Quick viewing(Text Mode)

Almost Reason Enough for Having Eyes

Almost Reason Enough for Having Eyes

Almost Reason Enough for Having

Jay Neitz, Joseph Carroll and Maureen Neitz

t is a commonly held misconception that dogs are blind. The truth is that a dog’s ability to see is similar to that of most other mammals,1 although different from that of humans with nor- I mal vision. Cross-species comparisons of differences in color vision have helped us appreciate our own ability to see colors, as well as to understand the underlying mechanisms of color vision. Among humans with normal color vision, there are differences in color that can also help us under- stand the fundamental processes of color vision.2,3 Over the past decade, comparisons and contrasts made at the cellular and molecular level among different color vision systems have brought us to a new under- standing of color vision, one that differs in many unexpected ways from the tenets of conventional wisdom. With this comes a new appreciation of the fact that wavelength sensing, in the form of color discrimination, is a fundamental visual capacity that serves to greatly expand the ’s information- gathering power. Color vision has played an important role in how the biological machinery for seeing has developed over the course of evolution.

26 Optics & Photonics News n January 2001 1047-6938/01/01/0026/8-$0015.00 ' Optical Society of America COLOR VISION

Neural mechanisms for seeing when the from the star stimu- zero input to maximum activity, in the color lates the photoreceptors. Activation of this white labeled line; sensations from dark The underlying machinery of a biological labeled line is responsible for the percep- grays to would correspond to graded system for seeing color is composed of two tion of white. Another line, the labeled line levels of activity in the black labeled line. stages: the first consists of light-sensitive for “blackness,” carries output collected , which is determined by the wave- receptors and the second consists of the from the very same photoreceptors. length content of colors, is the property neural components needed to process, However, this line derives from neural cir- that allows us to perceive colors as ranging partition, and encode information about cuits with physiologically different com- from through , and . wavelength that the photoreceptors col- ponents that invert the signal relative to In the human , the second lect. Our understanding of how visual the whiteness labeled line. Because the sig- information is encoded in humans has nal carried by the blackness line is sign- reversed, its activity is maximally silenced developed from studies of human percep- A short glossary of terms tion coupled with research on the physiol- when the star illuminates the photorecep- 4-7 tors. Conversely, the neural line labeled Percept: A mental impression of some- ogy and behavior of . The cen- thing perceived by the senses, viewed as tral idea is that, from the retinal image, “blackness” is activated when the contrast the basic component in the formation of neural elements extract the relative re- is reversed, for instance when we fixate on concepts; a sense datum. sponses of neighboring receptors and a printed black dot against the background Labeled line code: A coding mecha- process the information through a se- of a white page. Since one of the two out- nism used by the nervous system for stim- quence of stages to ultimately encode the put lines is inverted relative to the other ulus quality. Information in a neural mes- information using a system of “labeled (although both access the same photore- sage depends on the of nerve fibers (axons) that are active. The identity or lines.” ceptors), the activities of the two lines are label for an active line tells the "mean- The concept can be illustrated in the always opposed. The activity in these two ing" of the activity.Thus, for example, at a case of the percepts of parallel output lines is presumably respon- more basic level, increased discharge in that, at higher levels of visual processing, sible for the percepts not only of black and axons coming from the ear gives rise to are carried by a pair of labeled lines. white, but also of all of the intermediate the sensation of sound only; stimulation of the from the eye evokes a Consider the small cluster of photorecep- . There is not general agree- visual sensation only. Similarly, within the tors that are illuminated by a bright star as ment about how the activities of black and visual areas of the nerve fibers are we fixate on it: multiple neural output white labeled lines are related to percepts tuned, for example, so that red objects lines gather information from the illumi- of gray. However, one theory is that there increase discharge in one set of fibers and green objects in another set of fibers. It is nated receptors, as well as from their is no labeled line for gray. Rather, “gray” is the label on the active fibers that immediate neighbors. In the brain, one set the percept that occurs when neither line informs the brain that the sensation is one of nerve fibers that carry output from this is active. A series of sensations from light of redness or greenness. cluster of receptors constitute the labeled gray to white would correspond to line for “whiteness;” this line is activated increasing levels of activity, ranging from

