Target configuration determines how and what we learn during sensorimotor adaptation

Abstract

Motor adaptation--the process of correcting movement errors through feedback and practice--is a fundamental human capacity that keeps our actions well-calibrated amid changes in the environment and the body. However, how training context--specifically, the configuration of targets in the workspace--shapes how we learn and what we learn during motor adaptation remains unknown. To investigate this, we conducted two reaching experiments in which participants experienced a visuomotor gain perturbation, with feedback scaled to 1.3x (Exp 1) or 0.7x (Exp 2) the actual movement distance. In both experiments, participants were assigned to either the Extent Group, which trained with targets of varying amplitudes in a single direction, or the Angular Group, which trained with targets of equal amplitude in different directions. We found marked differences in how the two groups learned: the Angular Group learned more implicitly than the Extent Group, as evidenced by larger post-perturbation aftereffects when participants were instructed to forgo re-aiming strategies. Just as striking were the differences in what the two groups learned: the Angular Group learned a translation rule, which generalized to new directions but not amplitudes, while the Extent Group learned the imposed gain rule, which generalized to new amplitudes but not directions. Together, these findings underscore the importance of training context in determining how and what we learn.