Balanced motor primitive can explain generalization of motor learning effects between unimanual and bimanual movements

Motor learning in unimanual and bimanual planar reaching movements has been intensively investigated. Although distinct theoretical frameworks have been proposed for each of these reaching movements, the relationship between these movements remains unclear. In particular, the generalization of motor learning effects (transfer of learning effects) between unimanual and bimanual movements has yet to be successfully explained. Here, by extending a motor primitive framework, we analytically proved that the motor primitive framework can reproduce the generalization of learning effects between unimanual and bimanual movements if the mean activity of each primitive for unimanual movements is balanced to the mean for bimanual movements. In this balanced condition, the activity of each primitive is consistent with previously reported neuronal activity. The unimanual-bimanual balance leads to the testable prediction that generalization between unimanual and bimanual movements is more widespread to different reaching directions than generalization within respective movements. Furthermore, the balanced motor primitive can reproduce another previously reported phenomenon: the learning of different force fields for unimanual and bimanual movements.