Item

Abstract

Throughout development, we learn through trial and error in the face of uncertainty. This ‘reinforcement learning’ supports youth’s ability to adapt to environmental demands. Theory suggests that early experiences optimize learning to prioritize socially relevant cues across multiple levels of analysis. Yet, experimental paradigms often use abstract stimuli and limit measures of learning to behavioral summaries such as task accuracy. This study uses computational models grounded in reinforcement learning and dynamic systems frameworks to examine how socioemotional context modulates cognitive and psychophysiological learning processes. In a sample of 56 youth aged 12-15 years, a within-person experimental design quantified continuous behavior and heart rate (~700 observations of behavior and heart rate, per person) during an adapted reinforcement learning task with stimuli that varied in socioemotional relevance. A series of Bayesian multilevel models revealed that, compared to traditionally-used non-emotional stimuli, learning from stimuli high in socioemotional arousal enhanced behavioral performance. Specifically, computational models suggested that socioemotional salience may elicit faster value-updating processes and a qualitative shift to more exploitative decision-making. Underlying psychophysiological engagement was modulated by heightened sensitivity to learning from rewards, showing faster value-updating in the context of rewards aligned with more optimal psychophysiological flexibility and organization. Together, these findings clarify the contexts and modulatory processes that shape developmental learning mechanisms.