Abstract
People can perform similar movements based on distinct perceptual variables: For example, they can reach to the positions of objects, or they can make the same reach in order to move the hand forward a specified distance. Our experiments used a transfer-of-adaptation paradigm to test the idea that these two types of reaches are mediated by a common, early perceptual representation. Methods: Subjects reached in virtual reality to positions of visual targets or to match lengths of visible rods. Both tasks used the same set of reach endpoints and similar timing constraints; trajectories were almost indistinguishable. Feedback was given at the end of each reach by a visible marker indicating final hand position or by a second rod whose length indicated distance reached. During some runs, subjects received misleading position- or length-based feedback about the distance component of hand position. We examined the transfer of the resulting visuomotor adaptation between tasks. Results: Subjects showed near-complete adaptation to misleading positional feedback, but a few did not adapt to length-based feedback. Most subjects showed little or no transfer of adaptation between tasks; a small number showed partial or complete transfer. Examination of differences in reach latency between the tasks suggests that subjects who showed transfer of adaptation were using overlapping perceptual strategies in the two tasks; future experiments will examine the possibility that their perceptions of location were less accurate in the virtual reality than those of subjects who did not show transfer of adaptation. We will also attempt to determine what is being adapted in our tasks by examining the limb- and posture-specificity of the adaptation. Conclusions: Positional and length-based reaches are not mediated by a common, early perceptual representation; instead, information about position and length are kept separate until after visuomotor adaptation occurs.