Optics & Photonics News n January 2001 27 COLOR VISION stage for hue has two compo- However, the two systems diverge at the ed to the human S cone, the other to the nents8: each is responsible for a pair of sen- level of the cone photoreceptor output: human M and L cones. The photoreceptor sations, and just as in the case of the black- one system draws output from the S cones, basis of wavelength discrimination medi- white system, the sensations in each pair comparing it to the L plus M cone ated by this system is illustrated in Fig. 1. are opposed to one another. One hue sys- responses; the other extracts wavelength Over a wide band of the spectrum, each tem is responsible for perceiving the hue- information from a comparison of the rel- wavelength of light produces a unique pair blue and yellow; the other is responsi- ative L and M cone responses.9 The fact ratio of activation in the two different ble for the hue-pair red and green. Each of that only one system depends on S cones is photoreceptors. The neural circuits these two systems is believed to share significant because S cones differ from M responsible for color vision compute the properties with the black-white system: and L cones both in physiology and in reti- ratio of absorption of the two types of each presumably draws from a common nal distribution. We have also learned that cone to provide a continuum of wave- set of photoreceptors, but outputs the the underlying neural circuits of the sys- length information through the region information through different where the absorption spectra of the pig- neural components to consti- ments overlap. Figure 1 represents an tute the different labeled lines. attempt to illustrate the color world of The blue-yellow and red- creatures with dichromatic color vision. green systems are remarkably Wavelengths in which the relative absorp- different and separate. Togeth- tion by the S cones is highest are shown as er, the two systems extract in- appearing blue to the (by analogy formation used for color vision to the human blue-yellow system) and from the three traditionally rec- wavelengths in which the relative absorp- ognized classes of cone pho- tion by the other cone type (L in the dog) toreceptors: short-, middle-, is highest are shown as appearing yellow. and long-wavelength-sensitive, From the middle of the spectrum to the abbreviated S, M, and L. ends, changes in color are seen solely as changes in saturation (i.e., vividness of hue, or degree of difference from gray). In the middle there is a wavelength for which tems are quite different in design, opera- the outputs of the two cones are balanced tion,7 and vulnerabilities: the blue-yellow and the appearance is colorless (shown as system is more likely to succumb to gray in Fig.1), which corresponds to mini- insults, such as toxic exposure, eye disease mum saturation of both cone types. Thus, or a trauma such as retinal detachment; over an interval of about 100 nm, dichro- the red-green system, less susceptible to mats can discriminate among 20 to 30 gra- acquired disorders, is subject to an aston- dations of monochromatic light.1 Pre- ishing number of congenital defects that sumably, the neural circuit that compares rarely occur in the blue-yellow system. A the responses of the S and L cones to pro- comparative approach has provided clues duce an output with the sign “plus toward understanding these two systems S/minus L,” constitutes the neural sub- from an evolutionary perspective. In fact, strate for a labeled line for “blueness,” one with the exception of a few differences in which short-wavelength (blue) light driven by utility, in each case the differ- produces positive activity. The same ences between the systems can be inputs fed through a sign-reversing circuit explained by a surprisingly unique evolu- would constitute the line for “yellowness,” Figure 1. Absorption spectra of the two different tionary history. which is silenced by blue and activated by cone responsible for color vision in light at the other end of the spectrum. In dogs. The pigments have absorption maxima at 430 nm and 555 nm respectively and are commonly The blue-yellow physiological studies, the neural substrate referred to as short-wavelength sensitive (S) and color vision system for the “blueness” line has been well char- long-wavelength sensitive (L). The bar below the Among mammals, trichromatic color acterized in the , but interestingly, graphs illustrates how an equal energy spectrum vision is found only in humans and in a the sign-reversed “yellowness” line has might appear to a dichromat such as a dog.The short 7 wavelength end of the spectrum appears one hue, subset of the . Some mammals, proven more elusive. illustrated as blue, and wavelengths at the other end probably a minority, are monochromats of the spectrum appear as a different hue. Di-chro- that lack color vision.10 The most common Photopigments and their genes mats see only two different .Wavelengths toward form of color vision among mammals is Vision is initiated by the absorption of the middle of the spectrum vary in their saturation 10 (i.e., in their vividness of hue or in their degree of dif- , like that found in dogs, in light quanta by visual pigments within the ference from gray). A wavelength in the middle of the which only one of the two color vision sys- photoreceptors. These photopigments are spectrum would appear to a human with normal tems—the one homologous to the hu- composed of a (11-cis-reti- color vision but gray to a dichromat.Wavelengths at man “blue-yellow” system—is present. In nal) encased by, and covalently bound to, a the farthest extremes of the spectrum appear black because of the decline in absorption by the pho- mammals, this system makes use of two protein component (). In terrestrial topigments for those values. different photoreceptor types: one is relat- animals, the chromophore is the same for

