Generalization to local remappings of the visuomotor coordinate transformation

Abstract

During visually guided movement, visual representations of target location must be transformed into coordinates appropriate for movement. To investigate the representation and plasticity of the visuomotor coordinate transformation, we examined the changes in pointing behavior after local visuomotor remappings. The visual feedback of finger position was limited to one or two locations in the workspace, at which a discrepancy was introduced between the actual and visually perceived finger position. These remappings induced changes in pointing, which were largest near the locus of remapping and decreased away from it. This pattern of spatial generalization highly constrains models of the computation of the visuomotor transformation in the CNS. A simple model, in which the transformation is computed via the population activity of a set of units with large sensory receptive fields, is shown to capture the observed pattern.

Fig. 1.

Schematic predictions made by six qualitative models regarding the pattern of generalization that would result from a perturbation at a single point. Assuming that the central visual target has been remapped to a finger position to the right of the target, thearrows represent subsequent predicted changes in pointing behavior.

Fig. 2.

Apparatus used to introduce limited visuomotor remappings. The position of the finger was captured on-line by a computer, which calculated the perturbed finger position. The feedback of finger position was projected onto a screen as a cursor spot. Looking down at the mirror, the subjects saw the virtual image of the cursor spot, in the plane of the finger—the actual finger location was hidden from view. By controlling the illumination of the cursor spot, the visual feedback, and therefore the remapping, could be limited to particular areas of the workspace.

Fig. 3.

a, The position of the grid of targets is shown relative to the subject. Also shown, for thex-shift condition, is the perceived and actual finger position when pointing to the central training target. The visually perceived finger position is indicated by a cursor spot that is displaced from the actual finger position. b, A schematic showing the perturbation for the x-shift group. To see the cursor spot on the central target, the subjects had to place their finger at the position indicated by the tip of thearrow—a 10 cm, one-point visuomotor remapping.c, A schematic similar to b showing the perturbation for the one-point y-shift group.d, A schematic showing the perturbation and target numbering for the two-point y-shift group.

Fig. 4.

The targets (solid squares) and preexposure pointing locations are shown for all five groups as 95% confidence ellipses centered around the mean.

Fig. 5.

Target acquisition time as a function of trial during the exposure phase for the one-point x-shift (a), one-point y-shift (b), and two-point y-shift groups (c), plotted relative to their respective controls. For clarity, the SE bars are shown in one direction only.

Fig. 6.

Average change in pointing for the one-point (a) and two-point (c) control groups. Thearrows show the change centered on the visually presented target along with 95% confidence ellipses. Vector field of changes smoothed with Gaussian kernels for the one-point (b) and two-point (d) control groups.

Fig. 7.

Average change in pointing for the one-pointx-shift (a), one-pointy-shift (b), and two-pointy-shift (c) groups. Smoothed vector field of changes for one-point x-shift (d), one-point y-shift (e), and two-pointy-shift (f) groups. Proportion adaptation relative to the size of the perturbation for the one-point x-shift (g), one-pointy-shift (h), and two-pointy-shift (i) groups. In g, the lightest shade corresponds to 40% adaptation and the darkest shade corresponds to 11% adaptation; inh, the lightest shade corresponds to 16% adaptation and the darkest shade corresponds to 6% adaptation; in i, the lightest shadecorresponds to 58% adaptation in the positive ydirection, and the darkest shade corresponds to 42% adaptation in the negative y direction.

Fig. 8.

Pattern of generalization for the model under each of the three different experimental conditions. Average change in pointing for the one-point x-shift (a), one-point y-shift (b), and two-pointy-shift (c) conditions. This was computed by subtracting preexposure from postexposure pointing of the model to targets at 5 cm intervals over the workspace. Proportion adaptation relative to the size of the perturbation for the one-pointx-shift (d), one-pointy-shift (e), and two-pointy-shift (f) conditions.