CENTRAL MODIFICATIONS OF REFLEX PARAMETERS MAY UNDERLIE THE FASTEST ARM MOVEMENTS

Descending and reflex pathways usually converge on common interneurons and motoneurons. This implies that active movements may result from c hanges in reflex parameters produced by control signals conveyed by de scending systems. Specifically, according to the lambda-model. a fast change in limb position is produced by a rapid change in the threshold of the stretch reflex. Consequently, external perturbations may be in effective in eliciting additional reflex modifications of electromyogr aphic (EMG) patterns unless the perturbations art relatively strong. I n this way, the model accounts for the relatively weak effects of pert urbations on the initial agonist EMG burst (Ag1) usually observed in f ast movements. On the other hand, the same model permits robust reflex modifications of the timing and shape of the Ag1 in response to stron g perturbations even in the fastest movements. To test the model, we v erified the suggestion that the onset time of the Ag1, even in the fas test movements, depends on proprioceptive feedback in a manner consist ent with a stretch reflex. In control trials, subjects (n = 6) made fa st unopposed elbow flexion movements of similar to 60 degrees (peak ve locity 500-700 degrees/s) in response to an auditory signal. In random test trials, a brief (50 ms) torque of 8-15 Nm either assisting or op posing the movement was applied 50 ms after this signal. Subjects had no visual feedback and were instructed not to correct arm deflections in case of perturbations. In all subjects, the onset time of the Ag1 d epended on the direction of perturbation: it was 25-60 ms less in oppo sing compared with assisting load conditions. Assisting torques caused , at a short latency of 37 ms, an additional antagonist EMG burst prec eding the Ag1. The direction-dependent effects of the perturbation per sisted when cutaneous feedback was suppressed. It was concluded that t he direction-dependent changes in the onset time and duration of the A g1 as well as the antagonist activation preceding the Ag1 resulted fro m stretch reflex activity elicited by the perturbations rather than fr om a change in the control strategy or cutaneous reflexes. The results support the hypothesis on the hierarchical scheme of sensorimotor int egration in which EMG patterns and movement emerge from the modificati on of the thresholds and other parameters of proprioceptive reflexes b y control systems.