Online and post-trial feedback differentially affect implicit adaptation to a visuomotor rotation

Abstract

Multiple motor learning processes can be discriminated in visuomotor rotation paradigms. At least four processes have been proposed: Implicit adaptation updates an internal model based on prediction errors. Model-free reinforcement reinforces actions that achieve task success. Use-dependent learning favors repetition of prior movements, and strategic learning uses explicit knowledge about the task. The current experiment tested whether the processes involved in motor learning differ when visual feedback is altered. Specifically, we hypothesized that online and post-trial feedback would cause different amounts of implicit adaptation. Twenty subjects performed drawing movements to targets under a 45° counterclockwise visuomotor rotation while aiming at a clockwise adjacent target. Subjects received visual feedback via a cursor on a screen. One group saw the cursor throughout the movement (online feedback), while the other only saw the final position after movement execution (post-trial feedback). Both groups initially hit the target by applying the strategy. After 80 trials, subjects with online feedback had drifted in clockwise direction [mean direction error: 15.1° (SD 11.2°)], thus overcompensating the rotation. Subjects with post-trial feedback remained accurate [mean: 0.7° (SD 2.0°), TIME × GROUP: F = 3.926, p = 0.003]. We interpret this overcompensation to reflect implicit adaptation isolated from other mechanisms, because it is driven by prediction error rather than task success (model-free reinforcement) or repetition (use-dependent learning). The current findings extend previous work (e.g., Mazzoni and Krakauer in J Neurosci 26:3642–3645, 2006; Hinder et al. in Exp Brain Res 201:191–207, 2010) and suggest that online feedback promotes more implicit adaptation than does post-trial feedback.