Belief updating in uncertain environments are differentially sensitive to reward and punishment learning: Evidence from ERP

Abstract

Learning from rewards and punishments is crucial for adaptive decision-making, but it is still unclear how individuals integrate prediction errors to update beliefs in uncertain environments, particularly the distinctions between reward and punishment learning. We employed a probabilistic classification task and situated electroencephalography (EEG) signals within a hierarchical Bayesian framework, investigating distinctions between reward and punishment learning. Our findings indicate that participants exhibited superior learning performance in reward context compared to punishment context. Fitting the hierarchical Bayesian model revealed that punishment drives faster Bayesian belief updates, although these did not translate into improved behavioral outcomes. At the neural level, higher-level precision-weighted prediction error (pwPE2) was significantly positively correlated with FRN amplitude in the punishment context but not in the reward context, and the positive effect of pwPE2 on P300 amplitude was stronger in the punishment than the reward condition. These results provide electrophysiological signatures of punish context driving faster belief updates in uncertain environments. Our results provide novel evidence for a dual-process framework in reinforcement learning, underscoring distinct neural mechanisms underlying reward and punishment learning.