Multimicroelectrode investigation of monkey striate cortex: spike train correlations in the infragranular layers

In the infragranular layers of the striate cortex of three monkeys, we studied tangential neuronal interactions by analyzing cross-correlograms calculated from spike trains recorded with 30 closely spaced microelectrodes. 2. There are two major types of correlogram structures--"narrow" peaks a few milliseconds wide, sometimes accompanied by small lateral troughs, and "broad" peaks approximately 30- to 100-ms wide. Isolated troughs are rare. Both types of structures are superimposed in the same correlograms; they are not due to shared optical stimulation. 3. In layer VI, narrow peaks are largest in a short lateral range of approximately 220 micron, and they depend on ocularity. In layer V, the lateral range is greater, and the dependency on ocularity is weak. 4. In addition, narrow peaks are largest at distances of 160 micron if the angles of preferred orientation are similar. In layer VI, however, at tangential distances of 300-400 micron, peaks are smaller, and troughs are found more often, for neuron pairs with parallel orientations compared with those with orthogonal orientations. From the agreement of this finding with a cooperative theory, we conclude that orientation selectivity is shaped by collective interactions. 5. Broad peaks always depend on ocularity, and the associated lateral interaction range exceeds the maximum of 1 mm investigated. Their size sharply decreases with receptive-field distance. 6. Average mutual delays of spikes of neuron pairs, manifest as lateral displacements of broad peaks, are interdependent; the delay between neurons 1 and 3 is the sum of that of neurons 1 and 2 and of neurons 2 and 3. This feature permits to rank the neurons on a "delay scale." 7. We conclude from 5 and 6 above that broad peaks partly result from intraretinal interactions whose effects are transmitted to the cortex via slow and fast pathways. 8. Lateral troughs adjacent to narrow peaks provide evidence that neurons at the "slow" end of the delay scale inhibit those at the "fast" end, and to a lesser extent, nondirectional neurons inhibit directional ones.