J.S., Holroyd, C.B., Simons, R.F., 2006. The feedback-related negativity reflects the binary evaluation of good versus bad outcomes. Biol. Psychol. 71, 148–154. http://dx.doi.org/10.1016/j.biopsycho.2005.04.001. Holroyd, C.B., Coles, M.G., 2002. The neural basis of human error processing: reinforcement learning, dopamine, and the error-related negativity. Psychol. Rev. 109, 679–709. http://dx.doi.org/10.1037/0033-295X.109.4.679. Holroyd, C.B., Krigolson, O.E., 2007. Reward prediction error signals associated with a modified time estimation task. Psychophysiology 44, 913–917. http://dx.doi.org/10. 1111/j.1469-8986.2007.00561.x. Holroyd, C.B., Pakzad-Vaezi, K.L., Krigolson, O.E., 2008. The feedback correct-related positivity: sensitivity of the event-related brain potential to unexpected positive feedback. Psychophysiology 45, 688–697. http://dx.doi.org/10.1111/j.1469-8986. 2008.00668.x. Itagaki, S., Katayama, J., 2008. Self-relevant criteria determine the evaluation of outcomes induced by others. Neuroreport 19, 383–387. http://dx.doi.org/10.1097/WNR. 0b013e3282f556e8. Jia, S., Li, H., Luo, Y., Chen, A., Wang, B., Zhou, X., 2007. Detecting perceptual conflict by the feedback-related negativity in brain potentials. Neuroreport 18, 1385–1388. http:// dx.doi.org/10.1097/WNR.0b013e3282c48a90. Kimura, K., Katayama, J., 2013. Outcome evaluations in group decision making using the majority rule: an electrophysiological study. Psychophysiology 50, 848–857. http:// dx.doi.org/10.1111/psyp.12068. Kimura, K., Murayama, A., Miura, A., Katayama, J., 2013. Effect of decision confidence on the evaluation of conflicting decisions in a social context. Neurosci. Lett. 556, 176–180. http://dx.doi.org/10.1016/j.neulet.2013.09.020. Koban, L., Pourtois, G., 2014. Brain systems underlying the affective and social monitoring of actions: an integrative review. Neurosci. Biobehav. Rev. 46, 71–84. http://dx.doi. org/10.1016/j.neubiorev.2014.02.014. Koban, L., Pourtois, G., Bediou, B., Vuilleumier, P., 2012. Effects of social context and predictive relevance on action outcome monitoring. 12, 460–478. http://dx.doi.org/ 10.3758/s13415-012-0091-0. Latané, B., 1981. The psychology of social impact. Am. Psychol. 36, 343–356. http://dx.doi. org/10.1037/0003-066X.36.4.343. Latané, B., Darley, J.M., 1968. Group inhibition of bystander intervention in emergencies. J. Pers. Soc. Psychol. 10, 215–221. http://dx.doi.org/10.1037/h0026570. Leng, Y., Zhou, X., 2010. Modulation of the brain activity in outcome evaluation by interpersonal relationship: an ERP study. Neuropsychologia 48, 448–455. http://dx. doi.org/10.1016/j.neuropsychologia.2009.10.002. Li, P., Jia, S., Feng, T., Liu, Q., Suo, T., Li, H., 2010. The influence of the diffusion of responsibility effect on outcome evaluations: electrophysiological evidence from an ERP study. NeuroImage 52, 1727–1733. http://dx.doi.org/10.1016/j.neuroimage. 2010.04.275. Lott, A.J., Lott, B.E., 1965. Group cohesiveness as interpersonal attraction: a review of relationships with antecedent and consequent variables. Psychol. Bull. 64, 259–309. http://dx.doi.org/10.1037/h0022386. Miltner, W.H., Braun, C.H., Coles, M.G., 1997. Event-related brain potentials following incorrect feedback in a time-estimation task: evidence for a “generic” neural system for error detection. J. Cogn. Neurosci. 9, 788–798. http://dx.doi.org/10.1162/jocn. 1997.9.6.788. 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 486, 56–64. http://dx. doi.org/10.1111/j.1469-8986.2010.01136.x. Proudfit, G.H., 2015. The reward positivity: from basic research on reward to a biomarker for depression. Psychophysiology 52, 449–459. http://dx.doi.org/10.1111/psyp. 12370. Ridderinkhof, K.R., Ullsperger, M., Crone, E.A., Nieuwenhuis, S., 2004. The role of the medial frontal cortex in cognitive control. Science 306, 443–447. http://dx.doi.org/ 10.1126/science.1100301. Shaw, M.E., Shaw, L.M., 1962. Some effects of sociometric grouping upon learning in a second grade classroom. J. Soc. Psychol. 57, 453–458. http://dx.doi.org/10.1080/ 00224545.1962.9710941. Shestakova, A., Rieskamp, J., Tugin, S., Ossadtchi, A., Krutitskaya, J., Klucharev, V., 2013. Electrophysiological precursors of social conformity. Soc. Cogn. Affect. Neurosci. 8, 756–763. http://dx.doi.org/10.1093/scan/nss064. Talmi, D., Atkinson, R., El-Deredy, W., 2013. The feedback-related negativity signals salience prediction errors, not reward prediction errors. J. Neurosci. 33, 8264–8269. http://dx.doi.org/10.1523/JNEUROSCI.5695-12.2013. Walsh, M.M., Anderson, J.R., 2012. Learning from experience: event-related potential correlates of reward processing, neural adaptation, and behavioral choice. Neurosci. Biobehav. Rev. 36, 1870–1884. http://dx.doi.org/10.1016/j.neubiorev.2012.05.008. Yeung, N., Holroyd, C.B., Cohen, J.D., 2005. ERP correlates of feedback and reward processing in the presence and absence of response choice. Cereb. Cortex 15, 535–544. http://dx.doi.org/10.1093/cercor/bhh153.