DIMENSIONS AND PROPERTIES OF END-ZONE INHIBITORY AREAS IN RECEPTIVE-FIELDS OF HYPERCOMPLEX CELLS IN CAT STRIATE CORTEX

Subregions in the receptive fields of hypercomplex cells have been examined by a variety of quantitative methods with particular reference to the dimensions and properties of the end-zone inhibitory areas. These data have made it possible to construct detailed maps of the receptive-field organization of the two types of hypercomplex cell (I and II). The spatial extent of the end-zone inhibitory area is much greater than that responsible for discharge-region excitation. End-zone inhibition is, however, position dependent, the part of the area causing maximal inhibition lying precisely along the line of the most responsive part of the discharge region and just beyond its lateral border. Spatial summation of end-zone inhibition takes place along the line of its optimal stimulus orientation. Some simple and complex cells may have hypercomplex-type length-response curves in the nonpreferred direction of stimulus movement and vice versa for some hypercomplex cells. Whether these response patterns are due to the presence of direction-selective end-zone inhibition or not remains to be determined. While end-zone inhibition may be direction selective, it appears that it is usually nondirectional. Even when discharge region excitation is itself completely direction selective, the end-zone inhibition may be equally effective in both directions. Hence end-zone inhibition appears to be independent of the mechanism responsible for the direction selectivity of the discharge region. End-zone inhibition is stimulus orientation dependent, being maximal when the orientation is the same as the orientation that is optimal for the discharge region. When the stimulus is rotated away from the optimal, the strength of the inhibition progressively declines, falling to zero at 90° to the optimal. This property distinguishes end-zone inhibition from side band inhibition since the latter is not orientation sensitive. There may be considerable, or even total, spatial overlap between discharge-region excitation and end-zone inhibition, the spatial summation required for excitation being much less that that required to produce an inhibitory effect. The onset of inhibition on the length-response curve indicates that the effects of the spatial summation of inhibition now exceeds those of discharge-region excitation.