Maureen Neitz,
Daniel G. Green,
Jay Neitz
We so readily use our eyes to organize and process information, making it is easy to ignore the truly remarkable adjustments the eye makes to enable us to see. When our interest focuses on an object, such as a colorful bird, both eyes are directed so that an image of the bird is focused on the foveas of both the retinas. These two planar projections of the object are encoded and transformed in the retina into elaborate patterns of neural activity and relayed to various targets in the brain. As we orient and move around, the retinal projection of an object, its distance from us, its spectral content, and its intensity vary, and yet we continue to see the object as having the same shape, size, color, and brightness.
The eye often is compared to a camera, and because each shares common functions, this can be a useful starting point for understanding visual processes. The eye and the camera both have mechanisms for adjusting focus, setting exposure, and storing an image. Yet the differences between vision and photography are probably greater than the similarities. Photography is a process that captures a permanent record of the variations in intensity falling on a light-sensitive surface of the film. The process has similarities only to events that occur in the earliest stages of visual processing. The information we capture in a photograph is equivalent to visualizing the pattern of activity occurring at a particular instant in the mosaic of photoreceptors. Thus, the photograph itself replicates only the simplest aspects of vision. Without an eye to see it, a photograph is but a feeble shadow of the reality around us. When we look at a photograph, our eyes instantly see the color, form, and shape of objects from the real world.
The biologic processes transforming the shower of photons falling on a mosaic of photoreceptors into the stable and invariant experience of sight have long been of immense scientific interest. This chapter touches on what we have discovered about those processes, the information our eyes make available to us, and the details of the psychophysics of acuity, adaptation, and color.
VISUAL ACUITY
Ordinarily, seeing refers to our ability to recognize forms and patterns. An essential part of being able to see is the ability to appreciate the fine detail in a scene. Visual acuity, the ability to resolve fine detail in a pattern, is usually determined by reducing the size of a test pattern until the smallest detail in the pattern can just be resolved. Visual acuity can be expressed numerically in terms of the reciprocal of the size of the smallest resolvable detail. The size is expressed as the angle that the detail subtends at the eye of the observer. Figure 123.1a shows a Snellen letter and two other examples of acuity targets. Using such targets, visual acuity for normal observers ranges between 1.0 and 2.0 min?1. In conventional charts, with black patterns of various sizes on a white background, acuity is quantified in a slightly different fashion. The letters on this chart have been designed with the assumption that normal acuity corresponds to being able to resolve 1 min of arc (an acuity of 1.0 min?1). The size of each letter is such that its strokes will subtend 1 min of arc at a specified distance. The letter sizes can be thought of as being designated by these distances (Fig. 123.1b). This leads to the familiar fractional acuity notation, in which the numerator of the fraction indicates the viewing distance and the denominator the size of the letter. An observer who from 20 ft away can just recognize the line with letters having strokes of 1 min has a visual acuity of 20/20, an observer who requires letters twice that size has a vision of 20/40, and so forth.
|
|
|
|
FIGURE 123.1 (a) Examples of visual acuity targets. The smallest detail is indicated by arrows. (b) Spatial relationships that define visual angle. |
OPTICS
The first stage of visual processing is the formation of an image of the world on the mosaic of photoreceptors. Good vision depends on having a high-quality retinal image. Ideally, there is only a single distance plane where a given object is brought to sharpest optical focus. However, because we are tolerant of small amounts of optical blur, objects at a range of distances appear to be in sharp focus[1]. To bring targets closer or farther than this into focus, the lens of the eye must change its focal length through the process of accommodation. When the target is moved closer than the range of accommodation, the image plane falls behind the retina, and fine detail begins to be blurred. Even when an object is in best focus, there is loss of image detail owing to both aberrations and diffraction. These degradations in the sharpness of the retinal image are potentially more serious than focus errors because they are not correctable with the ordinary spherical and cylindrical lenses. The word aberration refers to a failure of the rays originating from a point source to be brought to a point focus. Inaccuracies and irregularities in the shapes of the curved refracting surfaces of the cornea and lens produce aberrations. Also, there is chromatic aberration because the refractive properties of the eye's dioptrics vary with wavelength, and different wavelengths are brought to focus at different points. Chromatic difference of focus amounts to a change of power of ?2D over the 400-700nm visible spectrum. It has been suggested that for large pupils it is the dominant aberration that limits retinal image quality.[2] Diffraction that occurs when light waves are abruptly truncated by an edge, such as the edge of the iris, also degrades the retinal image. As the result of interference between light waves and the edge of the pupil, a point, for example, will be imaged onto the retina as a fuzzy disk. The angular size of the disk varies inversely with pupil diameter.
The exact magnitude of the loss in image quality, resulting from focus errors and aberrations, also depends on the size of the pupil. Small pupils make diffraction effects worse; however, as the pupil becomes smaller, depth of focus increases and a small pupil also tends to reduce the deleterious effects of aberrations by limiting the area of the optically imperfect cornea and lens that are involved in producing the image. Over the physiologic range of pupil sizes (2 to about 7 mm), the balance between the effects of diffraction and those of aberrations occurs when the pupil is ?3 mm in diameter, approximately the size that it tends to achieve under normal bright light conditions.[3,4] Under these conditions, the quality of the retinal image is quite high and deviates only slightly from an ideal system limited only by diffraction.
