Descifrando la mente social en la depresión: una disociación computacional entre la confianza explícita y la...

This study employed a computational psychiatry approach to investigate how depressive symptoms impact belief updating about others' trustworthiness across different social contexts. Using a modified Trust Game with trustworthy (75% cooperative) and untrustworthy (25% cooperative) partner conditions, we examined both explicit belief reports and implicit computational processes in 169 participants. Linear mixed-effects models revealed that higher depression scores predicted lower posterior trust beliefs uniformly across both social conditions, indicating a generalized negative updating bias. Computational modeling further demonstrated a context-dependent pattern in likelihood bias (p (O|H)), with depressive symptoms associated with reduced bias specifically in the untrustworthy partner condition. This reflects cognitive inflexibility in processing negative social evidence and reveals the computational mechanisms underlying dichotomous thinking in depression. Additionally, depression scores correlated with reduced cooperative behavior exclusively in the untrustworthy condition. These findings reveal a novel dissociation in depression: while explicit social beliefs show pervasive negativity, implicit computational biases emerge specifically in threatening social contexts, supporting theoretical frameworks that emphasize process dissociation in mood disorders. The study bridges social cognition and computational psychiatry by identifying specific computational markers of impaired belief updating in depression.