Abstract
A period of holding still follows every movement. It has been assumed that for the arm, moving and holding are functionally independent: movement is via an adaptive, feedback-dependent controller that generates commands to transport the arm, while holding is via setting of reflexes that produce a postural field at movement endpoint. This assumption predicts that commands that move the arm should not affect the postural field at movement termination. Surprisingly, we found that as the reach commands changed, so did the ensuing postural field. The postural commands depended on mathematical integration of the reach commands. Following damage to the corticospinal tract, despite severe reach impairments, the holding system faithfully integrated the imperfect reach commands on a trial-by-trial basis. Together, these findings suggest that holding the arm still is accomplished via a separate, likely subcortical structure that acts as a mathematical integrator of the commands generated by the cortical reach controller.