Comparison of Visual Feedback Conditions While Learning a Redundant Novel Visuomotor Task

Human movement relies on coordinating multiple degrees of freedom (DOFs) converging onto fewer DOFs of functional goals. We compared two forms of visual feedback as participants learned a novel motor task: using 18-DOF finger movements to control a 3-joint kinematic chain, ultimately to place a 2-dimensional cursor within targets on a screen. Eighteen participants were randomly assigned to either a Chain group (who could see the chain and cursor) or a No-Chain group (who saw only the cursor). The Chain group had faster performance early in training, despite the identical mechanics faced by the two groups. Both groups improved with training, converging by the end of day one, and retained that performance on the second day. However, neither group learned to aim initial movements more directly to the target. Variance within task and null spaces decreased at similar rates, counter to the minimal intervention principle which posits greater reduction in the task-relevant variability. These results suggest that vision of intermediate control variables provides an early performance advantage, but that learners may adopt consistent control strategies that prioritize stability over directional optimization. These findings offer insights into the role of visual feedback during skill acquisition, with implications for designing feedback-based rehabilitation protocols.