IMPAIRMENTS OF REACHING MOVEMENTS IN PATIENTS WITHOUT PROPRIOCEPTION .2. EFFECTS OF VISUAL INFORMATION ON ACCURACY

1. The aim of this study was to determine how vision of a cursor indic ating hand position on a computer screen or vision of the limb itself improves the accuracy of reaching movements in patients deprived of li mb proprioception due to large-fiber sensory neuropathy. In particular , we wished to ascertain the contribution of such information to impro ved planning rather than to feedback corrections. We analyzed spatial errors and hand trajectories of reaching movements made by subjects mo ving a hand-held cursor on a digitizing tablet while viewing targets d isplayed on a computer screen. The errors made when movements were per formed without vision of their arm or of a screen cursor were compared with errors made when this information was available concurrently or prior to movement. 2. Both monitoring the screen cursor and seeing the ir limb in peripheral vision during movement improved the accuracy of the patients' movements. Improvements produced by seeing the cursor du ring movement are attributable simply to feedback corrections. However , because the target was not present in the actual workspace, improvem ents associated with vision of the limb must involve more complex corr ective mechanisms. 3. Significant improvements in performance also occ urred in trials without vision that were performed after viewing the l imb at rest or during movements. In particular, prior vision of the li mb in motion improved the ability of patients to vary the duration of movements in directional errors, path curvature, and late secondary mo vements. Comparable improvements in extent, direction, and curvature w ere produced when subjects could see the screen cursor during alternat e movements to targets in different directions. 4. The effects of view ing the limb were transient and decayed during a period of minutes onc e vision of the limb was no longer available. 5. It is proposed that t he improvements in performance produced after vision of the limb were mediated by the visual updating of internal models of the limb. Vision of the limb at rest may provide configuration information while visio n of the limb in motion provides additional dynamic information. Visio n of the cursor and the resulting ability to correct ongoing movements , however, is considered primarily to provide information about the dy namic properties of the limb and its response to neural commands.