Although committing an error or receiving a negative feedback is generally considered to be an unpleasant event, it is crucial for learning and for adjusting future behavior. Converging evidence implies that FRN (feedback-related negativity), an ERP component generated from the anterior cingulate cortex (ACC), is generally more pronounced for negative outcomes of our own performances. The FRN peaks approximately 200-300 ms after feedback onset, and is related to the learning of information that can guide subsequent behaviors [1]–[4]. As social creatures, humans learn not only from their own experiences but also by observing others' behaviors during social interactions [5]. A growing number of studies have demonstrated that observing another's monetary loss elicits an FRN effect that is similar, both in morphology and scalp distribution, to a loss that is directly experienced, a pattern commonly referred to as oFRN (observational-FRN)[5]–[7]. This phenomenon implies that the neural process underlying “learning by observation” resembles the process underlying “learning by doing”. This similarity allows us to avoid negative results that are potentially dangerous without having to experience them directly.

By manipulating the relationship between an observer and another person, previous studies suggested that the oFRN might be driven by two processes. First, one may evaluate the observer's outcome from an egocentric perspective when one's own benefit is involved. For example, the research in which an observer and a performer were in competition so that the observer lost when the performer won and vice versa, showed that the oFRN in response to wins of the performer was similar to that in response to losses of the observer[8], [9]. Even participating in one's own separate gambling game alongside the observational task could affect feelings about others. Leng and Zhou (2010) explored the differing neural responses to friends and strangers when the observer was engaged in the same gambling game, and failed to find a differentiation of FRN responses between friend and stranger observations This finding was interpreted as shown that participation itself might draw their attention from other's performance to their own benefit, potentially diminishing the expected differences in neural response toward friends versus strangers[10].

Second, an observer may evaluate a performer's outcome empathically and emotionally when it has no consequence for the self. Modulation of the oFRN occurs when the observer's outcomes are unrelated to the performer's, and the existence of a long-term relationship has been found to modulate the magnitude of oFRN when self-involvement is not a factor in the experiment. Ma and colleagues removed the observer's own gambling from the experimental design, so that participants only observed and evaluated the outcomes of friends versus strangers. Once the observers' egocentric focus on their own results was no longer a factor, there was a larger oFRN effect associated with friends' outcomes than with strangers' [11]. It seems like that the greater the self-benefit is involved, the more that egocentrism and cognitive evaluation reduce the emotional response toward others' results. The less self-benefit is involved, the more emotion influences the oFRN.

However, the design that does not include participation in the same gambling game cannot exclude personal involvement as an influence on oFRN. It also could be argued that it is the potential benefits gained from the success of the other person in a long-term relationship that results in the differentiation seen in the oFRN phase. For example, a friend might pay for drinks after winning a game. Therefore,it is unclear whether the stronger response to friends' losses and gains is due to the social context or to this potential personal benefits.

Another possible interpretation of the larger oFRN in response to friends' outcomes is that one may generally attend more to friends than to strangers because of higher familiarity. The purpose of the present study was to investigate the degree to which the oFRN is moderated by affective preference when controlling for potential benefit and familiarity. A variation on the trust game was used to establish a short-term affective preference for two strangers [12]. In the game, one confederate was fair and generous, while the other was unfair and selfish, leading to varying affective preferences on the part of the participant. Then, the two confederates played a gambling game individually while the participant observed. Event-related potentials (ERPs) were recorded during the observation phase. This paradigm allowed us to explore how a learned affective preference influences neural responses toward liked and disliked strangers. It was expected that the learned affective preference would result in a larger oFRN toward the likable player than toward the unlikable one.

Gender differences were also examined in the current study. Although females are better at recognizing emotions and at expressing themselves emotionally, males show greater responses when presented with threatening emotions such as fear, anger and dominance (see a review, [13]). For example, females tend to be sensitive to both positive (happy) and negative (angry) emotions, but males are primarily sensitive to the negative one (i.e., angry) [14]. This pattern might be due to sex differences in hormones, chromosomes, and brain structures [13]. Besides gender differences in the emotion expression and recognition, females often have a stronger empathic responses than males and score higher on self-reported empathy [15]. When accessing others' emotions females showed increased activation in the right inferior frontal cortex while there is no differential activations in males. Females also recruit areas of the human mirror system, which has been claimed to be related to the capacity for empathy, to a higher degree than males during both self assessing and other assessing tasks [16]. Furthermore, from an evolutionary perspective, males might benefit more than females from competition. For example, the copulation frequency of male elephant seals during breeding seasons is related directly to success in male-male competition. Low competitive males that reached maturity are prevented from mating by the highest ranking males[17]. There is also evidence that in humans, an intergroup conflict has profound effects on males in particular compared with females [18], [19].

Considering the fact that females showed stronger emotion sensitivity and empathy responses compared to males, it might be easier for them to distinguish the results of fair players and unfair players, whereas males might be especially more sensitive to unfair players only, and might even use an egocentric evaluation path to process unfair players' results, and present a reversed oFRN pattern. This hypothesis is further supported by Singer and colleges' findings that women showed more empathy toward fair versus unfair players' pain, whereas men's reward region was activated even when they faced unfair players' losses[12]. Thus, we predict that compared to women, men would show stronger sensitivity to unfair players' results.