The density of packing in the mosaic of foveal cones is also an important limiting factor. Each photoreceptor samples the local intensity at a point in the retinal image. Consequently, the size and the density of packing of the receptors must be adequate if we are to appreciate the fine detail in the retinal image. The foveal cones are very thin and tightly packed, making them particularly well suited to encode the fine detail in the image (Fig. 123.2a). The effect of the mosaic of foveal cones on vision is illustrated in Figure 123.2b. Acuity test patterns close to the limits of resolution have been drawn as a pattern of stimulated and unstimulated foveal cones. Because the density of receptors in the mosaic of foveal cones is just barely adequate to reproduce the image of these targets, one can readily appreciate that visual acuity is close to the limits set by the retinal mosaic. Of the two factors, optics and retinal packing, which is of primary importance? A direct experimental answer to this question has been obtained by using laser-generated interference fringes. Interference fringes, which are not images of objects but patterns resulting from the intrinsic wave properties of light, are not degraded by the eye's optics.[5]Consequently, it is possible to produce exceedingly fine high-contrast gratings directly on the retina. Visual acuity is ?50% higher with interference fringes (20/ to 20/10).[4] Thus, the resolution limit set by the packing density of the cones is similar to, but slightly higher than, the limit set by the optics of the eye. As a result, under ideal conditions, an observer with excellent vision can just resolve fine detail whose angular subtense approaches that of a single cone.[6]
|
|
|
|
FIGURE 123.2 (a) Section through cone inner segments at the center of human fovea. The bar indicates 2 min of arc (10 ?m distance on retina). The particular retina illustrated had the lowest peak density of the four retinas studied. (b) 20/20 Snellen letter drawn as a pattern of stimulated and unstimulated cones. |
To transmit to the brain the information about fine detail that is available at the level of the photoreceptors, there must be at least as one ganglion cell for each cone. For the fovea, where acuity is highest, there are between two and three ganglion cells transmitting information to the brain for every cone. A neural network compares the number of photons absorbed by a cone to the average number absorbed by its neighbors and projects the result via two 'private lines' to the brain in the form of one ON- and one OFF-center midget ganglion cell. These midget ganglion cells are responsible for transmitting the information about fine details in the image.
The response properties of the ganglion cells represent a first level of processing of photoreceptor signals into visual percepts. Each cone provides input to several types of ganglion cells each specialized to carry specific information about the visual stimulus. As shown in Figure 123.2, when part of the image of a black horizontal stroke of the Snellen E falls on a cone, its OFF-center ganglion cell fires, signaling the presence of a dark area in the image. When the image of the white background between two strokes of the E falls on a cone, the ON-center ganglion cell fires, signaling the presence of a light area of the image. The ratio of ganglion cells to cones is greater than 2:1 in the fovea because foveal cones connect to other types of ganglion cells that collect signals from larger numbers of cones and carry other types of information. There are ganglion cells that transmit the presence of a hue, for example, 'blueness', and other ganglion cells transmit information responsible for the percept of movement.
RETINAL POSITION
Visual acuity falls rapidly as the focal point moves away from the fovea, as would be expected from the decrease both in density of cone photoreceptors and in the relative number of ganglion cells available to carry information from the retina. The exact shape of the fall in acuity with eccentricity depends on the type of target used, but acuity falls roughly to half at 1° and to one-fourth at 5°.[7] To separate the optical factors from retinal factors, one can use interference fringes formed directly on the retina. With the stimulus near the fovea, the fall in acuity for interference fringes parallels the fall in cone density.[6]Beyond ?5° of eccentricity, the rate of decrease is too great for which the peripheral cone spacing happens to be the limiting factor. At these eccentricities, the fall-off in resolution and the estimated decrease in spacing between ganglion cells seem to agree reasonably well.[8]
CONTRAST
Although visual acuity is frequently used to characterize an individual's ability to see, there is considerably more to functioning in everyday life than being able to resolve fine detail in well-illuminated, high-contrast black-and-white patterns. In the real world, one must detect and recognize a variety of targets varying in contrast, size, and shape.
In recent years, we have increasingly come to appreciate the importance of the effects of contrast on visual performance.[4,9-15] To study the effects of contrast, an approach borrowed from optical engineering has been quite useful. The optical-transfer function is frequently used to characterize the imaging abilities of cameras and television systems. The idea of using the optical-transfer function stems from two key facts about linear systems: (1) any stimulus can be considered to be a sum of sinusoidal components, and (2) purely sinusoidal patterns are imaged by optical systems in a uniquely simple way. That is, the spatial variations in the image of a sinewave target are also sinusoidal, of the same spatial frequency but of reduced contrast. Thus, for any grating, a contrast reduction factor (and sometimes a phase shift factor) completely describes the object to image transformation. This number as a function of frequency defines the optical-transfer function.[16] Because any pattern of luminances can be described by a sum of sinewaves in two dimensions, knowing how sinewaves are imaged (i.e., the transfer function), it is possible to calculate the image that will be formed by any arbitrary pattern. That is, the optical-transfer function completely describes the imaging properties of the eye, and using the transfer function it is possible to quantify the quality of an image and to give a detailed account of how one or another factor influences the image quality. Figure 123.3a shows measurements of the optical-transfer function for an observer in good focus with a range of pupil sizes. Because at 30 cycles/degree the bars of the grating are of the same width as the strokes in the 20/20 letter; the curves show that near the 'normal' limit of resolution, the image of a grating is reduced in contrast by ?50%.
|
|
|
|
FIGURE 123.3 (a) Optical-transfer functions for the in-focus eye of a normal observer at several pupil sizes (. 2 mm; ?, 2.8 mm; ?, 3.8 mm; ?, 5.8 mm). b Contrast-sensitivity functions for 10 normal observers. Each point plots the contrast at which an observer could just detect the sinewave grating. |
In an attempt to extend the idea of an optical transfer to the processing occurring in the retina and brain, the contrast-sensitivity function has been evolved. To characterize the eye's ability to process spatial information, the observer is presented with a periodic pattern, usually a sinewave, and adjusts the contrast of the target until it can be just detected. The reciprocal of the threshold contrast defines the contrast sensitivity. Contrast sensitivity is then plotted as a function of the grating fineness, with the fineness of the grating expressed in terms of the number of bars in 1° of visual angle (cycles/degree). The shape of the contrast-sensitivity function reflects, both the optical-transfer function of the eye's dioptric apparatus and the neural processing of spatial frequency information.
