Abstract
Social presence affects behavioral performance and cognition, producing both facilitation and inhibition. However, how it modulates distinct cognitive components to shape goal-directed behavior remains unclear. This study investigated the neurocomputational mechanisms underlying social presence effects on instrumental learning by integrating behavioral, computational, and electroencephalographic (EEG) analyses. Participants performed a probabilistic learning task alone or under observation. Results indicated that social presence improved learning performance, especially with high feedback validity. This social facilitation effect was associated with more flexible behavioral adaptations in response to performance monitoring, evidenced by changes in win-stay/lose-switch strategies and post-error slowing. EEG analyses revealed that social presence promoted reliance on internal performance monitoring while diminishing weighting of external task feedback, reflected by increased correct-related negativity amplitudes following correct actions and reduced theta power to feedback. Furthermore, computational modeling revealed that social presence enhanced choice perseverance and working memory. Together, these findings provide insights into neurocomputational mechanisms through which social presence shapes goal-directed behavior, highlighting interactions between socially induced motivation and task-specific cognitive processes including performance monitoring, choice perseverance, and working memory. This work advances the mechanistic understanding of social facilitation and informs interventions leveraging social contexts to optimize learning.