Abstract
Every movement ends in a period of stillness. In current models of reaching, the commands that hold the arm still at a target, depend on the spatial location of the target, but not on the reach commands that moved the arm to the target. Contrary to this assumption, in a series of tasks in humans and non-human primates, we find that the commands that hold the arm still have a consistent mathematical relationship to the reach commands that moved the arm. With damage to the corticospinal tract, reach commands become impaired as expected, but remarkably, the mathematical dependence of hold commands upon the now imperfect reach commands remains intact. Therefore, we find that the brain uses a design principle in which its holding controller is not driven by the spatial location of the target. Rather, holding is obtained via mathematical integration of moving, potentially through a subcortical structure.
Footnotes
We have added intramuscular EMG recordings in 2 monkeys performing a point-to-point reaching task to several targets in the vertical plane. We show the link between the integral of moving activity and the final holding activity, in each muscle. The text and figures have been updated accordingly. New authors have been added to the work.
