Abstract
Our ability to learn and generalize new motor skills is a fundamental aspect of human behavior. In previous work using a visuomotor adaptation task, we demonstrated that observational learning benefits the transfer of acquired motor skills to novel contexts (Martinez, Wang & Song 2019). Recent studies have shown that visuomotor adaptation relies on a combination of an explicit strategy, such as aiming, and implicit motor error reduction (Taylor, Krakauer & Ivry 2014). However, it is unclear how observation contributes to each of these components, and in turn, influences generalization. To address this question, we asked participants to perform three sequential visuomotor adaptation phases: observation, training and generalization. In the observation phase, each participant first watched a short movie in which an actor demonstrated a visuomotor adaptation task. Next, all participants performed the same training and generalization phases. Specifically, they adapted to a 45° cursor rotation to one target location (training), and then attempted to transfer their motor adaptation to untrained target locations (generalization). During these phases, we displayed a ring of numbers around the starting location, with the target location embedded at zero. The participants used this ring to report their aiming strategy. The presence or absence of this ring during observation was a between-subjects independent variable (ring vs. no-ring group). We found that the ring group generated a uniform aiming strategy across directions during generalization. The no-ring group created aiming strategies limited to the learned target location. Overall, these results suggest that visually congruent observation helps form an aiming strategy and determines how well this strategy can be transferred across contexts.