Reinforcement motor learning after cerebellar damage is related to state estimation

Abstract

Growing evidence suggests that the cerebellum may play a role in neural mechanisms of reinforcement learning, but the nature of its contribution remains unclear. Recent work posited that cerebellum-dependent sensory prediction contributes to reinforcement learning in motor contexts by enhancing body state estimates, which are necessary to solve the credit-assignment problem. The objective of this study was to test the relationship between the predictive component of state estimation and reinforcement motor learning in individuals with cerebellar degeneration. Individuals with cerebellar degeneration and neurotypical control participants completed two tasks: a reinforcement learning task that required them to alter the angle of reaching movements and a state estimation task that tested the somatosensory perception of active and passive movement. The state estimation task permitted calculation of the active benefit shown by each participant, which is thought to reflect the cerebellum-dependent predictive component of state estimation. We found that the cerebellar and control groups showed similar magnitudes of learning with reinforcement and active benefit on average, but there was substantial variability across individuals. Using multiple regression, we assessed potential predictors of reinforcement learning within the cerebellar group. Our analysis included active benefit, clinical ataxia severity, movement variability, movement speed, and age as independent variables. We found a significant relationship in which greater active benefit predicted better learning with reinforcement. No other variables showed significant relationships with learning. Overall, our results support models that posit an indirect role for the cerebellum in reinforcement motor learning: one in which it may contribute to state estimation.