Typical contrast-sensitivity functions are shown in Figure 123.3b. The contrast sensitivity, the reciprocal of the just-visible contrast, peaks at an intermediate spatial frequency. Gratings that are coarser or finer than this optimal frequency require more contrast to be seen. The point where the contrast-sensitivity function intersects the horizontal axis establishes the highest spatial frequency that an observer can detect. Because one cycle of a grating contains a dark and a light bar, a 30 cycles/degree grating has bars that subtend 1 min of arc and corresponds in size to the strokes in the 20/20 letter.
Thus, in the same way that the optical-transfer function characterizes the imaging properties of a lens, the contrast-sensitivity function describes a patient's ability to see. In general, by considering the spatial frequency content of targets and knowing an observer's ability to process sinewave information at a variety of spatial frequencies, one obtains information relevant to assessing an observer's functional visual capabilities.
Contrast sensitivity provides information that supplements what one can obtain from acuity measurements alone. For example, two patients can have the same acuity and yet have different middle- and low-frequency contrast sensitivities. The differences in middle- and low-frequency sensitivity can have dramatic effects on the patients' visual performances. Moreover, contrast-sensitivity loss not only can influence the detection of larger low-contrast targets but also seems to affect tasks that one might have thought would require only good acuity. It has been reported that patients with only slight losses in acuity but with losses in contrast sensitivity over a wide range of spatial frequencies can experience difficulties in reading.[17,18] The explanation, in part, is provided by Rubin and Legge's[19] finding that peak contrast sensitivity is an important determinant of how rapidly low-vision observers can read letters. On the basis of these and other studies, it seems likely that contrast-sensitivity determinations will continue to be useful in assessing functional visual capacities, and in predicting such things as the benefit that might be derived from a particular low-vision aid.
INTENSITY
Our ability to see is strongly dependent on having adequate light. At the lowest light levels, a few quanta absorbed in an area containing 500 rods can produce a visual sensation. At these levels, some semblance of form vision is possible, but acuity is exceedingly low. Even at an intensity of ?10?5cd/m2, which is about a log unit above absolute threshold, visual acuity is only ?20/1000.[20] As luminance of the target is increased, the ability to resolve details continuously improves up to luminances of ?10cd/m2, where it reaches a plateau (Fig. 123.4).[21] A number of factors contribute to the dependence on target luminance. With the dimmest targets, vision is mediated by rods 4° or more from the fovea.[22] As stimulus intensity increases, there is movement of fixation from the periphery toward the fovea. At lowest levels of illumination, where the stimuli are so weak that not every receptor will absorb a quantum, the quantal nature of the stimulus can severely limit the eye's ability to appreciate details and contrast. Smaller pupil sizes that accompany higher luminances can also contribute improved vision because the quality of the retinal image with a fully dilated pupil is not optimal. As light intensity increases, there is a changeover from vision mediated by rods with high sensitivity but poor resolution to vision mediated by cones with intrinsically better resolution. This switch from rods to cones frequently produces a sharp and rather abrupt increase in the acuity-intensity relationship.
|
|
|
|
FIGURE 123.4 Visual acuity intensity functions under free viewing conditions. |
EYE MOVEMENT
The eyes are continuously executing small involuntary movements. The pattern of involuntary eye movements is a mixture of tremor, slow drifts, and rapid saccades. By analogy with a camera, one might think that any movement would be detrimental. In fact, eye movements seem to be an essential prerequisite to normal vision. If the retinal image is stabilized, within seconds the visual image fades.[23,24] After a minute or so, only a very blurred, cloudy version of the original scene persists.[25] These subjective reports of the appearance of stabilized images might suggest that high-frequency information disappears more quickly than coarser detail, but systematic measurements with gratings seem to suggest the reverse.[26] The key point is that for vision to be possible, there have to be eye movements. The eye movements move the retinal image, and as a result, the photoreceptors are subjected to continually changing spatial and temporal transients.
LIGHT AND DARK ADAPTATION
We are largely unaware of the great variations in illumination that occur in the real world because the eye light-adapts. By increasing (or decreasing) sensitivity when the environmental intensity increases (or decreases), our perception of the visual world remains relatively constant. The term light adaptation is used to describe the changes in visual sensitivity produced by steady background lights. An everyday example of this change is the disappearance of the stars with the coming of dawn. This, of course, is not due to a change in the amount of light emitted by the stars but rather is a result of the desensitization produced by the veil of scattered light in the sky. Dark adaptation is the recovery of sensitivity with time after exposure to a background that reduces sensitivity. Light adaptation works well in the natural world because we are attempting to sense objects around us that are not self-luminous but rather are illuminated from a distant source and seen by reflected light (Box 123.1).
BOX 123.1
|
Adaptation and Weber's law |
|
If E is the intensity of the illumination locally, and ? (x,y) is the spatial variation in the reflectivity of the objects in a scene, the equation L = ? (x,y)E gives the point-by-point variations in the luminance from particular region of the scene. The information extracted by the visual system from the image on the retina is an impression of ? (x,y), which is relatively independent of E. The eye does this by adjusting its sensitivity according to Weber's law, which states that sensitivity varies inversely with E, the intensity of the illumination. When E increases by, say, a factor of 10, the retinal gain decreases by this same factor, and consequently the neural signal, which depends on the product of retinal illumination by retinal gain, remains unchanged. |
Increment thresholds and dark adaptation are the two classic adaptation paradigms that have been used to reveal the changes in sensitivity occurring in the eye. In both, the dependent measure is usually an observer's ability to detect a small, briefly presented test flash. To determine increment thresholds, the small incremental test stimulus is usually presented on a spatially uniform background field, which is systematically varied in intensity.
