Computational Mechanisms of Self-enhancement During Social Comparison and their Relationship to Internalizing Symptoms

Internalizing disorders like anxiety and depression commonly feature cognitive biases in self-evaluation, particularly in the context of social comparison. Despite the role of such biases in the severity and prognosis of internalizing conditions, limited work has identified computational mechanisms underlying self- evaluative biases in social contexts. In a sample of N = 292 participants, the present study applied hierarchical Bayesian computational modeling to a trait-evaluation task where individuals choose whether positive and negative traits better describe themselves or a close friend. We found that individuals generally engage in self-enhancement, more efficiently processing information that supports positive self-schema. However this effect flips as individuals report more symptoms, such that it becomes more difficult to integrate evidence in support of a pos- itive self-concept. These findings suggest that altered process- ing of both positive and negative self-referential information is a transdiagnostic mechanism driving aberrant self-evaluation in internalizing disorders.