ABSTRACT
The choice of a plant model is important for the upstream visuomotor transformation because it defines the motor information content of its inputs. Eye position is then subtracted off again during the saccade to give the instantaneous motor error command that drives the burst neurons until the saccade reaches its target. In the latter model, retinal error is simply mapped directly onto an initial motor error command, and saccade guidance is handled downstream with the use of displacement feedback. When the head is motionless, the oculomotor system’s solution to this problem is of course Listing’s law. The fundamental reason that we cannot get rid of it, or the degrees of freedom problem is that retinal error is by definition two-dimensional and expressed in oculocentric coordinates, whereas motor error must be three-dimensional and craniotopic in order to correctly control movements of the eyes relative to the head.
