Visual cortical mechanisms underlying selectivity for direction and speed of moving images, for image length, and for image orientation will be studied in the cat. We will test the hypothesis that directional, length, and orientation properties depend on the same nonlinear suppressive interaction mechanism, but distributed differently in space and time for each property. We will measure, separately for each receptive field (RF) property, nonlinear interactions between laterally- and endwise-separated regions of receptive fields by using a new two-dimensional stimulus composed of small, optimally oriented bars, each element of which is modulated randomly and independently of all others. Two modeling techniques will be used to assess the nature of hypothesized suppressive interactions. 1.) We will investigate the connectivity of the visual system from the periphery to the level of the cortex by modeling it as a sequence, or cascade, of candidate linear and nonlinear transformations. System identification techniques will be used to infer the location and nature of these transformations. Thus isolated, each of the three RF mechanisms will be evaluated for dependence on the proposed suppressive mechanism. 2.) Analytic membrane- and psychophysical-movement models proposed by others will be compared with our physiological results by calculating model interactions for tests with our random stimulus. Finally, to continue our long-term goal to establish the structural basis for RF mechanisms, we will, when possible, fill measured cells with peroxidase (HRP), and correlate their synaptic morphology (at the electron microscopic level) with physiologically measured suppressive interactions and with proposed membrane mechanisms.