LIGHT ADAPTATION
In controlled laboratory conditions, one typically finds that when a small test flash is added to steady background over a considerable range of background intensities, the intensity of the just-detectable increment is approximately proportional to background intensity. This property, that the ratio of the test flash intensity to the background intensity is roughly constant, is Weber's law relationship (Box 123.1). There are, however, failures at both ends of the scale, when the background is either very bright or very dim. At the low end, Weber's law fails when the background becomes so dim that it no longer affects sensitivity. At the other extreme, a bright background can overload the system. If the background is sufficiently intense, a new phenomenon called saturation occurs. That is, in order to be seen, the increment needs to be made considerably brighter than one would predict from Weber's law.[27] In addition, increment threshold curves frequently have a discontinuity. Rod signals mediate low-intensity vision, and cones subserve the upper range of intensities. As a result, plots of increment threshold as a function of intensity yield curves that frequently are divisible into two distinctly different portions, with a kink in the curve marking the changeover from predominantly rod to predominantly cone control of sensitivity (Fig. 123.5).
|
|
|
|
FIGURE 123.5 Increment threshold functions. The lower limb is for rods, and the upper limb is for cones. |
The foregoing description of the behavior of the visual system during adaptation says little about what is happening in the eye during these changes. Because exposure to light bleaches visual pigment, the simplest possible mechanism imaginable to account for losses in sensitivity would be a decrease in the potency of the test probe caused by the direct removal of visual pigment. This, however, is not the case. Many years ago, Rushton[28] showed that a background can be so dim that only a few rods can absorb photons and yet significantly elevate threshold. Only one rod in 50 needs to absorb a photon of light for the threshold of a stimulus that falls on the receptors not directly affected by the background to be elevated by a factor of four. Even at background levels where rods saturate, only small amounts of visual pigment are bleached.
Cones may be a bit different in this regard. It is not completely clear to what extent backgrounds can desensitize the photoreceptors themselves. In addition, with cones the adapting backgrounds can bleach pigment, and so the depletion of visual pigment may contribute to sensitivity loss in cones at high light levels.
DARK ADAPTATION
It takes time for the threshold to reach a new equilibrium value after a background abruptly changes from one intensity to another. The recovery of sensitivity in the dark after prior exposure to a bright stimulus is called dark adaptation. The speed with which a new equilibrium is reached depends on the direction and magnitude of the change. In general, reduction in sensitivity occurs quickly relative to restoration of sensitivity. That is, it takes a matter of seconds to adapt to a brighter background, whereas adjustment to decreases in background intensity is slower. In particular, the recovery to total darkness may proceed very slowly. After exposure to a bright stimulus, the exact time needed to adjust to a dimmer background depends on whether rods or cones are being tested and on the intensity of the prior exposure. In the extreme situation, in which one is plunged from a very bright environment into complete darkness, it can take up to an hour for rod sensitivity to recover fully. Measuring the dark-adaptation curve is the standard method for tracking this recovery process. One plots the threshold as a function of time after the termination of the conditioning light stimulus. The shape of the dark-adaptation curve depends on the test stimulus parameters, such as size, color, and retinal location, as well as conditioning stimulus parameters, such as its intensity, duration, and color. Fortunately, to a great degree these multiple factors can be reduced to just two principal determining variables. These are the extent to which the test light stimulates rods and cones and the quantity of visual pigment that has been bleached by the exposure to the conditioning stimulus (see Box 123.2).
BOX 123.2
|
Adaptation and the Dowling-Rushton Law |
|
The Dowling-Rushton Law is the empirical relationship between bleached pigment and sensitivity. Rather than being linear it is, to a good approximation, logarithmic, given by the equation L It = ? (1 - ?) where It is threshold, ? is the proportion of pigment, and a is a constant. |
If the size, color, and retinal location are arranged so that the test flash stimulates both rods and cones, and significant amounts of rod and cone pigment are bleached by the conditioning stimulus, then dark adaptation proceeds in two distinct phases. In the first phase or branch, typically lasting ?5-10 min, the threshold is determined by the cones as they recover their sensitivity. The time course of the process parallels the regeneration of cone pigment. Later, the rods recover sufficiently for their thresholds to be lower than those of the cones, and they mediate threshold, giving rise to a rod branch in the curve. Complete recovery of rod sensitivity takes as long as it takes for rod pigment to regenerate. During this time, sensitivity can increase by a factor of 10000 (Fig. 123.6).
|
|
|
|
FIGURE 123.6 A typical dark-adaptation curve. The first limb of the curve reflects recovery of cones, and the second, slower limb reflects rod recovery. |
COLOR VISION
Humans value color perception highly as a sensory capacity. Few of us would accept a black and white television or monochrome computer monitor in place of color ones even in exchange for large savings in cost. One reason color is important to us is that it has a powerful effect on our emotions. This is presumably, in part, because some components of our color vision system are evolutionarily ancient, predating other sensory capacities. Color cues are associated with time of day, the season, and an organism's position and orientation in space. Color can signal the presence of an injury or illness, the presence and or quality of food, and the identity of a mate. In our modern world color-coding is extremely important in transmitting information visually. Objects identified by color among a large number of distracters can be located nearly instantly in visual search and color is invaluable in perceptual grouping and segmenting objects.
