Abstract
It has been widely accepted that the CNS develops a representation (model) of the environment, but what is not entirely clear is the coordinate reference frame used. We explored how visual feedback influenced the coordinate frame in which the CNS stores and recalls these memories of learned skills in a reaching-generalization task. Four groups of subjects trained to perform reaching movements in a perturbing force field, two with aligned (first person) visual feedback and two with non-aligned (vertical screen). After 170 trials of practice, we asked subjects to extrapolate (generalize) what they learned to a new part of the workspace in novel force environments (endpoint-based versus joint-based extrapolated force fields). Regardless of the test condition, all subjects improved their ability to generalize skills to the new workspace. There was evidence that the extrapolation of their learned skills was based on both object-centered and joint-based coordinates. Consistent with previous studies, subjects performed significantly better in joint-extrapolated force field, but only if the visual feedback was vertical. Subjects performed equivalently in both force fields with aligned (first person) feedback. These findings suggest that the type of visual feedback biases the way subjects perform, and that learning results can be significantly influenced by feedback.