The potential dynamic role of red nucleus in contact placing (CP) was investigated in intact awake cats. The activity of individual rubral neurons was recorded and the timing of their discharges was compared with electrical activity in selected contralateral forelimb muscles and with associated changes in joint angles characteristic of CP of the forelimb or hindlimb. The series included 273 rubral neurons (RN) that were recorded extracellularly in five cats; 62% were identified as rubrospinal tract neurons (RTN). Latencies of antidromic invasion following stimulation in the caudal contralateral medulla ranged from 0.25 to 3.3 ms. In some units, invasion occurred intermittently regardless of the stimulus strength above threshold. Most RN discharged tonically in the absence of obvious movement at, e.g., rates of 5-60 spikes/s. Spike train analyses were performed on 21 RN with stable resting discharge. Although the interspike interval distribution usually resembled that of a gamma process, the discharges were nonrandom. Usually, successive intervals were not independent, long and short intervals tending to be followed by intervals longer and shorter, respectively, than the mean interval. The sequence of intervals often shows episodes of relatively periodic discharge with mean interval differing between episodes. Among 120 RN that altered in firing rate during the first lifting-withdrawal phase of CP, 97 (81%) initially accelerated in rate. Variability in the delay between forepaw contact and the initial RN discharge permitted the distinction between contact-locked responses (53%), which began with a short latency, e.g., 20 ms, after contact and peaked usually within 50 ms. When CP failed to occur to a trial, contact-locked responses were further subdivided into a noncontingent type in which the RN response occurred regardless of whether CP occurred or not (type A) and a contingent type in which the RN response was significantly reduced or even absent when CP failed to occur (type B); delayed movement-locked responses (47%), which could be identified only when CP movements were delayed after contact. Delayed movement-locked responses were either types A or B, according to whether the RN response preceded (e.g., by 40-160 ms), or started during the electromyogram (EMG) response or the movement, respectively. Thus response types 1B and 2A can predict the success or failure of subsequent CP. Among an additional 17 RN initially responding after the paw cleared the top corner of the placing apparatus, the firing rate accelerated.(ABSTRACT TRUNCATED AT 400 WORDS)
Nonparametric estimation of characteristics of the interspike interval distribution
