| The goal of this project was to assess how neurons in primary motor cortex (MI) of monkeys responded to perturbations at the shoulder and elbow as they performed a simple postural task. For this purpose, single cell and electromyographic (EMG) recordings were obtained as four monkeys wore a robotic exoskeleton (KINARM) that maintained their arm in the horizontal plane. Each monkey was trained to maintain its hand at a central target. After a random amount of time, a 300ms load was applied to the shoulder and/or elbow. Three load conditions were applied at each joint (flexor, extensor and no load) creating a total of nine conditions, including loads applied at both joints, loads at only a single joint as well as a no load condition.;We found that just over half the cells recorded in the task modulated their activity with the mechanical loads applied to the shoulder and/or elbow. Approximately 2/3 of the responsive cells modulated their activity with loads applied at a single joint, while the remaining one third of cells were sensitive to loads applied at both the shoulder and elbow joints. We examined the timing of these cell responses and found a broad range of onset times. Cells responsive to loads applied to both joints displayed an interesting coupling in their activity: cells tended to be maximally active for loads that caused flexion at one joint combined with extension at the other joint. This tendency was also observable in the distribution of cell preferred directions in torque space, causing a significant bimodal distribution. Finally, with the use of newly developed directional tuning scores, we were able to highlight one of the most interesting findings of this study: many cells showed early bimodal responses, which became unimodal over time. Qualitatively similar observations were made for the EMG data collected during the task. These findings provide new insights into how multi-joint perturbations are represented in MI. |
