Conditioning of single units in visual association cortex: cell-specific behavior within a small population.

These experiments examine the interrelationships between the activity of adjacent neurons during learning. Does learning depend on coherent behavior in a population of neurons or does it depend on particular neurons engaging in a particular activity at specific times? A second purpose was to examine specificity in response modification as a function of reinforcement contingency. Cells from visual association cortex of locally anesthetized, paralyzed cats and rabbits were studied with extracellular microelectrodes capable of recording single and multiunit activity, as well as local field potentials. Multiunit records were fractionated by amplitude "windows" discrimination. Pavlovian discriminative conditioning procedures were used to evaluate selective plasticity. Cells that were activated by at least two different visual stimuli were selected. Only one of the effective stimuli was paired with foot-shock (reinforcement). Of the 180 cells or cell clusters studied, 27% exhibited conditioned modification to the reinforced stimulus (CS+) and 19% changed their response pattern to the unreinforced stimulus (CS-). None of the well isolated cells showed conditioning to both CS+ and CS-. Thus, cellular plasticity was specific to reinforcement contingency. These results provide a first demonstration of reinforcement-dependent functional distinctiveness at the neuronal level. Some cells showed no alteration of response pattern despite a most prolonged conditioning procedure. Neighboring cells, responsive to the same stimuli, revealed increases or decreases in firing rate, selective changes in the latency or amplitude of single response peaks, or the appearance of one or more new peaks as a function of conditioning. Rarely did adjacent cells show the same type of alteration when alteration occurred; there was no general tendency toward coherent firing patterns as conditioning proceeded.