Control and Learning of Dynamics in Human Movement
Johns Hopkins University
Movement of our limbs and interaction with our surroundings are complex tasks, requiring precise coordination and control. In addition to controlling the complicated dynamical system of one's body, people also modify their movements based on the dynamics of their environment and the objects with which they interact. This research investigates different aspects of dynamics -- body, environment, and object dynamics -- in upper limb motor control through behavioral studies in various subject populations. In the first study, we examine the learning of arm movements to altered environment dynamics, introduced abruptly or gradually. We compared the performance of patients with cerebellar damage to that of age-matched controls. We found that the force field direction greatly affected the patients' ability to learn, regardless of whether the forces were introduced gradually (small error) or abruptly (large error). Patients performed better when the dynamics of the environment helped them compensate for movement dynamics associated with reaching. Thus, with cerebellar damage, the ability to adapt reaching movements to new environment dynamics becomes largely affected by the inherent dynamics of the arm. The second study focuses on the ability to modulate body dynamics, effectively adjusting the mechanical impedance of the arm, to ensure stability. Although cerebellar patients are impaired at adapting their movements to new environment dynamics, we wanted to test if cerebellar damage also affects the ability to control limb impedance in the presence of perturbing forces. In a postural maintenance task, we found that people were generally able to modify their endpoint (hand) stiffness for the two types of directional perturbations. However, the ability to selectively control the endpoint stiffness geometry was detrimentally affected by both cerebellar damage and age. Lastly, we investigate the effect of object dynamics on grip force control in healthy people during object interaction. When people manipulate objects in virtual environments or via robotic teleoperation systems, they receive limited somatosensory feedback about the (virtual or remote) objects. We examine the effects of force feedback of object dynamics, accuracy demands, and training on grip force control during indirect object interaction. People failed to couple grip and load force (typical for direct object manipulation) when force feedback was not provided, and exerted more grip force without force feedback and when accuracy demands were high. The results also highlight the importance of the force feedback condition during training, as it can affect subsequent grip force control.
human motor control, haptics