Distinct optimization of motor control and learning: Self-paced motor learning achieves different motor patterns from overtrained

Abstract

Self-paced movement, refined through extensive training and optimization, serves as an ideal model for investigating internal objectives within the sensorimotor system without external constraints. Previous studies have demonstrated that self-paced movement is consistent within individuals but variable across individuals, emphasizing a robust and unique optimization in self-paced motor control. While motor learning is also an optimization process, it remains unclear whether the internal objectives overlap between motor control and learning, or, more specifically, whether motor patterns identified in self-paced motor control are replicated in motor learning. Given that optimality principles typically address redundancy, human behavioral experiments were designed to equate movement distances for both motor control and learning in self-paced movements and to thus compare redundant motion parameters, movement velocity and duration. We dissected participants reaching movements at their preferred pace, controlling a visual cursor to a target on a display over specified movement distances, through a detailed analysis that that accommodates nonlinear interaction and significant inter-individual baseline differences of movement velocity and duration. In self-paced motor control, participants adjusted both velocity and duration in response to changes in target distances. However, when visuomotor shift perturbations were applied to the cursor in the longitudinal direction, triggering motor learning, participants adapted their movement distance solely by modulating velocity, while leaving duration unchanged. These results suggest that distinct optimization strategies are employed in motor control and learning to resolve movement redundancy. Significance statementNatural behaviors are thought to be well-optimized. Self-paced movements--an important form of such behaviors--showing significant stability within individuals but variability across individuals. This suggests that the sensorimotor system in the central nervous system operates through a unique and robust optimization process for self-paced motor control. Self-paced motor control is further refined through motor learning. However, it remains unclear whether motor control and learning share a common objective function. Our findings demonstrate that laboratory-based motor learning does not replicate the motor patterns observed in overtrained, self-paced movements performed under natural conditions. This suggests that distinct optimization mechanisms are at play in motor control and learning, indicating that these processes are not contiguous.