28 Optics & Photonics News n January 2001 COLOR VISION all pigments but the vary, providing attribute the re- different surroundings for the chro- maining 48% ami- mophore that tune the absorption maxi- no acid difference ma of the to distinct spec- to the effects of the tral positions. All opsins are believed to genetic drift that have evolved from a single common ances- has occurred over tor11: they belong to an enormous family the eons since the of receptor molecules that may presum- original duplication ably be traceable in evolutionary origin to event. This degree a single progenitor. of difference can be Over the last 15 years, molecular genet- used to estimate the ic methods have been used to deduce the amount of time that amino acid sequences of the photopig- has elapsed since ment opsins from humans and a wide our ancient ances- 11-15 tors evolved two pigments. Figure 2. Calibration of the molecular clock for variety of animals. The deduced amino estimating times of divergence for photopigments. acid sequences are strikingly different for If molecular divergence of the pig- The estimated number of years (in millions) since the two pigment classes that underlie ments is our correlate of evolutionary each animal and man shared a common ancestor dichromatic color vision: over 50% of the time, then an estimate of how long ago the is plotted versus the percent difference between two pigments diverged could be made if the amino acid sequence of the animal s amino acids that make up each pigment and man s. The straight line is the best fitting are different. The two pigments are nearly we had a calibration for the molecular regression line fit to the data points. The yellow identical in their purpose except that in clock. One available data set that can be shaded area shows estimated limits for the regres- order to provide the basis for color vision, used is the sequences of the photopig- sion. It can be extrapolated from this that the S they have to be tuned to different regions ments contained in the rod photorecep- and L (or M) photopigments appeared as separate molecules over a billion years ago. This suggests of the spectrum. Producing a spectral sep- tors responsible for low light vision. These that the photopigment basis for color vision aration requires a relatively small number have been extensively studied in a wide appeared very early in evolution. of amino acid differences. We assume that variety of species,15 and presumably rod and cone pigments would be subject to the the high degree of difference is an indica- eyes.16 Although we don’t have direct evi- same evolutionary constraints. tion that, in evolutionary terms, the S and dence predating the fossil record, it is pre- In Fig. 2, the percentage identity be- L photopigments are not closely related. sumed that eyes or light-sensitive precur- tween human rhodopsin and sors to eyes have been present throughout from a variety of different species is plot- The evolution of color vision the history of the existence of free-moving ted against estimates of the time elapsed It is believed that humans evolved from an organisms that lived in environments with ancestor with only one type of photore- since each animal and humans shared a 22 18-21 sufficient available light. The important 16 common ancestor. The relationship is ceptor. As discussed above, color vision point is that it appears that the divergence relatively linear (Fig. 2) as would be systems work by comparing the relative of the S and L photopigments must have expected if the differences were indeed due number of photons absorbed by different been very nearly coincident with the to a genetic drift that had proceeded at a photopigments tuned to distinct regions appearance of the first eyes. It is logical constant rate. The best fitting regression of the spectrum. The three clases of mod- that a photosensing organ with only one line shows a change of about 4% per 100 ern human cone pigments are to receptor type must have emerged first, but million years. Extrapolation using this have evolved from the processes of gene the implication from examination of the 17 scale places the time elapsed since diver- duplication and divergence. Duplicated molecular data is that duplication and gence between the S and L/M cone pig- copies of an original cone pigment gene divergence of spectrally separate pho- ments at more than a billion years. would have been free to mutate separately topigments occurred virtually immediate- Allowing reasonable confidence limits for and diversify their absorption spectra, ly after emergence of a photosensing the extrapolation would place the likely providing the basis for color vision. organ. Adding the second pigment confers time of divergence somewhere in the Presumably it is the selective advantage the advantage of an extended spectral interval between 800 and 1100 million provided by color vision that has served to range, even without the neural circuitry years ago (MYA). The first appearance of maintain the separate S and L pigment for extracting wavelength information. the photoreceptive structures that were the genes in the genome. Not all the amino However, we argue here that the advan- precursors to the earliest eyes probably acid positions involved in spectral tuning tages to having color vision are great and appeared then. between S and L pigments have been iden- there would be strong selective pressure Even on the vast scale of evolution, tified; however, it is known that only seven favoring evolution of those circuits 800-1100 million years is a very long peri- amino acid changes are required to make posthaste. Thus, color vision, at least in od of time. Mammals first appeared about the 30 nm difference between M and L some crude form, may be as old as 220 MYA; the earliest land animals photopigments, and that just two changes vision itself. account for most of that shift. Extrapolat- crawled out of the sea 370 MYA; the earli- ing from this finding, we can predict that est vertebrates appeared about 495 MYA. approximately 6% difference in amino The fossil record doesn’t go back much The importance of color vision acid sequence would be required to make further than 600 million years, but some of There is a historical tendency to focus on the 100 nm shift between S and L. We the earliest fossils do show the presence of the visual system’s function as light detec-

Optics & Photonics News n January 2001 29 COLOR VISION

tiple pigment system even in the earliest organisms. We have to appreciate that wavelength sensing is as fundamental to vision as is light detection.