With regard to mechanism, color vision is based on three types of cone photoreceptors which are the basis for all vision with the exception of vision under very dim light conditions, which is dependent on rods. Information about pattern, luminance, and color are all extracted from a mosaic of three types of cone, one class most sensitive to short wavelength light (S), a second class (M) most sensitive to middle wavelength light and a third (L) most sensitive to long wavelength light (Fig. 123.7). Perception of black, white, and gray, the hues of red, green, blue, and yellow and their patterns in the retinal image are extracted by different types of ganglion cells each specialized to carry specific information about the visual stimulus from one mosaic of cones.
|
|
|
|
FIGURE 123.7 (a) Illustration of the arrangement of the three types of cones, L-, M-, or S-, which represent classes of photoreceptors that are primarily sensitive to long wavelength light (L), middle wavelength light (M), and short wavelength light (S) within the visible spectrum. In this diagram the cones are colored blue, green, or red in order to represent the different photoreceptor classes. S cones represent a minority, ?5% of the total. The arrangement in a trichromat is very close to random for L and M cone classes. Midget ganglion cells have center surround-receptive fields with the center derived from a single cone (for example, within the inner ring of the two black concentric circles). The six adjacent cones are the most important contributors to the surround (for example, the cones within the outer ring of the two black concentric circles). In a normal trichromat the surrounds of many L or M cones will have some cones of a different type than the center. Thus most midget ganglion cells have spectrally opponent responses to diffuse red and green lights; they are either excited by red and inhibited by green or vice versa. (b) Arrangement of cones in a deuteranope, a dichromat, with only two different cone classes L and S. In a dichromat, usually all the cones in the surround are of the same class as the center input to a midget ganglion cell. Such ganglion cells do not respond to diffuse colored lights; they are specialized to signal patterns of light and dark on the retina. |
A great deal is known about the neural types and their interconnections in the retina responsible for luminance, color, and form, and there is a growing body of information about the higher visual centers; yet, how these operate to give us vision remains a fascinating puzzle. Clues toward solving the puzzle come from information about the evolution and development of visual system and about its anatomy and physiology. In contrast to our persisting ignorance about color vision circuitry, the last 20 years has seen an explosion of information about the cone photopigments. Many of the long standing questions about these pigments and their role in normal vision and vision disorders have now been answered. In humans there are three types of cone photopigment, one for each class of cone (see Box 123.3). The L and M cone opsins and the genes that encode them are unusually variable, presumably due to the unstable tandem arrangement of the genes and their unique evolutionary history. Among the mammals, red-green color vision first evolved in a primate ancestor. Strong selective pressure favoring trichromatic color vision acted on primates in the wild, minimizing the prevalence of mutant L and M opsin gene arrays despite the extreme instability inherent in tandemly duplicated genes. However, in humans, selection against mutant X-linked pigment genes has been relaxed. The variability that has resulted includes gene arrangements responsible for color blindness, which is the most common of all human single locus genetic disorders. The increasing proportion of mutant cone pigment genes is also presumably the root of a growing number of other problems and complications of human vision.
BOX 123.3
|
Normal color vision: terms and genetics |
|
Normal human color vision is trichromatic, mediated by three well-separated classes of cone photoreceptor commonly referred to as the blue, green, and red cones. Calling them short-, middle-, and long-wavelength sensitive, abbreviated S, M, and L cones minimizes confusion that can arise from giving them color names. Photopigment molecules within each cone are responsible for the spectral properties of the cones. Each photopigment molecule is composed of two parts; a protein termed the opsin, and an 11-cis-retinal chromophore. The official names for the genes encoding the L, M, and S cone opsins are OPN1LW, OPN1 MW, and OPN1SW respectively. Both OPN1LW and OPN1 MW are on the X-chromosome at position Xq28, OPN1SW is located on chromosome 7 at 7q32.1. The location of OPN1LW and OPN1 MW on the X-chromosome accounts for the great gender difference in the prevalence of color vision deficiencies. |
INHERITED RED-GREEN COLOR VISION DEFICIENCY
Inherited color vision deficiency occurs at an extraordinarily high frequency in human populations although the prevalence varies with ethnicity and race.[29] Caucasians exhibit among the highest rates with 7-8% of males affected, and native Fijians have the lowest rate with 0.82% of males affected, while Japanese and Africans have intermediate rates with 4.17% and 2.61% of males affected, respectively. Compared to many other common inherited recessive disorders, such as cystic fibrosis and sickle cell anemia, color vision deficiency is unusual in occurring at an exceptionally high frequency with no compelling evidence of a strong heterozygote advantage to explain why. For example, cystic fibrosis is the most-common life-limiting autosomal recessive disorder among humans, estimated to occur at a rate of about one in 3200 live births.[30] Heterozygotes for cystic fibrosis are protected against heat- and disease-induced dehydration, thereby providing them with a survival advantage over the history of human existence. Another example is sickle cell anemia in which heterozygotes are protected against the severe pathogenesis of malaria.[30] The answer to why color vision deficiency is so prevalent in human populations lies in the evolutionary origin and arrangement of the OPN1LW and OPN1 MW genes in the human genome, and on the strength of natural selection on trichromatic color vision.
