BRIGHTNESS PERCEPTION, ILLUSORY CONTOURS, AND CORTICOGENICULATE FEEDBACK

A neural network model is developed to explain how visual thalamocorti cal interactions give rise to boundary percepts such as illusory conto urs and surface percepts such as filled-in brightnesses. Top-down feed back interactions are needed in addition to bottom-up feed-forward int eractions to simulate these data. One feedback loop is modeled between lateral geniculate nucleus (LGN) and cortical area V1, and another wi thin cortical areas V1 and V2. The first feedback loop realizes a matc hing process which enhances LGN cell activities that are consistent wi th those of active cortical cells, and suppresses LGN activities that are not. This corticogeniculate feedback, being endstopped and oriente d, also enhances LGN ON cell activations at the ends of thin dark line s, thereby leading to enhanced cortical brightness percepts when the l ines group into closed illusory contours. The second feedback loop gen erates boundary representations, including illusory contours, that coh erently bind distributed cortical features together. Brightness percep ts form within the surface representations through a diffusive filling -in process that is contained by resistive gating signals from the bou ndary representations. The model is used to simulate illusory contours and surface brightnesses induced by Ehrenstein disks, Kanizsa squares , Glass patterns, and cafe wall patterns in single contrast, reverse c ontrast, and mixed contrast configurations. These examples illustrate how boundary and surface mechanisms can generate percepts that are hig hly context-sensitive, including how illusory contours can be amodally recognized without being seen, how model simple cells in V1 respond p referentially to luminance discontinuities using inputs from both LGN ON and OFF cells, how model bipole cells in V2 with two colinear recep tive fields can help to complete curved illusory contours, how short-r ange simple cell groupings and long-range bipole cell groupings can so metimes generate different outcomes, and how model double-opponent, fi lling-in and boundary segmentation mechanisms in V4 interact to genera te surface brightness percepts in which filling-in of enhanced brightn ess and darkness can occur before the net brightness distribution is c omputed by double-opponent interactions.