Characterizing bilateral M1 excitability changes after learning a novel spatiotemporal coordination pattern

Abstract Bimanual coordination is essential for many skilled everyday actions and requires maintaining specific spatiotemporal relationships between the two hands according to task demands. However, the contribution of the primary motor cortices (M1s) to learning such relationships is only partially understood. Conventional bimanual tasks necessarily engage both M1s in controlling the moving hands, making learning-related changes difficult to distinguish from those associated with movement execution. To reduce this effector-related confound, we used a unimanual and bimanual rhythmic coordination paradigm in which participants learned the same novel 90° relative-phase relationship by coordinating either hand with a virtual partner and compared these conditions with conventional bimanual practice. In three experiments, performance accuracy and stability improved during practice and improved further after a 6-h retention interval with comparable time-dependent changes across unimanual and bimanual trained groups. In Experiment 2, motor-evoked potentials measured before and after training revealed an overall increase in corticospinal excitability across both M1s under unimanual and bimanual training. In Experiment 3, bidirectional dual-site paired-pulse stimulation revealed effector-dependent changes in interhemispheric inhibition, in which bimanual practice produced no reliable directional modulation, whereas unimanual practice produced asymmetric changes that depended on the trained hand. Neither corticospinal excitability nor interhemispheric inhibition showed a consistent relationship with behavioral improvement. Together, these findings demonstrate comparable behavioral learning of a novel spatiotemporal coordination relationship across bimanual and unimanual contexts, while showing that behavioral improvement is not straightforwardly reflected in M1 neurophysiological changes. Causal intervention studies are therefore needed to determine whether and when M1 contributes directly to acquisition of novel coordination relationships.