Genetic evidence indicates that all vertebrate opsin genes evolved from a common ancestor through a process of gene divergence and duplication.[31] Most mammals have two types of cone photoreceptor, one maximally sensitive to short-wavelength (S) or in some cases ultraviolet (UV) light and another maximally sensitive to light in the middle-to-long wavelengths (M/L).[32] The opsin components of the photopigment molecules that determine the spectral properties of the cones are encoded by an autosomal gene in the case of the S or UV opsin, and a gene on the X-chromosome in the case of M/L opsin. Together, the two cone types form the basis for dichromatic color vision. In New World primates, trichromatic color vision was acquired through evolution of allelic diversity in the X-chromosome opsin gene, which produced variety in spectral sensitivity of the encoded photopigments.[33] Males have only one X-chromosome, but females have two. In female New World monkeys who are heterozygous at the X-chromosome opsin gene locus, X-inactivation segregates expression of the alleles into separate populations of cones, producing three cone types. The heterozygous females have trichromatic color vision,[34] indicating that they have all of the components necessary to form fully functional circuits for trichromatic color vision. In Old World primates, trichromatic color vision arose via a gene duplication that placed two opsin genes together in tandem on the X-chromosome.[31] X-inactivation cannot segregate expression of tandem OPN1LW and OPN1 MW genes into separate populations of cones; however, a critical enhancer known as the locus control region (LCR) was not duplicated along with the opsin gene, thereby limiting the photoreceptor to expressing one opsin gene at a time. [35-38] Ultimately, in the adult cone photoreceptor, only one opsin gene is expressed to the exclusion of all others.[39]
Tandemly duplicated genes are inherently unstable because they are prone to unequal homologous recombination between misaligned arrays during meiotic cell division in females, which produces gene rearrangements that underlie inherited color vision deficiency,[31] as illustrated in Figure 123.8. Natural selection is expected to virtually eliminate visual pigment gene arrays that confer color vision defects; however, if selection is relaxed, then arrays causing color vision defects can accumulate in the population giving rise to an increase in color vision deficient males and in female carriers. In the US, about one in 12 males is affected by red-green color vision deficiency and one in seven females is a carrier.
|
|
|
|
FIGURE 123.8 The variety of normal and defective color vision phenotypes in humans is produced by unequal homologous recombination during meiotic cell division in females. Red and green arrows represent the OPN1LW and OPN1 MW genes, respectively; arrows that are half red and half green represent chimeric genes produced by a crossover between an OPN1LW and an OPN1 MW gene. The color of the arrowhead indicates whether the gene encodes an L pigment (red) or M pigment (green). (a) A crossover between an OPN1LW and OPN1 MW in the parental arrays produces two recombinant arrays. Only the two 5' genes (left-most) in the three gene array are expressed and both of these encode L-class pigments. The array produces dichromatic color vision if the encoded L-class pigments have identical spectral properties, or it produces anomalous trichromacy if the pigments differ in spectral properties. (b) A crossover in the region between the genes in one array and the region downstream of the last gene in another array produces a three gene array that encodes an L and an M pigment and thus will produce normal color vision. The other recombinant array contains a single gene, which encodes an L opsin and thus produces dichromatic color vision. (c) Recombination between an OPN1 MW in a three gene array and OPN1LW in a two gene array gives rise to two recombinant arrays, both of which give rise to color vision defects. One array will have two genes, an OPN1LW/OPN1 MW hybrid followed by an OPN1 MW gene, which if both genes encode pigments identical in spectral properties will cause dichromacy, or if the encoded pigments differ in spectral poperties, the array will cause anomalous trichromacy. Likewise the array with an OPN1LW gene followed by an L-pigment encoding chimeric gene will either cause dichromacy or trichromacy if the encoded pigments are identical or different in spectral properties, respectively. |
When color defective arrays initially began to accumulate in the population, female carriers would have had one normal array with one OPN1LW and one OPN1 MW gene and another array that either had one or three opsin genes (Fig. 123.8a,b). In either case, the unequal number of opsin genes on the two X-chromosomes produces instability because there is no perfect alignment of the two arrays during meiotic cell division. Recombination between two arrays with different numbers of opsin genes will give rise to variability in the number of opsin genes per X-chromosome and an increase in the prevalence of chimeric genes resulting from intermixing of OPN1LW and OPN1 MW genes (Fig. 123.8c). As will be discussed below, the diversity introduced by intermixing L and M opsin gene sequences underlies the variety of phenotypes associated with anomalous trichromacies.
Only among humans is there widespread variability in the number of visual pigment genes on the X-chromosome with a high frequency of arrays containing more than two opsin genes.[31,40] Since normal trichromatic color vision requires expression of only one L and one M opsin gene, this raises the question of whether the extra genes beyond the necessary two are expressed. Experiments have shown both by inference[41] and by direct analysis of opsin gene expression in human retinas from color deficient donors[42] that usually only the two opsin genes at the 5' end of the X-chromosome opsin gene array are expressed, although exceptions have been observed.[43] Thus, the order of the genes in the array on the X-chromosome and the spectral sensitivity of the photopigments encoded by the expressed genes play a central role in determining color vision phenotype.
CAUSES OF COLOR VISION DEFICIENCY
A common misconception is that inherited red-green color vision deficiency is a single entity; it is not. Instead it is a group of disorders that can be dichotomized at the first level according to what is missing to cause the perceptual loss, and at a second level according to the degree of color vision that remains (Box 123.4). The most common cause of color vision deficiency is the deletion of all OPN1LW (protan defects) or all OPN1 MW (deutan defects) genes. The degree to which color vision is impaired is determined by the spectral properties of the pigments encoded by the genes that remain.
BOX 123.4
|
Red-green color vision deficiency: classification and terminology |
||||||||||||||||||
|
Red-Green color vision defects fall into two categories depending on which cone class does not contribute to color vision. The noncontributing cone class is indicated by the prefixes:
Further categorization of color vision defects depends on whether the remaining color vision is based on two (dichromacy) versus three (anomalous trichromacy) spectrally distinct types of cones. The suffix -opia denotes dichromacy. The suffix -anomaly denotes anomalous trichromacy in which two of the cone classes are more similar in spectral sensitivity than the corresponding normal cones:
|
DICHROMACY
The most severe of the common inherited red-green color vision defects are the dichromacies, protanopia and deuteranopia in which color vision is mediated by just two pigments in two types of cone. Protanopia and deuteranopia (Box 123.4) each occur at a rate of ?1% in Caucasian males, and although dichromacy is much rarer in females, about one in 4000 females is affected.