The red-green color vision system The estimated time of divergence for the S and L/M opsins suggests that a biological apparatus capable of wavelength discrimi- nation may have been among the earliest sensory systems to emerge. The human blue-yellow color vision system has been called ancient.25 Here, it is argued that its roots may be traceable to a wavelength Figure 3. The geometric expansion of visual the internal quality of objects. The exam- sensor in very early organisms dating back capacity that accompanies each additional spec- ple given most often is spectral reflectance a billion years. Interestingly, the same kind trally different photopigment. Creatures with only 24 one type of photopigment are color blind: their as a signal indicating the ripeness of fruit. of analysis indicates that the red-green sys- visual world is restricted to a number of distin- However, in both the natural and the man- tem is as new as the blue-yellow system is guishable steps of gray on the order of 102 (left made world there are myriad examples: old. In humans, the L and M photopig- panel). Adding a second spectral type of cone pho- the color change indicating when meat is ments are individually polymorphic but toreceptor and the appropriate neural connec- on average they differ by about 15 amino tions adds another dimension to vision, expanding cooked, when skin is sunburned, when the number of intensity and wavelength combina- there is blood in urine, or when newsprint acids. Using the same calibration as for the tions that can be discriminated to about 10,000 is old (one of us says his color vision tells S vs. L/M pigments, and subtracting seven (middle panel).Adding a third photopigment makes him that he is getting old as he looks at the amino acids known to be involved in the trichromatic color vision possible and expands the color of his hair in the mirror). spectral difference between L and M, number of different colors to over one million (right panel). The information contained in light leaves eight that might be attributable to from the world around us is encrypted in genetic drift (8/364 = 2%). At 4% per 100 the infinite combinations of spectral con- million years, the divergence of the L and tor and object detector. If we focus on the tent and intensity. How much information M genes would be estimated to have been detection function of the visual system, can be extracted is determined by the abil- about 50 MYA. We don’t want to place too color can be seen as nonessential, a luxury ity of the visual system to discriminate much confidence in the accuracy of this enjoyed by few species and one that was among the different combinations. At any estimate; however, it is consistent with added as a later refinement to the more given adaptation level, the can recent calculations that place the split highly evolved eyes of creatures adapted to discriminate nearly 200 different levels of between New and Old World primates at a diurnal lifestyle. In his comprehensive gray (Fig. 3, Panel 1), which is small con- about 60 MYA.19 The divergence of the L volume on the biology of the eye, Walls sidering the infinite number of possible and M pigments can be argued to have concluded for example that color vision information-carrying combinations of occurred after that split since New World was rare among the mammals, although wavelength and intensity. The presence of primates usually have only one gene he recognized its presence among modern color vision tremendously expands the encoding a photopigment in the middle- 23 diurnal birds and fish. amount of extractable information. to-long wavelength range, while our near- Of course, the information embodied Consider the difference when just one er relatives, the Old World monkeys, have in the light that reaches our eyes reveals far photopigment is added, such as when we both L and M genes.10 more than the presence of objects: when compare vision to The addition of the third cone pigment we consider the intensity of light, its wave- dichromatic color vision. Starting with gene was a required step in achieving a length content, and the pattern and distri- one photopigment and adding a second functional red-green color vision system. bution of both, the amount of informa- spectral type does not just add a color, it From the standpoint of being able to tion light conveys is truly immense. For adds an entire dimension of vision. For the extract the information encoded in the example, the wavelength content of the dichromat, changes in wavelength pro- wavelength content of light, the addition available light changes with the time of duce up to 50 discernible steps. However, of another pair of neuronal lines in paral- day, the time of year, and the weather, and in the visual system, since black-white and lel with the black-white and blue-yellow thus carries information about each. blue-yellow are carried in separate parallel lines represents an enormous gain. Wavelength content changes with direc- systems, for each gray level there will be that since the lines are added in parallel, tion: it is different toward the horizon nearly 50 different possibilities on the the addition of each pair expands the compared to overhead or toward the blue-yellow scale. Thus, adding the second number of discriminable wavelength com- ground, and thus carries information pigment and the appropriate processing binations geometrically. Humans can dis- about position and orientation in space. machinery increases the discriminable tinguish close to 100 steps of spectral For aquatic animals, wavelength content combinations geometrically from near 200 change contributed by the activity of the changes as a function of water depth. to near 10,000 steps (Fig. 3, Panel 2). It is redness and greenness labeled lines. Reflected wavelength content carries a little wonder that there would have been a Multiply that times the approximately remarkable amount of information about strong evolutionary push to adopt a mul- 10,000 colors that can be distinguished

