Abstract
Our ability to distinguish colors derives from some combination of responses of spectrally opponent and non-opponent neurons. Models of how such responses lead to sensations must satisfy constraints imposed by psychophysics. We have been using a specific form of magnitude estimation to specify color appearance and to relate these sensations to physiology. Color appearance of equi-luminant stimuli can be scaled by using the four necessary and sufficient hue terms Red, Yellow, Green, and Blue, plus saturation. The hue terms correspond to separate physiological mechanisms that determine the mutually exclusive sensations of R and G, and of Y and B; saturation corresponds to the relative magnitudes of responses of chromatic and achromatic mechanisms. We have used our scaling techniques to specify how appearance changes with viewing conditions, including: stimulus size, intensity, and duration; retinal locus; adaptation to different illuminants; and variations in appearance across the adult life-span and across the general population. Across these conditions, one hue sensation stands out as extremely robust: the redness that we see at the longer wavelengths. Its boundary, which corresponds to the spectral locus of unique yellow, is largely invariant. Physiologically, this boundary is set by the opposed inputs from L and M cones to the Red-Green mechanism. The stability of this boundary is most surprising, given the polymorphism of L and M cone types, their ratios in individuals, and their retinal distributions