The planning of visually guided arm movements: Feedback perturbation and obstacle avoidance studies

With little conscious effort, the central nervous system (CNS) is capable of coordinating visual information from the distal environment with the complex commands required to organize action. One account of this remarkable ability is that CNS hierarchically organizes movement control, with an effector independent planner superordinate to the variable details of movement execution. This thesis is concerned with the extent to which arm movements are preplanned in extrinsic, Cartesian space. Two fundamental questions are addressed: what aspects of movement are centrally planned, and by what criteria is the plan chosen?

The first part of this thesis presents two sets of visual feedback perturbation experiments supporting the notion of Cartesian planning of movement trajectories. A novel prism adaptation technique is employed to investigate the relationship between the static visuomotor map and path planning. Reorganization of the mapping from extrinsic space to intrinsic coordinates is found to result in a corresponding adaptation in movement path. The planning of movement velocity is addressed by perturbing the visual feedback velocity without altering the path. While the production of rhythmic arm movements are insensitive to such perturbations, discrete pointing movements are susceptible to perturbations in the feedback velocity. When the feedback is artificially skewed, subjects adapt their behavior by skewing movement velocity in the opposite direction.

The second part of the thesis focuses on the planning of obstacle avoidance movements. The obstacle rotation paradigm is introduced, and a series of experiments in two and three dimensions show that such movements are not consistent with strictly Cartesian planning of end-point trajectories: the movement path varies in a systematic manner as the orientation of the movement changes. Two stability models are presented to account for these observations, one based on the arm's kinematics and one on its inertial properties. Finally, evidence is presented ruling out both perceptual and execution level explanations of the observed path variation. Together, these results show that the structure of the actuator is taken into account to optimize the movement path. In the final chapter, a modified hierarchical model is discussed which accounts for the results of both parts of the thesis. Copies available exclusively from MIT Libraries, Rm. 14-0551, Cambridge, MA 02139-4307. Ph. 617-253-5668; Fax 617-253-1690.