30 Optics & Photonics News n January 2001 COLOR VISION using the combination of the other sys- vertebrates have more advanced color evolution of color vision.29 Diurnal pri- tems, and the addition of the red-green vision than mammals is that the earliest mates have highly acute spatial vision. The system boosts the number of “colors” we mammals appeared at the height of the high spatial resolution is served by a reti- can see to upwards of one million (Fig. 3, dinosaurs’ dominion of the earth; the nal pathway that makes use of neurons Panel 3).26 mammals found a niche available in being with unusually small receptive fields, The geometric expansion of visual nocturnal, and they developed a retina called “midgets,” each of which directly capacity that accompanies each added that was highly adapted for vision under contacts a single cone. Through connec- cone receptor type is of inestimable signif- very low light levels.23 In icance. To fully appreciate it, one need look optimizing the eye for night no further than a comparison of normal vision, two of the four and defective color vision in humans. photopigment genes found Congenital red-green color vision defects among the other vertebrate are extremely common, affecting about lines were apparently lost in 8% of males and 0.4% of females in the mammals. Modern birds, rep- United States.27 About 25% of the people tiles, and fish enjoy color with red-green color vision defects are vision that is rooted in a dichromats; most of them suffer from multi-cone color vision sys- what is essentially a reversal of the gene tem that predates the diver- duplication that was originally responsible gence of those vertebrate for allowing . A gene classes. The nocturnal ances- deletion has left the human dichromat tors of modern primates were with a single photopigment gene on the X- reduced to dichromacy, and . Human dichromats are the blue-yellow system is the commonly referred to as “color blind,” only color vision machinery although their blue-yellow system is they share with other verte- intact. They enjoy color vision similar to brates. Primates achieved that of many of our mammalian relatives, trichromacy by inventing it but far more limited than that of trichro- separately. mats: compared to the trichromat the dichromat’s palette is estimated to be Unique processes create “missing” nearly one million colors. As the neural circuits for would be expected, this loss disadvantages red-green color vision the dichromat in many everyday tasks; An obvious question at this selected examples are illustrated in Fig. 4. point is why trichromatic color vision has evolved only Beyond trichromacy in the primates when other The ability to distinguish two million col- mammals have adopted diur- ors may seem impressive, but it is still nal lifestyles. If added dimen- small compared to the infinite combina- sions of color vision confer tions of wavelength and intensity in our such a tremendous advantage, environment. By extrapolation from the why wouldn’t all diurnal dramatic expansion in discrimination mammals have evolved sever- capacity accompanying the evolution al higher dimensions of color from dichromacy to trichromacy, we can vision? Diurnal mammals guess that the geometrical increase in dis- probably emerged within tinguishable combinations would also be about the last 100-150 million impressive if an organism had four cone years. This is a very short peri- pigments wired in a system to provide a od of time in which to evolve fourth dimension of color vision. Perhaps a higher dimension of color to a tetrachromat, a mere trichromat vision, particularly if many would seem as color blind as a dichromat statistically unlikely muta- Figure 4. Many real-world tasks are trivial for a trichromat but does to a trichromat. It has become tional changes are required to impossible for a person with a color vision defect (a dichromat). increasingly apparent over the last 20 years develop the neural circuits The top three pairs of panels are images that have been digitally that most non-mammal diurnal verte- required to process and parti- altered to simulate the visual world of a dichromat (a human brates, such as birds and fish, have four tion color information. deuteranope). Below each pair of pictures is a question for which cone pigments. They have an added pig- Primates have unique reti- the answer is either A or B. Individuals with normal color vision should take the test without referring to the bottom three pairs ment that is sensitive to light. nal features, evolved for differ- ent purposes, that may have of panels (the original, unaltered photographs).The questions are Some species have been demonstrated to trivially easy for a trichromat (the correct answers are B,A, and 28 afforded a shortcut unavail- achieve tetrachromatic color vision. The B). These examples illustrate the importance of color vision in explanation usually given for why other able to other mammals for the everyday life and the difficulties associated with .

