A Multi-Stage Neural Network Model for Human Color Vision

Charles Q. Wu

Abstract: The contemporary “Standard Model” for human color vision is a two-stage model: The first stage consists of three types of receptors at the retina of the eye, and the second stage consists of three opponent-color neural channels: Red-Green, Blue-Yellow, and Black-White. In this paper I call upon the phenomena of complementary afterimages and of the “flight of colors” to show that this model is not an adequate explanation for these color phenomena in specific and for color vision in general. To remedy the theoretical inadequacy of the “Standard Model”, I propose a neural stage for color complementarity directly corresponding to our color sensation. Mapping onto the anatomical organization of human visual system, I further suggest that layer 4C in the primary visual cortex is the neural substrate directly responsible for color complementarity in particular and for color appearance (that is, color consciousness) in general.