Input-output properties of motor unit responses in muscles stretched by imposed displacements of the monkey wrist.

Reflex responses are "servo-like" where the output is graded with the input or are "triggered" where the output is independent of input once an input threshold is exceeded. Imposed displacements of monkey or human upper limb joints result in prolonged EMG output in the muscles stretched by the displacements. The longer-latency portions of the prolonged output have been variously reported to be servo-like or to be triggered in nature. In monkeys and humans, angular wrist displacements imposed by step loads result in three peaks (M1, M2 and M3) in the gross EMG recorded from the stretched muscles. Each gross EMG peak largely results from the firing of a separately-responding subpopulation of single motor units (SMUs). We studied the responses of SMUs to loads that were presented to the monkeys in a random order as to magnitude, duration and onset time. Average response histograms were constructed for the SMU responses for individual step load magnitudes. Averages were also constructed for the simultaneously-recorded gross EMG responses for each step load magnitude. The input parameters used were the initial velocity of displacement or the magnitude of step load, while the output was taken as the probability of firing/millisecond/presentation above baseline for the SMUs or the area under the response peaks above baseline for gross EMG. The results establish: 1) That it is not possible to unambiguously determine the input-output properties of the responses to imposed displacements utilizing the analysis of gress EMG activity due to the response characteristics of the various subpopulations of motor units contributing to the gross response. 2) That the SMU activity during all of the peak intervals is monotonically graded with increases in magnitude of the step load or the initial velocity of displacement. Hence, the long-latency portions of the EMG responses are servo-like in nature and are not preprogrammed or triggered responses. 3) That the gain (output/input) of the gross EMG responses almost entirely reflects the variation in the number of motoneurons recruited by changes in magnitude of the step loads rather than variation in the firing rates of motoneurons during the reflex responses.