Optics & Photonics News n January 2001 31 COLOR VISION tions that presumably evolved to amplify with the consensus of activity may be ants are expressed in separate populations spatial contrast, a signal of opposite sign is selectively weakened or lost. It is possible of cones, making it certain that some collected from the cone’s surrounding that the nervous system could use a simi- females do indeed have four spectrally dif- neighbors. If a new spectral class of cone lar preexisting capacity for “neural learn- ferent cone types and the photoreceptor were added to the retina, the output col- ing” to extract information from our chro- basis for tetrachromatic color vision. If the lected by the midget system would auto- matic experience to mold the higher corti- mechanisms described above are responsi- matically be one that compared responses cal circuits for red-green color vision.5 ble for creating the neural circuits for a from spectrally different cone types, exact- third dimension of primate color vision, ly as is necessary for color vision. If the then theoretically they should work the newly added cone type were randomly same way to provide a fourth dimension interspersed in the photoreceptor mosaic, of color vision for women heterozygous any given cone of the new type would have In the future… for the L pigment gene. Is it true that some some probability of having all of its neigh- women have super color vision? From the bors be of the opposing type. Thus, for it may be possible to description of the geometric increase in example, if the central cone were M and use gene therapy to capacity that comes from an additional the surrounding cones L, a circuit for red- replace missing pigment one might think that if such green color vision would be created women were among us, we would surely serendipitously, without a delay stemming photopigments in know about it. However, there are reasons from the evolution of genetically specific the eyes of color this might not be the case. One is that the neurons to wire the appropriate connec- most common variants of the L pigment tions in the retina. Even in cases in which blind humans. are only separated by a few nanometers; the surrounding cones were mixed in type, color vision based on such small spectral the averaged signal from the surround differences would produce a limited num- would be different enough from the cen- ber of additional distinguishable steps. tral cone to provide a usable chromatic Alternatively, it is possible that such small signal. In contrast, in mammals other than Directions in color vision differences may not be enough to drive primates most neurons responsible for research cortical neural mechanisms when spatial contrast connect to a larger num- Although vision scientists have yet to pro- they have to compete with the existing ber of cones. Random connections that duce definitive evidence for the wiring of well-separated L vs. M inputs. There is a collect from a large number of cones color vision being done by the oppor- chance that in some women, the two L would produce very similar average tunistic use of chance connection and pigment variants will be separated by responses for both center and surround, neural learning, several very exciting lines about 10-12 nm. These women are the and spectrally opponent cells would not be of research related to these possibilities are best candidates for having true tetrachro- likely to arise by chance. underway. These experiments take advan- matic color vision, and technology is now In primates, it seems plausible that evo- tage of naturally occurring large variations available that would allow them to be lution exploited existing circuits for spatial in the genes for the human L and M pig- identified by direct examination of their analysis to create the first stage of color ments. photopigment genes. There are some processing that occurs in the retina. But Evolution is opportunistic. The gene women who claim to have a superior color how is the partitioning maintained and duplication that eventually gave rise to the sense, and to be able to make color dis- the processing extended at higher visual human L and M photopigment genes tinctions lost to other people. It would be centers in the cortex? One possibility is placed the duplicated gene adjacent to the an easy matter to examine the genes of that once again the newly evolved primate original, producing a highly unstable these women to see if they encode spec- red-green color vision system may have genetic arrangement. If the evolutionary trally well-separated L pigments. Jordan taken advantage of pre-existing properties process had been more long-sighted, this and Mollon,35 in a search for an extra of the nervous system in order to avoid the probably would not have been the chosen dimension of color vision among female necessity of creating genetically specified arrangement. The configuration allows carriers of color vision defects, a group circuits that would exclusively recognize the L and M genes to misalign during that is expected to have a high frequency of and connect to outputs carrying specific meiosis and recombine, consequently tetrachomats, found one woman in a sam- chromatic information. The circuits of the intermixing their sequences. This has ple of 31 who behaved as would be pre- mammalian are known to be formed chimeric genes,31 which can pro- dicted if she were a tetrachromat. It would molded by visual experience. This neural duce photopigments with absorption be interesting to know if her genes predict- plasticity has been best studied in its role spectra that are intermediate between ed the presence of pigment variants that in the development of circuits responsible those of the original L and M genes. In the are particularly well separated spectrally. for coordinating input from the two eyes.30 normal population, variants are particu- There is amazing variation in the ratio The principle is that several synaptic larly common for the L gene.32-34 Thus of L to M cones in the human retina.36 inputs might initially converge on a single females, who have two X , This was demonstrated recently in dra- neuron in the cortex, but that over time can have genes encoding spectrally differ- matic fashion when adaptive optics was the synapses are strengthened for those ent L pigments on each X-chromosome. used to obtain the first images of the cone inputs that are concurrently active, while The process of X-chromosome inactiva- mosaic in the living human eye.37 One those inputs that are not well correlated tion insures that the two L pigment vari- might predict that these differences in reti-

