To make thousand colors out of three receptors - how color vision works
To us, color vision seems to be the most important aspect of visual perception. This subjective impression reflects the vast information that we extract from a scenery by analyzing its colors. We would be handicapped in finding ripe red fruits within a jungle of green leaves or in reading traffic lights if we were only able to dist inguish light and dark. Because of the advantages of color vision, this visual capability developed early on in evolution in many animal groups. In the first part of this text, you will find some information about how color vision works; the second part deals with defective color vision. The retina of vertebrates hosts two different types of light-sensitive cells or photoreceptors - rods and cones. Rods owe their name to their elongated outer segment (see figure 1A). Rods are exquisitely light sensitive; they evolved for night vision. Rods are able to respond to single photons, for example coming from a faint star. Light adaptation allows rods also to respond to 10,000 times higher light intensities occuring in a night with a bright full moon or at dawn. Cones owe their name to the shape of their outer segment which reminds us of a pine tree's cone. Cones appeared earlier during evolution than rods. They are less light sensitive and adapt over a wider range of light intensities than rods. Cones evolved for vision in day light and, moreover, to allow color vision. To discriminate colors, the brain needs to compare the signals from at least two cone types with different sensitivity to light of a particular color. Most mammals use this simple two-cone system to discriminate colors. However, humans use three different types of cones for color discrimination. The cones in humans are either sensitive to light of long wavelength (L), middle wavelength (M), or short wavelength (S). The excitation spectra of rods and cones in humans are shown in figure 1B.
|
| |
|
Figure 1: The different photoreceptor types and their spectral sensitivity. (A) Rods and cones in the primate retina (adapted from: Kolb, Fernandez, Nelson. Anatomy and physiology of the retina. In Webvision: http://webvision.med.utah.edu). (B) Spectral sensitivity of human photoreceptors (adapted from: Dowling (1987). "The Retina: an approachable part of the brain." The Belknap Press of Harvard University Press, Cambridge). |
A simple model of color vision would be as follows: an object appears uniformly white, gray or black if the different cone types are excitated equally. If they become differently excitated, the object will appear colored. This simple hypothesis, however, is in conflict with the fact that the color that we perceive of an object is influenced by the color of the background, a phenomenon that is called simultaneous color contrast. You can easily test this background effect by inspecting the spots in figure 2.
|
| |
|
Figure 2: The color of an object depends on the background. |
Although the spectral composition of the light coming from the
three spots is the same, the spot appears gray on white background, reddish on blue
background, and bluish on red background.
Figure 3: A model for color vision (adapted
from: Gouras. Color vision. In Webvision:
http://webvision.med.utah.edu
But why is the mechanism that extracts color information so complicated? Why does the brain consider the background to attribute a color to an object? It is because the brain has to accomplish an extremely complicated task that we call color constancy: a ripe banana appears yellow to us, irrespective of whether we see it under the blue sky of a bright sunny day or in a room at candle light. We are so used to this visual performance that we take it as granted. But consider that the spectral composition of the light reflected by the banana differs dramatically between the two conditions: during daylight, the spectral composition of the reflected light is dominated by short wavelengths, under candle light it is dominated by long wavelengths. As described above, our brain compares the spectral composition of light reflected by the object with that of light reflected by the background. As changes in illumination change the spectral composition of light reflected by the object and by the background in the same way, this kind of comparision always reveals the same result, i.e. color constancy. Simultaneous color contrast is likely to be a corollary of color constancy.
No matter how good a model might explain the psychophysical aspects of color vision, it can only be valid if it is consistent with the neuronal processing of cone exitation. Indeed, cell physiologists have identified several cell types that process cone excitation consistent with the model shown in figure 3. So called color opponent ganglion cells in the retina process cone excitation consistent with Hering’s idea of a blue/yellow and a red/green channel. These cells receive input from cones that are organized in a center-surround receptive field. Certain ganglion cells receive excitatory input from red sensitive L-cones in the center and inhibitory input from green sensitive M-cones in the surround (red ON/green OFF). Other ganglion cells receive inhibitory input from L-cones in the center and excitatory input from M-cones in the surround (red OFF/green ON). Similarly, ganglion cells with either ON or OFF M-cone centers and antagonistic L-cone surrounds have been identified. The receptive fields of the red/green opponent ganglion cells are summarized in figure 4A.
|
Blue/yellow opponent ganglion cells have also been
identified, but in contrast to red/green opponent ganglion cells, their
antagonistic mechanisms seem to overlap. They receive input from blue sensitive
S-cones and mixed input from L- and M-cones that becomes maximal with yellow
light. The receptive fields of the blue/yellow opponent ganglion cells are
summarized in figure 4B.
