Systematic modulation of sensorimotor learning by domain-specific working memory

People differ in how quickly they learn and adapt sensorimotor skills, and these differences have been linked to individual variation in working memory capacity (WMc). In tasks that can be supported by cognitive strategies, visuospatial WMc has been proposed as a key contributor. However, it remains unclear whether this association reflects domain-specific mechanisms or domain-general executive resources, and whether it extends beyond spatial memory to other visual features. Here, we systematically tested whether domain-specific or domain-general WMc predicts adaptation across three visuomotor adaptation (VMA) tasks known to differentially engage explicit and implicit learning systems. After obtaining independent measures of spatial and feature-based WMc, healthy subjects completed (1) a standard VMA task, which engages both explicit and implicit learning systems; (2) a delayed feedback VMA task, which isolates explicit learning; and (3) a clamped feedback VMA task, which isolates implicit learning. Our results provide converging evidence in support of a domain-specific association between spatial WMc and individual differences in learning. In Experiments 1 and 2, greater spatial WMc was associated with robust increases in explicit learning whereas featured-based WMc was not associated with learning outcomes. Surprisingly, in Experiment 3, greater spatial WMc was associated with reduced implicit learning, suggesting an interaction between latent cognitive capacities and implicit learning not accounted for by traditional models. These results shed light on the precise cognitive mechanisms underlying sensorimotor adaptation and provide novel insight into domain-specific links between spatial WM and motor learning.