Does orientation domain inhibition play a role in visual cortex plasticity?

Abstract

Does visual input play an instructive role in refining cortical function, or does it merely permit genetically programmed maturation to run its course? It has been suggested that normal stimulus input, which is well-matched for the two eyes, ensures that binocular units mature with well-matched stimulus preferences (Pettigrew, 1974); lack of correlated input certainly destroys matched preferences for orientation (Blakemore and Van Sluyters, 1974, esp. Fig. 4). In this note I consider the goodness of optimal stimulus orientation match achieved by binocularly activated cells in the normal adult cat. One would expect that a neuron which responds vigorously to input from either eye, and which responds very selectively to stimulus orientation, would benefit more from visual input and so achieve a closer match between orientation preferences for inputs from the two eyes than would a cell only poorly excited by one eye or tuned only very broadly. This expectation is not realized. I suggest an hypothesis which still permits an instructive role for visual input. The data, collected for other purposes (Nelson et al., 1977), seemed sufficiently interesting to warrant more detailed consideration here. Data were obtained from 74 cells in visual area 17 in anesthetized and paralyzed cats; 92 % of the cells had receptive fields within 10 ~ of the projected position of the area centralis. Orientation tuning was assessed at essentially the same time for each eye using a multihistogram method (Henry et al., 1973). Curves were typically based on 24 orientations (taken in steps of 3 ~ for simple family cells or 6 ~ for complex family), each tested more than 20 times (for further details, see Nelson et al., 1977). Following mild Gaussian smoothing and spline interpolation, several measures were obtained from these curves. Binocularity is defined as 2N/ (N+D) , where N and D are the peak responses of the non-dominant and dominant eyes. A class IV unit (Hubel and Wiesel, 1962) would be 100% binocular. Units with only an inhibitory contribution or no response at all from one eye (class I or VII) were not included in the study. Mean half-width at half-height is the half-width at half-height above spontaneous activity averaged for the two eyes' responses. Orientation disparity is given by the difference between the peak orientation of the two curves