Abstract
1. Subjects sat with their right hands in a mould which was attached to the shaft of a servo-controlled motor. 2. There were two different tasks. In one, the subject attempted to maintain a constant wrist flexing force, while the motor imposed flexion-extension movements. In the other, the subject attempted to maintain a constant wrist position while the motor exerted variable flexion-extension forces. 3. The 'maintain force' task was punctuated by randomly timed ramp extension movements to test the stretch reflex activity. The 'maintain position' task was punctuated by force changes which were the same as those recorded during the ramp extensions mentioned above. In this way it was possible to test stretch reflexes with the same disturbance during the 'maintain force' and 'maintain position' tasks. 4. When, after some practice, subjects had become good at maintaining the constant force, later components of the stretch reflex (after 40 ms) were found to be smaller than when they were maintaining a constant position. 5. Attempts to maintain a constant position often involved co-activation of the wrist extensor muscles. This could be prevented by local anaesthesia of the radial nerve. The long-latency stretch reflex of the flexors was still task dependent after this had been done. 6. The roles of co-activation and of task-dependent reflex responses are discussed.