A State Space Model of a Visuomotor Rotation Experiment

Abstract

We present a new linear time-invariant (LTI) state space model to explain adaptation in human motor control. We focus on a visuomotor rotation experiment in which a human subject must rapidly move a cursor on a horizontal screen through a target disk. The hand itself is occluded from view, while the cursor is rotated relative to the hand angle by an amount of r degrees. Our model is based on the application of well-known techniques from control theory, in particular regulator theory. The model is simple, yet it reveals a plausible architecture for the high level computations underlying human motor control, including a representation of the internal model. It is a two state LTI model, where each state has a physical interpretation