The information of neighbouring ganglion cells of either the red/green opponent or the blue/yellow opponent type is combined by so called double-opponent cells that reside in the primary visual cortex.
| |
|
Figure 5: Pathway from the retina to the primary visual cortex in humans (adapted from: Hubel (1988). Eye, Brain and Vision. Scientific American Library, New York). |
These cells also have an antagonistic center-surround receptive field organization, but each cone type operates in all parts of the receptive field and has different actions (excitatory or inhibitory) in either the center or surround. For example, L-cones in the center of the receptive field of a double-opponent cell (responding to the object) are excitatory, and L-cones in the surround (responding to the background) are inhibitory. This inhibition is adequate to normalize the object's brightness relative to the background. Therefore, double-opponent cells might establish normalization required for color constancy and simultaneous color contrast. Figure 6 illustrates how the receptive field of a double-opponent cell might be organized.
| |
|
Figure 6: Putative organization of the receptive field of a double-opponent cell in the primary visual cortex of primates. |
Selective loss of color vision
Some people see the world only in shades of gray. The underlying disease, colorblindness, is either acquired through lesions within the visual cortex or it is inherited. Our research is concerned with the detailed mechanisms of inherited forms of colorblindness.
Genetic analysis identified mutations in several genes that have been associated with colorblindness. All these genes code for proteins that serve specific functions in cone photoreceptors. Therefore, in people with inherited colorblindness, color vision is impaired at the very first stage of vision, i.e. by the dysfunction of cones.
Defective color vision is a surprisingly common disase: about 8 % of all males and 0.5 % of all females are affected. But by far the majority of the affected people only have problems to discriminate red and green colors and, therefore, are not really colorblind. This type of color vision deficiency is inherited and is caused by mutations in the genes encoding the visual pigments of red sensitive L-cones (L-pigment) or green sensitive M-cones (M-pigment). The genes for the L-pigment and the M-pigment reside adjacent to each other on the X chromosome. It is the combination of close neighbourhood and high similarity of the two genes that leads to the high frequency of mutations in the L-pigment and the M-pigment genes. In germ line cells, a process called homologous recombination leads to the exchange of genetic material between the two copies of a chromosome (figure 7A) in order to increase the genetic variation in the offspring. Due to the high similarity between the L-pigment and the M-pigment gene, as well as between the flanking regions, in some instances this process leads to unequal recombination. Unequal recombination can result in the loss or the duplication of the M-pigment gene, or in hybrids formed of the L- and M-pigment gene (figure 7B).
| |
|
Figure 7: Recombination in the region of the L-pigment and M-pigment gene. (A) Equal recombination of two X chromosomes (X 1 and X 2) in the region of the L-pigment gene (red box) and the M-pigment gene (green box). (B) Unequal recombination causes mutant cone pigment genes or the duplication and loss of the M-pigment gene. |
Most mutations in the cone pigment genes are recessive, i.e. only if both copies - or alleles - of the L-pigment gene (or both alleles of the M-pigment gene) are mutated, L-cone function (or M-cone function) is impaired. Because females have two copies of the X chromosome and males only have one, the chance to carry at least one normal L- and M-pigment allele is larger for females. This explains the lower frequency of color vision deficiencies in the female population.
While the three cone types differ in their visual pigments, the other components of the photoelectrical transduction are identical for all cone types. Thus, mutations in genes coding for these elements affect all cone types, and, therefore, may lead to the complete loss of cone function and of color vision. Genetic analysis identified such mutations in people with inherited colorblindness, a disease also known as achromatopsia. The majority of these mutations have been identified in the genes that code for CNGA3 and CNGB3, the two subunits of the cyclic nucleotide-gated (CNG) channel in cones. CNG channels serve to generate the electrical signal in phototransduction. It is likely that colorblind people with mutations in the genes that code for CNGA3 or CNGB3 lack functional cone CNG channels and, therefore, lack functional cones.
Only one out of about 20,000 people suffers from achromatopsia; this equals a frequency of 0.005 %. Because mutations in the cone CNG channel genes are recessive and because the genes reside in chromosomes of which males and females have two copies, both genders are affected by achromatopsia with the same frequency.
Achromatopsia has obtained some popularity through the book The Island of the Colorblind by Oliver Sacks. Oliver Sacks describes his visit of Pingelap, an atoll in the West Pacific. About 5 % of the pingelapease people are colorblind. The high frequency of the disease has been traced back to 1775, when a typhoon decimated the population of the atoll, leaving only a handful of survivors for the repopulation. At least one of the survivors must have carried the mutation that is responsible for the disease. Recently, the genetic analysis of pingelapease people identified this mutation in the gene coding for CNGB3.
Because people with achromatopsia have no functional cones, they are left with the extremely light sensitive, low acuity rod system. As a consequence, they try to avoid bright light and their visual acuity is poor. Figure 8B might give an idea of an achromats perception.
|
| |
|
Figure 8: Putative perception of an achromat. (A) A scenery with colored balloons as it appears to a person with normal cone function, and (B) as it might appear to an achromat. |
last change 07.01.2004 | Thomas Gensch | Print