Correlations between activity of motor cortex cells and arm muscles during operantly conditioned response patterns.

Monkey motor cortex cells were recorded during isolated, isometric contractions of each of four representative arm muscles -- a flexor and extensor of wrist and elbow -- and comparable response averages computed. Most cells were coactivated with several of the muscles; some fired the same way with all four and others with none. Results suggest that many precentral cells have a higher order relation to muscles than motoneurons. Operantly reinforced bursts of cell activity were associated with coactivation of specific muscles, called the cell's "motor field"; the most strongly coactivated muscle was usually the one whose isolated contraction had evoked the most intense unit activity. During active elbow movements most cells fired in a manner consistent with their isometric patterns, but clear exceptions were noted. Differential reinforcement of unit activity and muscle suppression was invariably successful in dissociating correlations. The strength of each unit-muscle correlation was assessed by the relative intensity of their coactivation and its consistency under different response conditions. Several cells exhibited the most intense coactivation with the same muscle during all conditions. Thus, intensity and consistency criteria usually agreed, suggesting that strong correlations so determined may operationally define a "functional relation". However, correlations in the sense of covariation are neither necessary nor sufficient evidence to establish anatomical connections. To test the possibility of direct excitatory connections we stimulated the cortex, but found lowest threshold responses in distal muscles, even from points where most cells had been strongly correlated with proximal muscles. Post-spike averages of rectified EMG activity provided scant evidence for cell-related fluctuations in firing probabilities of any muscles.