32 Optics & Photonics News n January 2001 COLOR VISION nal architecture would lead to dramatic color blind humans. Furthermore, if the (Cambridge University Press, New York,1999). differences in color vision. However, two neural circuits can handle even higher 18. A. Janke, X. Xu and U.Arnason, The complete people who had very different L/M cone dimensions of color vision that could mitochondrial genome of the wallaroo (Macropus ratios (one was 1:1 and the other 4:1) did come from artificially adding a fourth robustus) and the phylogenetic relationship among Monotremata, Marsupialia, and , not have measurably different color vision cone type, it is possible that gene therapy Proceedings of the National Academy of Sciences for parameters predicted to be most affect- could also be used to extend normal USA 94, 1276-81 (1997). 38 19. U.Arnason,A. Gullberg and A. Janke, Molecular ed by the ratio difference. One explana- human color vision. From witnessing how timing of primate divergences as estimated by two tion offered was that plasticity has allowed strongly people are driven to have a mon- nonprimate calibration points, Journal of Molecular the nervous system to use information itor that can output the highest amount of Evolution 47, 718-27 (1998). 20. R.M. Nowak,Walker s mammals of the world (The gathered from experience to make com- color information, we expect that if there Johns Hopkins University Press, Baltimore, MD, pensating changes for the differences in were not associated risks, a therapy for 1999). cone ratio. This lends credence to the idea color blindness would be widely adopted. 21. C.Tudge,The Variety of Life (Oxford University Press, New York,2000). that neural plasticity is responsible for set- Would trichromats have their vision 22. J.R. Cronly-Dillon, in and Visual ting up the cortical connections for color expanded to tetrachromacy if a safe proce- System, J.R. Cronly-Dillon and R.L. Gregory, eds. 15- 53 (CRC Press, Inc., Boca Raton, FL, 1991). vision. In some of the best-studied sys- dure were readily available? 23. G. L.Walls,The vertebrate eye and its adaptive radi- tems, neural plasticity disappears after a ation (The Cranbrook Institute of Science, critical period during early childhood. Acknowledgment Bloomfield Hills, MI, 1942). 24. P.Sumner and J. D. Mollon, as signal However, in recent years there has been This work was supported by NIH grants of ripeness in fruits taken by primates, Journal of growing evidence for plasticity of the adult EY09303, EY09620 and EY01921, and Experimental Biology 203, 1987-2000 (2000). cortex. If the circuits responsible for red- 25. J. D. Mollon, Tho she kneel d in that Place where Research to Prevent Blindness. We would they grew... .The uses and origins of primate colour green color vision are plastic, is it possible like to thank P. M. Summerfelt for her vision, Journal of Experimental Biology 146, 21-38 that they could be among those that invaluable help in preparing the manu- (1989). 26. J.D. Mollon, Color vision: Opsins and options, remain plastic throughout life? Recently, script and figures. Proceedings of the National Academy of Sciences Yamauchi et al. presented evidence in fa- USA 96, 4743-5 (1999). vor of that theory.39 They reported long- 27. J. Pokorny,V.C. Smith, G.Verriest and A.J.L. Pinkers, References Congenital and Acquired Color Vision Defects term changes in color vision in subjects 1. J. Neitz,T. Geist, and G.H. Jacobs, Color vision in (Grune and Stratton, New York,1979). who had chromatically altered their visual the dog, Visual Neuroscience 3, 119-25 (1989). 28. C. Neumeyer, in Evolution of the Eye and Visual environment for relatively short periods of 2. J. Neitz and G.H. Jacobs, of the System (J.R. Cronly-Dillon and R.L. Gregory, eds.) long-wavelength cone in normal human color 284-305 (CRC Press, Inc., Boca Raton, FL, 1991). time. vision, Nature 323, 623-5 (1986). 