The role of aversive learning in social interactions

Sammanfattning: It could be argued that our survival as humans hinges on our ability to interact socially with others. Our social interactions are influenced by evaluations of each other: we cooperate with those we like and avoid or are aggressive towards those we dislike or are afraid of. The aim of this thesis was to investigate how we come to learn to fear or dislike other individuals based on who they are; and how such learned evaluations influence actual social interactive behaviors. One elegant way to study how humans respond and react to threats in the environment is classical fear conditioning, where we can study how emotional values are created, upheld and changed. Research using classical fear conditioning has found that people are predisposed to develop stronger associations between threatening events and certain categories of stimuli (e.g., snakes, angry faces, and faces of individuals belonging to social out-groups). These biased aversions tend to persist even when circumstances change and the threat is no longer present. Though the fear system underlying this type of learning may be useful under some circumstances, it may also be at the root of some persistent social problems affecting modern societies (e.g., xenophobia). To address these questions experimentally, this thesis aimed to identify how we learn to associate threats to different social groups (e.g., racial and hierarchical) (Study I & Study II); whether learned aversions influence anti-social interactive behaviors (Study III); and to study the mechanisms of maladaptive reciprocal punishments in dyadic interactions (Study IV). In Study I, we found that activity linked to both conditioned fear and perception of racial out-group members jointly contributed to the expression of race-based biases in learning and behavior. Importantly, we showed that brain activity in the fear-learning-bias network was related to participants' discriminatory interactions with new out-group members at a later time. In Study II, we investigated the interaction between learned social dominance and social out-group (i.e., ethnicity) threats to understand if dominance hierarchy knowledge (i.e., observation of threats) can change direct experience with out-group members. We found a dissociation between implicit and explicit measures of out-group biases, such that implicit measures (i.e., Implicit Association Task and skin-conductance responses) of the participants revealed out-group biases, whereas their explicit measures (i.e., modern racial prejudice scale and a social interactive task; the modified ‘Cyberball’ game) did not. In Study III, we found that learned aversions influenced future retaliation in a social context. Our results suggest that aggressive traits, when paired with aversive learning experiences, enhance the likelihood to act anti-socially toward others. In Study IV, we demonstrate that participants punish co-players, despite the cost of receiving punishment back. These findings describe a form of self-punitive behavior previously documented in animals. Participants’ tendencies to administer shocks were exacerbated when the co-player initiated punishment, indicating that a small initial offense motivated punishing behavior over time. This finding suggests a simple experimental model of a vicious cycle of punishments. Together these findings highlight the role of aversive learning in social interactions.