Copyright © 1999 Massachusetts Institute of Technology
Color vision is our ability to distinguish and classify lights of different spectral distributions. The first requirement of color vision is the presence in the RETINA of different photoreceptors with different spectral absorbances. The second requirement is the presence in the retina of postreceptoral neuronal mechanisms that process receptor outputs to produce suitable chromatic signals to be sent to the VISUAL CORTEX. The third requirement is a central mechanism that transforms the incoming chromatic signals into the color space within which the normal observer maps his sensations. A good deal is known of the neurophysiology and anatomy of the first two requirements, but very little is known of the third.
The visible range of light extends over a wavelength band from 400 to 700 nanometers (nm), from violet to red. Color vision has evolved in a number of species, including bees, fish, and birds, but among mammals only primates appear to use color as a major tool in their processing of visual scenes. Most natural colors, as opposed to those in the laboratory, are spectrally broad-band. There are many systems for specifying their spectral composition, ranging from simply the physical distribution across the spectrum, through industry standards based on descriptions of spectral mixtures as functions of three variables (for the three different photopigments in the human eye), to color spaces that attempt to reproduce the way in which we perceptually order colors (Wyszecki and Stiles 1982). For a cognitive psychologist, there are two different aspects of color vision: we are not only very good at distinguishing and classifying colors (under optimal conditions wavelength differences of just a few nanometers can be detected), but also color vision is very important in visual segmentation and OBJECT RECOGNITION.
Humans and Old World primates possess three different photoreceptor, or cone, types in the retina. A fourth type of photoreceptor, rods, is concerned with vision at low light levels and plays little role in color vision. The idea of three different cone types in human retina was derived from the empirical finding that any color can be matched by a mixture of three others. It was first proposed by Thomas Young in 1801, and taken up by Hermann von HELMHOLTZ later in the nineteenth century. The Young-Helmholtz theory came to be generally accepted, and the spectral absorbances of the three photopigments were determined through psychophysical measurements. However, it is only in recent years that it has become possible to directly demonstrate the presence of the three cones, either by measuring their spectral absorptions (Bowmaker 1991) or their electrical responses (Baylor, Nunn, and Schnapf 1987). The spectral absorbances peak close to 430, 535, and 565 nm, and preferred designations are short (S), middle (M), or long (L) wavelength cones, rather than color names such as blue, green, and red. It is important to realize that a single cone cannot deconfound intensity and wavelength; in order to distinguish colors, it is necessary to compare the outputs of two or more cone types.
There has been much recent interest in the molecular genetics of the photopigments in the cones, especially in the reason why about 10 percent of the male population suffer from some degree of red-green color deficiency (Mollon 1997). The amino acids in the photopigment molecule that are responsible for pigment spectral tuning have been identified. A change in just a single amino acid can be detectable in an individual's color matches; it is rare for such a small molecular change to give rise to a measurable behavioral effect.
The spectral absorbances of the photopigments are broad, and this means that the signals from the different cones are highly correlated, especially with the broad-band spectra of the natural environment. This correlation gives rise to redundancy in the pattern of cone signals, and so postreceptoral mechanisms in the retina add or subtract cone outputs to code spectral distributions more efficiently, before signals are passed up the optic tract to the CEREBRAL CORTEX (Buchsbaum and Gottschalk 1983). Again, the first suggestion of such mechanisms came in the nineteenth century, from Oswald Hering. He proposed that there were black-white, red-green, and blue-yellow opponent processes in human vision, based on the impossibility of conceiving of reddish-green or bluish-yellow hues, whereas reddish-yellow is an acceptable color. Red, green, blue, and yellow are called the unique hues; it is possible to pick a wavelength that is uniquely yellow, without a reddish or greenish component.
Three main VISUAL PROCESSING STREAMS leave the primate retina to reach the cortex via the lateral geniculate nucleus. Each of them is associated with a very specific retinal circuitry, with signals being added or subtracted very soon after they have left the cones (Lee and Dacey 1997). One carries summed signals of the M- and L-cones; it is thought to form the basis of a luminance channel of PSYCHOPHYSICS and is heavily involved in flicker and MOTION, PERCEPTION OF. It originates in a specific ganglion cell class, the parasol cells, and passes through the magnocellular layers of the lateral geniculate on the way to the cortex. A second channel carries the difference signal between the M- and L-cones, and forms the basis of a red-green detection channel of psychophysics. It begins in the midget ganglion cells of the retina and passes through the parvocellular layers of the lateral geniculate. The third channel carries the difference signal between the S-cones and a sum of the other two, and it is the basis of a blue-yellow detection mechanism. It begins in the small bistratified ganglion cells and passes through intercalated layers in the lateral geniculate. These red-green and blue-yellow mechanisms account well for our ability to detect small differences between colors. Neurophysiology and anatomy of these peripheral chromatic mechanisms have been established in some detail (Kaplan, Lee, and Shapley 1990; Lee 1996). It is thought that the S-cone and blue-yellow system is phylogenetically more ancient than the red-green one, and it is present in most mammals (Mollon 1991). Separate M- and L-cones and the red-green mechanism have evolved only in primates.
