PREDICTIONS OF A RECURRENT MODEL OF ORIENTATION SELECTIVITY

Recurrent models of orientation selectivity in the visual cortex postu late that an initially broad tuning given by the pattern of geniculate afferents is substantially sharpened by intracortical feedback. We sh ow that these models can be tested on the basis of their predicted res ponses to certain visual stimuli, without the need for pharmacological or physiological manipulations. First, rye consider a detailed recurr ent model proposed by Somers, Nelson and Sur [(1995) Journal of Neuros cience, 15, 5448-5465] and show that it can be simplified to a single equation: a center-surround feedback filter in the orientation domain, Then, we explore the responses of the simplified model to stimuli con taining two or more orientations, We find that the model exhibits pecu liar responses to stimuli containing two orientations, such as plaids or crosses: if the component orientations differ by less than 45 deg t he model cannot distinguish between them; if the orientations differ b y more than 45 deg the model overestimates their angle by as much as 3 0 deg, Moreover, the model cannot signal the presence of three orienta tions separated by 60 deg (it responds as if there were only two orien tations), and the addition of two-dimensional visual noise to an orien ted stimulus results in strong spurious responses at the orthogonal or ientation, We argue that the effects of attraction and repulsion betwe en orientations and the emergence of responses at off-optimal orientat ions are common to a,vide class of feedback models of orientation sele ctivity, These models could thus be tested by measuring the visual res ponses of cortical neurons to stimuli containing multiple orientations . (C) 1997 Elsevier Science Ltd.