Generalization to Local Remappings of the Visuomotor Coordinate Transformation Generalization to Local Remappings of the Visuomotor Coordinate Transformation Materials and Methods Procedures for the Two-point Groups Computational Model Training Protocol for the Model

To investigate the representation and plasticity of the visuomotor coordinate transformation we examined the changes in pointing behavior subsequent to local remappings. The visual feedback of nger position was limited to one or two locations in the workspace, at which a discrepancy was introduced between the proprioceptively-and visually-perceived nger position. Such a local remapping 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 how the central nervous system computes the visuomotor transformation. However, a simple model, in which the transformation is computed via the population activity of a set of units with large sensory receptive elds, is shown to capture the observed pattern. In order to reach or point to a visually perceived target, the central nervous system (CNS) must transform visual information into coordinates appropriate for movement (Andersen et al. This visuomotor coordinate transformations forms an integral part of the planning and on-line execution of movement. Both experimentally-imposed visual perturbations, such as those resulting from wearing prism glasses (for a review see Welch, 1986), and lesions to brain areas such as the posterior parietal cortex, result in motor deecits attributable to inaccuracies in transforming between visual and motor coordinate systems (Andersen, 1987). The question of how the vi-suomotor coordinate transformation is learned and represented remains of basic importance to understanding the motor system. One of the ways in which the representation of the visuomotor transformation has been probed relies on the analysis of pointing errors. Using this paradigm, studies by Soechting and Flanders (1989a, 1989b) have suggested that the visuomotor transformation from target location to the orientation of the arm relies on a linear approximation of the nonlinear function that relates the two coordinate systems. Subsequent studies, also based on the analysis of dependences among errors, suggest that the transformation from elevation and distance of the target to arm elevation may be computed separately from the transformation from target azimuth to arm yaw angle (Flanders and Soechting, 1990). In this paper we explore the representation of the visuomotor transformation through a diierent paradigm, which analyzes the pattern of adaptation arising from a local perturbation of the visuomotor relationship. This paradigm makes it possible to address two questions. First, what eeects do remappings at one location in space have on visuomotor behavior elsewhere. Second, how do these patterns of eeects constrain computational models …