Response patterns and force relations of monkey spinal interneurons during active wrist movement

Abstract

The activity of C6-T1 spinal cord neurons was recorded in three macaques while they generated isometric wrist flexion and extension torques in visually guided step-tracking tasks. Electromyographic activity (EMG) was recorded in </=12 independent forearm muscles. Spike-triggered averages (STAs) of rectified and unrectified EMG were used to classify neurons into four groups. Motoneurons (MNs) had a clear postspike motor unit signature in the unrectified STA of one muscle. Premotor interneurons (PreM-INs) had postspike effects in at least one muscle, with onset latencies of >/=3.5 ms from the trigger. Synchrony interneurons (Sy-INs) were non-PreM-Ins that had spike-related features with latencies <3.5 ms in at least one muscle. Unidentified interneurons (U-INs) showed no features in any of the STAs. A total of 572 task-related spinal neurons were studied; 29 cells were MNs, 97 PreM-INs, 32 Sy-INs, and 414 U-INs. MNs were activated predominantly in a tonic fashion during the ramp-and-hold torques and were active in one direction only. The most common response pattern for interneurons, irrespective of their class, was phasic-tonic activity, followed by purely tonic and purely phasic activity. Most interneurons (77%) were bidirectionally active in both flexion and extension. For all classes of interneurons, units with phasic response components tended to be activated first, before torque onset, followed by tonic units. The onset times of PreM-INs relative to onsets of their target muscles were distributed broadly, with a mean of -25 +/- 128 (SD) ms. For most neurons with tonic response components (all MNs, 71% of PreM-INs, 67% of Sy-INs, and 84% of U-INs), activity during the hold period was correlated significantly with the magnitude of static torque exerted by the monkey. The rate-torque regressions generally had positive slopes with higher mean slopes for extension than for flexion. The phasic response components were correlated significantly with rate of change of torque for a smaller percentage of tested PreM-Ins (50%), Sy-INs (83%), and U-INs (77%). In contrast to other premotor neurons [corticomotoneuronal (CM), rubromotoneuronal (RM), and dorsal root ganglion (DRG) afferents] previously characterized under similar conditions, a larger proportion of the spinal PreM-INs were activated after onset of their target muscles, probably reflecting a larger proportion of PreM-INs driven by peripheral input. The rate-torque slopes of PreM-INs tended to be less steep than those of CM and RM cells. Unlike the CM and DRG PreM afferents, which were activated unidirectionally, most spinal PreM-INs showed bidirectional activity, like RM cells.