Abstract
Visuomotor adaptation (VMA) is a critical process that enables people to recalibrate their sensorimotor mappings to overcome unexpected perturbations. However, the mechanisms underlying substantial individual differences in VMA learning outcomes remain unclear. Typically, VMA is accomplished through an interplay of explicit and implicit learning: effortful, explicit processes that emerge early in learning are later complimented by slower implicit processes driven by sensorimotor error. In Experiment 1, we examined whether VMA is supported by domain-specific or general working memory capacity (WMC). After obtaining independent measures of spatial- and object-based WMC, participants completed a standard 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. Our results demonstrated a domain-specific effect of WMC, with a strong association between higher spatial WMC and improved VMA outcomes; surprisingly, higher object WMC had no effect. In Experiment 2, we tested how spatial WMC specifically contributes to VMA via modulation of explicit learning mechanisms. We modified the VMA task to limit implicit learning by removing continuous feedback, providing endpoint feedback after a 500 ms delay. We found that higher spatial WMC again enhanced VMA learning, while object WMC counterintuitively inhibited it. To further link WMC to explicit strategy in VMA, we computed task-evoked pupil diameter (PD) as an indirect measure of cognitive effort and arousal during both tasks. These analyses revealed a consistent association between larger PD and improved learning. Notably, the relationship between PD and learning was significantly stronger among participants with low spatial WMC. Taken together, our results demonstrate a clear domain-specific contribution of WMC to sensorimotor learning, and further suggest that participants with low spatial WMC can improve their learning outcomes via additional cognitive effort.