In most cases, the direct cause of dichromacy is the deletion of the genes that encode one class of cone photopigment from the X-chromosome through unequal homologous recombination[31,44-47] as illustrated in Figure 123.8a,b. For example, in a recent study 53 of 55 protanopes lacked genes for L opsin, and 51 of 73 deuteranopes lacked genes for M opsin.[47]
PROTANOPIA
One common cause of protanopia is the deletion of all but one opsin gene on the X-chromosome with the one remaining gene encoding an M-class pigment (Fig. 123.8a).[31,48-49] Another common gene arrangement among protanopic men is an array that lacks all OPN1LW genes, and that has a chimeric OPN1LW/OPN1 MW gene in the first position and an OPN1 MW gene in the second position (Fig. 123.8c).[46,47] For protanopes with this arrangement, the chimeric gene and the OPN1 MW gene do not encode photopigments that differ in spectral properties, accounting for the protanopic phenotype.[47]
Occasionally, protanopes who have an apparently intact OPN1LW gene have been identified[31,46,47] and presumably the gene is either not expressed or does not encode a functional photopigment. In cases where this hypothesis has been tested experimentally, the OPN1LW gene has either been found to be displaced to the 3' end of the array where it is not expressed,[46,47] or it has been found to carry a particular deleterious combination of amino acids at polymorphic positions encoded by exon 3. The same combination was observed in OPN1 MW genes in deuteranopes where its effect on the cone mosaic and its contribution to dichromacy has been examined in more detail, as will be described below under deuteranopia.
DEUTERANOPIA
The majority (about two-thirds in some studies) of deuteranopes have undergone a deletion of all X-chromosome opsin genes except for one remaining OPN1LW gene, accounting for the phenotype.[31,44,50,51] Nearly one-third of deuteranopes have an OPN1 MW gene; however quite often they have a chimeric OPN1LW/OPN1 MW gene inserted between the OPN1LW and OPN1 MW genes, displacing the OPN1 MW gene to a nonexpressed position.[31,52] Another relatively common cause of deuteranopia is the presence of an inactivating mutation in the OPN1 MW.[47,53] By far the most common inactivating mutation found in OPN1 MW genes is a nucleotide change that results in the substitution of arginine for a highly conserved cysteine at position 203 (C203R) of the cone opsin molecule, preventing the opsin from folding properly.[54,55] Other inactivating amino acid substitutions have also been found but most are quite rare.[47]
Perhaps the most interesting inactivating mutation found in OPN1 MW, which as was alluded to above has also been found in OPN1LW genes, is a combination of amino acids at normally polymorphic positions. OPN1LW and OPN1 MW have been intermixed by recombination so that in the present day population of humans with normal color vision, there are 11 dimorphic positions among L and M pigments. As a result, there is tremendous variation in the amino acid sequences of the L and M cone photopigments found in humans with normal color vision.[56-59] A specific combination of amino acids at the dimorphic positions has been observed to always be associated with a color vision deficiency in which there is a perfect correlation between the opsin with the deleterious combination and the absence of function of the corresponding cone. High resolution adaptive optics imaging of the retina of a deuteranope, whose OPN1 MW gene specified an M pigment with the deleterious combination, was shown to have gaps in his cone mosaic, presumably where M cones had once been.[60] In addition, the deuteranope was shown to have a reduction in cone density by about one-third. When the gaps in his mosaic were modeled as M cones and cone density recalculated taking into account the modeled cones, the density estimate was normal. Taken together, these observations support the hypothesis that in some forms of color vision deficiency, the cause is a loss of photoreceptor cells that is due to a malfunction in the production or function of the photopigment.
Males who have opsin gene arrays in which the first two genes encode photopigments of the same functional class but with a difference in peak sensitivity of less than 2.5nm, occasionally perform as dichromats on standard color vision tests, including the anomaloscope color matching tests.[47] Thus, their performance is worse than would be predicted strictly by the complement of opsin genes they have. When there is a very small spectral separation on which the person must rely to make color discriminations, there appear to be a variety of other factors including physiological factors, personality factors, differences between naïve and experienced observers, and differences in the relative ratios of the underlying cone photoreceptors to name a few, that contribute to variability in phenotype.
In summary, the most severe red-green color vision defects, the dichromacies, are commonly explained by the straightforward deletion of cone opsin genes. Another relatively common cause is a point mutation that disrupts the function of the encoded opsin. Recent evidence indicates that there is a fundamental difference in the effects on the cone mosaic that results from these two mechanisms for color vision deficiency, specifically with regard to what happens to the subpopulation of cones that do not contribute to color vision. In the case of the single-gene dichromat, it appears that all of the cones that would have become L or M cones express the available X-chromosome opsin gene and so no cone photoreceptors are lost. In contrast, evidence indicates that in dichromats with two or more opsin genes in which the first or second gene encodes an opsin with an inactivating amino acid substitution(s), a subpopulation of photoreceptors express the mutant gene, which ultimately results in the death of the photoreceptor, giving rise to a lower than normal cone density and leaving gaps in the cone mosaic.[60] What occupies the gaps left by the absent cones remains unknown.