The study was approved by the Ethics Committee of the department of Psychology at South China Normal University. Written informed consents were obtained from participants, and they had the right to discontinue participation at any time.

Forty participants (20 male and 20 female) aged 19–25 years (Mean  =  22 years, SD  =  2.1 years) were recruited from South China Normal University, China. Five participants (three male and two female) were excluded from ERP data analysis, one because one fell asleep during the experiment and the other four because they correctly identified the other two “participants” as confederates. All participants had normal or corrected-to-normal vision and no history of neurobiological or psychiatric disorders. Participants were paid 30 yuan (about $4.50) for their participation.

Previous researches on the FRN effect under observation conditions have focused on manipulating the relationship between self and others by taking advantage of long-term friendships or by using a benefit-related competition situation. The present study, using trust game and gambling game, extends this work by examining the gender difference in how the short-term social interaction induces affective preference modulation in oFRN. The results suggest a gender-based difference in rapidly processing others' losses and gains. For females, the likable performers' losses elicited larger oFRN than did those of unlikable performers. Although males rated the fair player as more likable, they did not display significantly different responses toward the two performers. The results imply an existence of emotional and empathic explanation of oFRN.

The female having generally sensitive neural response for distinguishing likable from unlikable strangers is concordant with previous studies and implies an empathetic and emotional account of oFRN. It has been reported that female participants categorized a friend's loss as equivalently negative to their own even in a competitive situation[26]. The dual-processes hypothesis suggests that the other person's status as a personal acquaintance might have enhanced females' empathic responses, overriding the egocentric evaluative component for female participants[9]. In addition, empathy might be the cause of observational FRN[9]-[11], [26]. In the previous studies, the size of the FRN effect varied as a function of whether an action was performed by a friend or by a stranger[11]. In another study, females showed a smaller empathic response toward unfair players' pain[12]. In the present study, the greater FRN responses for the likable player in the observation condition, similar to responses induced by feedback information concerning the observer's own results, under the observation condition may be related to a stronger empathic response toward the likable player. When observing the unlikable players' results, the observer's reduced empathic response might result in a non-significant FRN effect. These results might also be supported by Wilson's selection theory in which the units of evolution is group instead of individual, and highly cooperative groups have an evolutionary advantage over poorly cooperative groups [27]. People tend to cooperate with in-group members whereas they compete with out-group members to benefit themselves. From this perspective, the human tribal inclination has evolved to help humans categorize individuals based on their group membership, and treat in-group members benevolently and out-group members malevolently[19]. The likable player in our study might be considered as an in-group member while the unlikable player might be treated as an out group member. Thus, we observed a disassociation between the responses to these two members.

Crucially, the current study revealed a gender effect in which females responded differently toward likable versus unlikable performers, whereas males did not. This inconsistent response toward two strangers suggests that, in females, not only long-term relationships but also the affective link established through short-term social interaction can affect early neural responses to the consequences of actions unrelated to one's own interests. We suggest that this phenomenan is caused by different empathic responses in females from that in males. The gender-based differences in empathy have been widely reported. Using self-report questionnaires, females showed superiority in empathy in the general population and even among persons with Asperger Syndrome(AS)/high-functioning autism(HFA) [15]. Using voxel-based morphometry analysis, concurrent with the dispositional empathy measures, researchers also found that young adult females had significantly larger gray matter volume locatged in the mirror-neuron system, which is highly related in empathy ability[28]. Females showed earlier and stronger brain processing of the action's purpose in the females brain compared with males in an ERP study[29]. Females were also more accurate when their feelings of the target[30].

Gender-based differences in emotion sensitivity and interpersonal sensitivity might also be a reason for the gender differences were found in the current study. Females often perform better in various emotional tasks than males. Studies focused on sex differences in facial expression processing found that female advantage in decoding of emotional cues both in adults and children[31].Consistently, females displayed superior performance in affective arousal and expression of emotion over males. For instance, increased corrugator activity for angry faces and increased zygomatic activity for happy faces were more pronounced for females compared to males[32]. Unpleasant and high arousing stimuli also evoked stronger N100 and N200 in females compared to males[33].

In addition, Singer and her coworkers found that males expressed a desire for revenge when observing players who had just betrayed them by administering pain[12]. However, males' hypothesized reversed responses toward unfair player did not occur in our study. It is possible that males did not process the information. This explanation might be supported by the shorter latency of oFRN in males compared with that in females. Females' automatic response toward others may lead to a longer latency in which to make sure that they could separate the likable player and unlikable one, whereas the males, who finished the evaluation phrase in a shorter time, could not finish an precise evaluation. Males' weaker empathic/emotional response and lower level of social sensitivity might not allow them to separate their representations of the two strangers within a few hundred million seconds. Finally, from an evolutionary perspective, males should show stronger sensitivity to unfair players' results, but our results were not completely in line with the evolutionary hypothesis. However, the rapid evaluation failure does not mean it will not work in the long run. Using more extreme targets (people one hates), the revenge effect on oFRN might be possible.

By manipulating observers' affective preferences toward different performers through trust games and by removing the effect of self-benefit, the present study found that females displayed a stronger FRN effect when observing likable players' outcomes compared to unlikable players' and a stronger P300 response toward unlikable ones; males did not display any difference in their responses toward the two strangers. These findings suggest that, at least in females, outcome evaluation can be affected by short-term affective preference even when the person being observed is a total stranger.