Abstract
In dynamic environments, the ability to recalibrate sensorimotor mappings in response to unexpected perturbations is crucial. Visuomotor adaptation is accomplished through a combination of explicit and implicit learning processes. Previous studies revealed a relationship between working memory (WM) capacity and visuomotor learning, such that higher spatial WM predicts faster learning rates. However, little is known about the mechanisms that underlie this association. We hypothesized that higher WM capacity facilitates learning by reducing the cognitive load associated with strategic control processes, and further predicted that task-evoked pupil diameter (PD) would track variability in the cognitive effort exerted to overcome initial motor errors. Participants completed a spatial WM task, followed by a visuomotor rotation task using a stylus and touch surface to reach for targets while the direction of the cursor was rotated 45° relative to the hand to measure visuomotor adaptation. Eye-tracking data were continuously recorded throughout the task to measure PD. We found that PD tracked visuomotor learning, such that the initial exposure to the 45° rotation was accompanied by a period of relatively large pupillary responses across participants. The magnitude of these responses then decreased linearly over the course of learning, possibly reflecting a transition from explicit to implicit learning mechanisms. Consistent with recent work, higher WM capacity was associated with faster adaptation. Notably, we found that the decreasing relationship between PD and time on task was more critical for participants with low WM capacity: While high WM predicted rapid learning and relatively stable PD, error-related increases and subsequent reductions in PD strongly predicted learning rates among low-WM participants. Taken together, our results suggest that PD may effectively index strategic contributions to visuomotor adaptation, and moreover, that engaging in effortful cognitive control may support reduced motor error among learners with low spatial WM capacity.