Kinematic representation of imposed forearm movements by pericruciate neurons (areas 4 and 3a) in the awake cat.

In eight awake cats, elbow flexion movements were imposed by a computer-controlled torque motor using three different classes of angular displacement inputs: force step-load displacements; sinusoidal displacements; and constant-velocity ramp displacements. Microelectrode recordings were obtained from 309 pericruciate neurons in areas 4 and 3a. Average response histograms for single-unit activity coupled with computer simulation of the imposed movements have shown in a neuronal population (n = 81), selected for receptive fields that were directly related to elbow movements, that both the magnitude and temporal features of the responses can be characterized by the coefficients of a third-order differential equation describing the movement's angular kinematics (i.e., position, velocity, acceleration, and jerk). To compare the responses of different neurons the coefficients were normalized to the angular velocity coefficient, which was assigned a weighted value of 1.0. The neurons' average responses were "predictable" by the normalized coefficients regardless of the imposed movements' temporal characteristics. Two distinct and spatially separate pericruciate areas containing neurons that responded to the imposed forearm movements were located: 1) one within area 4 at the lateral extent of the cruciate sulcus, which contained neurons that responded with predominant jerk and acceleration coefficients, exhibited either cutaneous or deep receptive fields, and demonstrated low microstimulation current thresholds to activate forelimb muscles; 2) a second, more laterally located area near the 3a/4 border in the postsygmoid gyrus, which contained neurons that responded with predominant velocity coefficients, and comparatively small jerk acceleration, and position coefficients, exhibited either cutaneous or deep receptive fields, and demonstrated high microstimulation thresholds (greater than 20 microA). Due to the sensitivity of the higher derivatives to changes in motion, the relative magnitude and time course of the average firing probability of area 4 neurons with prominent acceleration and jerk coefficients were dominated by these kinematic features during the more rapidly imposed movements. The findings are in accord with a hypothesis proposing that motor cortical neurons in area 4 form a sufficient substrate for a "predictive" feedback organization, and may constitute an essential component of a system capable of regulating errors in angular joint movements despite the relatively long conduction delays and the slow time course of muscle tension production inherent to mammalian neuromuscular systems.