1. The activity of single afferent units was recorded in cervical dorsal root ganglia (DRG) in two macaque monkeys as they generated alternating flexion and extension torques about the wrist during a step-tracking task. During these isometric and auxotonic muscle contractions, electromyographic (EMG) activity was recorded with electrode pairs in up to 12 independent forearm muscles. Spike-triggered averages (STAs) of rectified EMG activity were used to identify afferents that were associated with correlated facilitation of active muscles. 2. Our aim was to find peripheral afferents producing postspike effects in muscles and to compare their properties with those of corticomotoneuronal (CM) and rubromotoneuronal (RM) cells previously obtained under identical behavioral conditions. We documented the timing, magnitude and distribution of their postspike facilitation (PSF) of forearm muscles and investigated the response properties of task-related units. 3. Of 125 afferent units tested with STAs, 68 showed PSF of EMG activity in at least one muscle. Fifty-nine DRG units provided sufficiently long recordings to generate averages with greater-than-or-equal-to 2,000 triggers, the minimum number considered to demonstrate reliable effects. Of these 59 units, 29 (49%) were associated with facilitation of forearm muscle activity. 4. Many STAs showed a gradual increase in EMG activity starting before or near the afferent trigger spike; often superimposed on this broad facilitation was a sharply rising PSF starting at a longer latency. The earliest poststimulus facilitation evoked by single microstimuli delivered in DRG occurred in stimulus-triggered averages at a latency of 3.5 ms. In STAs the broad facilitation beginning at latencies shorter than the responses to electrical stimulation was attributed to synchronous discharges in other afferent units. The sharper postspike EMG increases occurring with latencies of greater-than-or-equal-to 3.5 ms were identified as PSF produced by the afferent. The PSF parameters documented in this study were measured after subtracting the effects of synchrony facilitation. 5. PSF of EMG activity began at a mean latency of 5.8 +/- 0.3 (SE) ms and peaked at a mean latency of 7.5 +/- 0.3 (SE) ms. In previous studies, the PSFs from CM and RM cells had mean onset latencies of 6.3 and 5.6 ms, respectively, and mean peak latencies of 10.2 and 9.1 ms. 6. A measure of the PSF amplitude is the mean percent increase (MPI), defined as the increase of the PSF above its base measured as a percentage of the prespike baseline mean. Muscles facilitated by the 29 afferent units that produced PSF had an average MPI of 4.6 +/- 0.3% (SE). This is smaller than the MPIs reported for RM (5.1%) and CM (7.0%) cells (which were not similarly corrected for possible synchrony facilitations). 7. Units that facilitated EMG activity usually facilitated a subset of the coactivated muscles. The 29 adequately tested afferents facilitated an average of 46% of the synergistically coactivated muscles. This is comparable to the percentage of muscles facilitated by CM (40%) and RM (50%) cells. The average number of muscles facilitated per afferent was 2.6. 8. During the step-tracking task the monkeys generated ramp-and-hold torques about the wrist. The afferent units that showed PSF either increased their discharge during wrist flexion alone (10 units) or during extension alone (13 units), or were bidirectionally active (6 units). The most common response pattern, seen in 52% of units, was a tonic discharge. Twenty-one percent of the units had a phasic-tonic discharge, and 27% had a phasic discharge. The latter were often bidirectionally active. 9. The time between the first change in afferent unit activity and the onset of activity of its facilitated target muscles was calculated for 79 unit-muscle pairs from averages of isometric responses. Many afferent units began to discharge before the onset of their target muscle activity. On the average, afferent unit discharge began 52 +/- 13 (SE) ms before activation of the target muscle. The onset latencies of the phasic cells tended to be earliest (- 150 +/- 59 ms), followed by phasic-tonic (- 72 +/- 14 ms) and tonic cells (- 24 +/- 17 ms). In previous studies phasic CM and RM cells also tended to discharge earliest within their populations. 10. The effects of torque pulse perturbations, introduced during auxotonic contractions, were tested in seven units. In general, stretching the muscles facilitated by the afferent unit produced a burst in unit activity followed by an EMG burst, whereas shortening the target muscle produced a pause in afferent unit and EMG activity. 11. The results reveal some salient differences between the discharge properties of DRG afferents and supraspinal premotoneuronal (PreM) cells in motor cortex and red nucleus. All three PreM populations include tonic, phasic-tonic, and phasic cells but in significantly different proportions. The PreM afferents were typically active only with their target muscles, like CM cells and unlike RM cells. The hypothesis that these PreM afferents arose from muscle spindles is supported by their short-latency PSF, their responses to perturbations, and their onset before activity of their target muscles; the early activation would further suggest that gamma-motoneurons can activate spindles before alpha-motoneurons become active.
