Individual differences and the neural representations of reward expectation and reward prediction error

Amygdala tractography predicts functional connectivity and learning during feedback-guided decision-making

Flexibly adapting behavior in dynamic environments relies on fronto-limbic networks that include the amygdala, orbitofrontal cortex, and striatum. Animal work demonstrates that interactions among these regions are critical for flexible feedback-guided learning, but it remains unknown to what extent such anatomical-functional interactions operate in humans. Here, we use connectivity analyses in both structural and functional MRI to further our understanding of how brain circuits work in conjunction to promote goal-directed behavior. In particular, fiber tracking based on diffusion-weighted imaging provides information about anatomical connectivity between brain structures, and functional MRI provides estimates of functional connectivity between structures. We found that, during a feedback-guided reversal learning task, the strength of estimated white matter tracts from the amygdala to the hippocampus, orbitofrontal cortex, and ventral striatum predicted both how subjects adapted their behavior following positive and negative feedback, and the functional connectivity (estimated from functional MRI time series) between the amygdala and these regions. In addition, we identified a dissociation between an amygdala-hippocampus circuit that predicted response switching, and an amygdala-orbitofrontal cortex circuit that predicted learning following rule reversals. These findings provide novel insights into how the anatomy and functioning of amygdala-related brain circuits mediate different aspects of feedback-guided learning behavior.

Reinforcement learning signals predict future decisions

Optimal behavior in a competitive world requires the flexibility to adapt decision strategies based on recent outcomes. In the present study, we tested the hypothesis that this flexibility emerges through a reinforcement learning process, in which reward prediction errors are used dynamically to adjust representations of decision options. We recorded event-related brain potentials (ERPs) while subjects played a strategic economic game against a computer opponent to evaluate how neural responses to outcomes related to subsequent decision-making. Analyses of ERP data focused on the feedback-related negativity (FRN), an outcome-locked potential thought to reflect a neural prediction error signal. Consistent with predictions of a computational reinforcement learning model, we found that the magnitude of ERPs after losing to the computer opponent predicted whether subjects would change decision behavior on the subsequent trial. Furthermore, FRNs to decision outcomes were disproportionately larger over the motor cortex contralateral to the response hand that was used to make the decision. These findings provide novel evidence that humans engage a reinforcement learning process to adjust representations of competing decision options.