29. P.Lennie, P.W. Haake, and D.R.Williams, in 3. G. Jordan and J.D. Mollon, Rayleigh matches and Computational Models of Visual Processing (M.S. unique green, Vision Research 35, 613-20 (1995). Landy and J.A. Movshon, eds.) 71-82 (MIT Press, The future of technology 4. R. L. DeValois, I.Abramov and G. H. Jacobs, Analysis Cambridge, 1991). and color vision of response patterns of LGN cells, J. Opt. Soc.Am. 30. D.H. Hubel, Eye, brain, and vision (W.H. Freeman 56, 966-77 (1966). and Company, New York,1988). The development of color monitors for 5. R. L. DeValois and K.K. DeValois, A multi-stage 31. M. Neitz and J. Neitz, Molecular genetics of color computers has recapitulated the evolution color mode, Vision Research 33, 1053-65 (1993). vision and color vision defects, Archives of of human color vision. Monitors were 6. P.Lennie in Principles of Neural Science, E.R. Ophthalmology 118, 691-700 (2000). Kandell, J.H. Schwartz, and T.M. Jessell, eds., 572-89 32. J.Winderickx, L. Battisti,Y. Hibibya,A.G. Motulsky originally . Although they (McGraw-Hill, 2000). and S.S. Deeb, Haplotype diversity in the human did not evolve through a two phosphor 7. D.M. Dacey, Parallel pathways for spectral coding in red and green opsin genes: evidence for frequent stage that would have been the exact primate retina, Annual Review of Neuroscience sequence exchange in exon 3, Human Molecular 23, 743-75 (2000). Genetics 2, 1413-21 (1993). equivalent of dichromacy, they did evolve 8. L.M. Hurvich and D. Jameson, An opponent 33. S. Sjoberg, in Cellular Biology, Neurobiology & through stages in which the number of process theory of color vision, Psychological Anatomy 143 (Medical College of Wisconsin, Review 64, 384-404 (1957). Milwaukee, 1998). different colors was very limited. Today, 9. P.K.Kaiser and R.M. Boynton, Human Color Vision 34. M. Neitz, S. Balding, S. Sjoberg, and J. Neitz, the most highly evolved monitors, capable (Optical Society of America,Washington, D.C., 1996). Topography of L & M cone pigment gene expres- of simultaneously producing millions of 10. G.H. Jacobs, The distribution and nature of colour sion in human retina, (to be published). vision among the mammals, Biological Reviews 68, 35. G. Jordan and J.D. Mollon, A study of women het- colors, are matched with human color 413-71 (1993). erozygous for colour deficiencies, Vision Research vision capacity. Evolution can occur 11. J. Nathans, D.Thomas, and D.S. Hogness, Molecular 33, 1495-508 (1993). quickly when a change is highly favored. genetics of human color vision: the genes encoding 36. J. Carroll, C. McMahon, M. Neitz and J. Neitz, blue, green, and red pigments, Science 232, 193- Flicker-photometric electroretinogram estimates In our survey of the local computer mon- 202 (1986). of L:M cone photoreceptor ratio in men with pho- itor population, all are of the millions-of- 12. M. Neitz, J. Neitz, and G.H. Jacobs, Spectral tuning topigment spectra derived from genetics, J. Opt. of pigments underlying red-green color vision, Soc.Am.A 17, 499-509 (2000). colors type. We might infer from the stun- Science 252, 971-4 (1991). 37. A. Roorda and D.R.Williams, The arrangement of ningly rapid change in the monitor popu- 13. S.S. Deeb,A.L. Jorgensen, L. Battisti, L. Iwasaki, and three cone classes in the living human eye, Nature lation that in the modern world, a com- A.G. Motulsky, Sequence divergence of the red and 397, 520-2 (1999). green visual pigments in great apes and humans, 38. D. Brainard et al., Functional consequences of the puter monitor’s ability to produce mil- Proceedings of the National Academy of Sciences relative numbers of L and M cones, J. Opt. Soc. lions of colors has an extremely high USA 91, 7262-6 (1994). Am.A 17, 607-14 (2000). adaptive significance. It is possible that in 14. D.J. Hunt, et al., Molecular evolution of trichromacy 39. Y.Yamauchi et al., Is unique yellow determined by in primates, Vision Research 38, 3299-306 (1998). the relative numbers of L and M cones?, the future, technology may be available to 15. S.Yokoyama, Molecular evolution of vertebrate Investigative Ophthalmology and Visual Science 41, do for the eye what has been done for visual pigments, Progress in Retinal and Eye S526 (2000). Research 19, 385-419 (2000). color monitors. If the neural circuits for 16. C.W.Oyster,The Human Eye: Structure and Function color vision are sufficiently plastic, it may (Sinauer Associates, Inc., Sunderland, MA, 1999). Jay Neitz, Joseph Carroll and Maureen Neitz 17. L.T. Sharpe,A. Stockman, H. Jagle and J. Nathans in work in the Department of Ophthalmology and in be possible to use gene therapy to replace Color Vision:From Genes to Perception (K.R. the Department of Cell Biology, Neurobiology missing photopigments in the eyes of Gegenfurtner and L.T. Sharpe, eds.) 3-52 and Anatomy at the Medical College of Wisconsin in Milwaukee, Wisconsin. Jay Neitz s e-mail

Optics & Photonics News n January 2001 33

Top View