Modularity for sensorimotor control: evidence and a new prediction.

By combining a few modules, the CNS may learn new control policies quickly and efficiently. Support for a modular organization of the motor system has recently come from the observation of low dimensionality in the motor commands. However, stronger evidence would come from testing the predictions on the effect of an intervention on the mechanisms required to implement a modular controller. Thus, the authors propose to test the predictions of modularity on motor adaptation. They argue that unlike a nonmodular controller, a modular controller must adapt faster to a perturbation that is compatible with the modules (i.e., one that can be compensated by reusing the same modules), than to an incompatible perturbation (i.e., one that requires new modules).