The lateral geniculate nucleus is not thought to substantially modify color signals from the retina, but what happens in the cerebral cortex is uncertain. It is possible to determine the spectral sensitivity of the opponent processes of Hering (Hurvich 1981), and it turns out that our perceptual color opponent space has different spectral properties compared to the cone difference signals that leave the retina, and that we use to distinguish small color differences. Put another way, the unique hues do not map directly onto the cone-opponent signals emanating from the retina. Transformation of retinal signals to produce our perceptual, opponent color space can be modeled (Valberg and Seim 1991; DeValois and DeValois 1993), but how this occurs neurophysiologically is unknown; it is even uncertain whether we should look for single cells in the cortex that code for Hering's opponent processes or unique hues, or whether these are an emergent property of the cortical cell network (Mollon and Jordan 1997).
It is generally thought that a high concentration of color-specific neurons are found in the cytochrome oxidase blobs in primary visual cortex (see Merigan and Maunsell 1993 for review). Some of these cells resemble retinal ganglion cells in their spectral properties; others do not (Lennie, Krauskopf, and Sclar 1990). These color-specific signals are then passed on to the temporal visual processing stream. Color signals do not seem to flow into the parietal visual processing stream, which has much to do with motion perception and is dominated by the magnocellular pathway.
Many cognitive aspects of COLOR VISION emerge at the cortical level. For example, so-called surface colors depend on the context in which they are seen; the color brown depends on a given spectral composition being set in a brighter surround. If the same spectral composition were viewed in a black surround, it would appear yellowish. The limited range of colors that can be seen in isolation, in a dark surround, are called aperture colors. Another important emergent feature is color constancy, our ability to correctly identify the color of objects despite wide changes in the spectral composition of illumination, which of course changes the spectral reflectance of light from a surface. Color constancy is not perfect, and is a complex function of the spectral and spatial characteristics of a scene (Pokorny, Shevell, and Smith 1991). It has been proposed that color constancy emerges in a secondary visual cortical area, V4 (Zeki 1980, 1983), but this remains controversial. Lastly, color also plays an important role in many higher order visual functions, such as SURFACE PERCEPTION or object identification. The neurophysiological correlates in the cortex of these higher order color vision functions are unknown, and are likely to be very difficult to ascertain. It seems probable that many of these functions are distributed through cortical neuronal networks, rather than existing as specific and well-defined chromatic channels, as in the retina.
Baylor, D. A., B. J. Nunn, and J. L. Schnapf. (1987). Spectral sensitivity of cones of the monkey Macaca Fascicularis. Journal of Physiology 390:145-160.
Bowmaker, J. K. (1991). Visual pigments and color vision in primates. In A. Valberg and B. B. Lee, Eds., From Pigments to Perception. New York: Plenum, pp. 1-10.
Buchsbaum, G., and A. Gottschalk. (1983). Trichromacy, opponent colours coding and optimum colour information transmission in the retina. Proceedings of the Royal Society of London B 220:89-113.
DeValois, R. L., and K. K. DeValois. (1993). A multi-stage color model. Vision Research 33:1053-1065.
Hurvich, L. M. (1981). Color Vision. Sunderland, MA: Sinauer Associates.
Kaplan, E., B. B. Lee, and R. M. Shapley. (1990). New views of primate retinal function. Progress in Retinal Research 9:273-336.
Lee, B. B. (1996). Receptive fields in primate retina. Vision Research 36:631-644.
Lee, B. B., and D. M. Dacey. (1997). Structure and function in primate retina. In C.R. Cavonius, Ed., Color Vision Deficiencies XIII . Dordrecht, Holland: Kluwer, pp. 107-118.
Lennie, P., J. Krauskopf, and G. Sclar. (1990). Chromatic mechanisms in striate cortex of macaque. Journal of Neuroscience 10:649-669.
Merigan, W. H., and J. H. R. Maunsell. (1993). How parallel are the primate visual pathways? Annual Review of Neuroscience 16:369-402.
Mollon, J. D. (1991). Uses and evolutionary origins of primate color vision. In J. R. Cronly-Dillon and R. L. Gregory, Eds., Evolution of the Eye and Visual System. London: Macmillan, pp. 306-319.
Mollon, J. D. (1997). . . . aus dreyerly Arten von Membranen oder Molekülen: George Palmer's legacy. In C. R. Cavonius, Ed., Color Vision Deficiencies XIII . Dordrecht, Holland: Kluwer.
Mollon, J. D., and G. Jordan. (1997). On the nature of unique hues. In C. Dickinson, I. Murray, and D. Carden, Eds., John Dalton's Colour Vision Legacy. London: Taylor and Francis, pp. 381-392.
Pokorny, J., S. K. Shevell, and V. C. Smith. (1991). Colour appearance and colour constancy. In P. Gouras, Ed., Vision and Visual Dysfunction, vol. 6: The Perception of Colour. London: Macmillan, pp. 43-61.
Valberg, A., and T. Seim. (1991). On the physiological basis of higher color metrics. In A. Valberg and B. B. Lee, Eds., From Pigments to Perception. New York: Plenum, pp. 425-436.
Wyszecki, G., and W. S. Stiles. (1982). Color Science -- Concepts and Methods, Quantitative Data and Formulae. Second edition. New York: Wiley.
Zeki, S. (1980). The representation of colors in the cerbral cortex. Nature 284:412-418.
Zeki, S. (1983). Colour coding in the cerebral cortex: The reaction of cells in monkey visual cortex to wavelengths and colors. Neuroscience 9:741-765.
Copyright © 1999 Massachusetts Institute of Technology