The Impact of Different Learning Processes on Acquisition, Transfer, and Proprioception in Complex Motor Tasks

Abstract

Motor skill learning involves multiple mechanisms, including use-dependent learning (UDL), reinforcement learning (RL), and error-based learning (EBL). These operate over different time scales and neural pathways, contributing uniquely to skill acquisition, consolidation, and transfer. Here we asked how these mechanisms support the acquisition of a spatially complex maze navigation task in five groups of healthy young adults. Groups received one of five types of feedback during training of their unseen dominant hand: UDL (no feedback), RL (binary success/failure feedback with a static threshold), RLA (binary feedback with an adaptive threshold), RLB (binary feedback with adaptive threshold and brief flash of cursor feedback), or EBL (continuous real-time cursor feedback). Skill, transfer, and proprioceptive acuity were assessed pre- and post- training using a speed-accuracy function (SAF) for each hand and a two-alternative forced-choice shape discrimination task. Results showed that UDL and RL groups exhibited no improvement post-training, while EBL, RLA, and RLB groups demonstrated accuracy improvements. EBL and RLB participants experienced a significant reduction in movement variability, with EBL showing a greater decrease compared to UDL. The left-hand SAF revealed improvements in accuracy across all groups except UDL. All groups showed reduced variability in the left hand, suggesting intermanual transfer, with EBL transferring more variability improvements than UDL. No significant proprioceptive changes were observed in any group. These findings provide new insights into motor skill learning, emphasizing that even minimal feedback can facilitate complex skill acquisition and transfer and has significant implications for studies where error-based learning may not be applicable. New and NoteworthyMinimal, adaptive binary feedback can effectively support the acquisition and transfer of spatially complex motor skills. While use-dependent and static reinforcement learning failed to enhance performance, adaptive reinforcement and error-based feedback significantly improved accuracy and reduced movement variability. Notably, these gains transferred to the untrained hand, highlighting the potential of reinforcement-based strategies in contexts where error-based learning is limited or unavailable, offering important implications for rehabilitation and motor training design.