ANOMALOUS TRICHROMACIES
The milder forms of red-green color vision deficiencies are the anomalous trichromacies. As the term for their condition implies, affected individuals have trichromatic color vision, but it is not based on L, M, and S pigments like normal color vision. Classical descriptions of anomalous trichromacy postulated the existence of 'anomalous pigments' such that in addition to S cones, protanomalous individuals are said to have normal M and anomalous L pigments while deuteranomalous individuals are said to have normal L and anomalous M pigments. Results of molecular genetic analyses have provided insight into what the anomalous pigments are, and as a consequence it has become clear that the classical concept of the 'anomalous pigment' is unbefitting.[52] For example, 'anomalous L' pigments are often indistinguishable from normal M pigments in spectral sensitivity and in amino acid sequence, and there is similar overlap between 'anomalous M' pigments and normal L pigments.[51,52] The 'anomalous pigments' have been generated by recombination between the ancestral OPN1LW and OPN1 MW genes; however, over evolutionary time, multiple rounds of recombination have intermixed the ancestral pigment genes so thoroughly that among modern human males with normal color vision, there is a family of photopigments specified by the OPN1LW genes that differ in amino acid sequence and in the wavelength of peak sensitivity (Fig. 123.9). Similarly, there is a family of pigments encoded by the OPN1 MW genes found in color normal individuals (Fig. 123.9). The variant forms of the L and M photopigments underlying normal color vision overlap with the forms that correspond to what were classically termed the 'anomalous pigment'. Referring to the photopigments underlying anomalous trichromacy according to their spectral sensitivities promotes a clearer understanding of the cause of the differences between normal versus anomalous trichromacy.
|
|
|
|
FIGURE 123.9 During human evolution, recombination has intermixed the OPN1LW and OPN1 MW genes so thoroughly that there is not just one L- and one M-pigment encoded. Instead, the OPN1LW and OPN1 MW genes found in males with normal color vision encode a family of L-cone photopigments (L-class) and a family of M-cone photopigments (M-class). The OPN1SW gene has not been observed to vary, and hence among humans there is only one know functional S-cone photopigment. |
The genes can be categorized as encoding an L-class or an M-class pigment by the sequence of exon 5. Whether the encoded pigment will have peak sensitivity near 560nm or near 530nm is determined by the amino acids at two of the polymorphic amino acid positions encoded by exon 5 (Fig. 123.10).[61-63] Amino acids encoded by five other polymorphic positions encoded by exon 2, 3, and 4 produce spectral variants of the L-class. Only the polymorphic amino acid positions specified by exon 3 and 4 produce shifts in spectral peak of M pigments, and thus, there are fewer variant forms of M- than of L-class pigments[62] (Fig. 123.9). In addition, the spectral shifts produced by the polymorphisms in the M-class pigments are relatively small compared to shifts made by the same amino acid substitutions at the corresponding positions of the L pigments. From the deduced amino acid sequences of the pigments encoded by the genes in the first two positions of the X-chromosome opsin gene array, the spectral separation of the pigments underlying color vision can be predicted.
|
|
|
|
FIGURE 123.10 Spectral tuning of the L and M cone photopigments is achieved through amino acid substitutions at a limited number of positions. The balls represent the amino acids that comprise the L and M opsins. The photopigments are seven transmembrane proteins. The red balls indicate amino acids positions 277 and 285. Substitutions at these positions distinguish the L-class from the M-class pigments. The yellow balls indicate amino acid positions at which substitutions produce spectral shifts, and that are responsible for variability in the absorption spectrum among the L-class, among the M-class, and between the L- and M-classes. Substitutions at the yellow positions have a relatively small effect on the absorption spectrum in comparison to the red positions. The blue balls indicate amino acid positions that are variable among L- and M-class pigments but that do not influence the absorption spectrum. |
It is the large difference in spectral absorption between L and M cones that underlies the excellent color discrimination in the red-to-green region of the visible spectrum that is enjoyed by people with normal color vision. Individuals with deuteranomaly lack M cones but they have two distinct classes of L-cone that are different enough in spectrum to provide the basis for limited color vision in the red-green region of the spectrum.[62-64] Likewise, individuals with protanomaly lack L cones but they have two different classes of M-cone. An anomalous trichromat with a large spectral difference between the L or M cone subtypes has the basis for much better color vision than a person with two cone subtypes that are nearly identical.[44,51,64-65] Spectral separations of 5nm or larger give rise to only a very mild color vision deficit,[64] and affected individuals often perform as nearly normal in standard color vision tests. When the underlying pigments are separated in peak sensitivity by between 2.5 and 5nm, color vision is more than mildly impaired, but is nonetheless quite excellent compared to smaller separations. When color vision in the red-green region of the spectrum is mediated by cones that differ in peak sensitivity by fewer than 2.5nm, color discrimination is quite impaired, and some affected individuals' performance on standard color vision tests is indistinguishable from the performance of a dichromat.[64]
PROTANOMALY
Protanomalous individuals usually have a rearranged visual pigment gene locus in which the first gene is an OPN1LW/OPN1 MW chimera with exon 5 derived from the parental OPN1 MW gene and thus it encodes a pigment of the M class. The chimeric gene is followed by an OPN1 MW gene (Fig. 123.8c). Amino acid sequence differences between the M-class pigments encoded by the two genes produce a spectral difference between them, enough to support a small degree of trichromatic color vision (Fig. 123.10).[51,52,64]
DEUTERANOMALY
Deuteranomalous individuals usually have a rearranged visual pigment gene locus so that the first gene is a normal OPN1LW gene, which is followed by a chimeric gene that is an OPN1 MW/OPN1LW gene that was produced by unequal homologous recombination and that encodes an L-class photopigment (Fig. 123.8a). It is not uncommon for the chimeric gene to be followed by one or more additional chimeric genes or normal OPN1 MW genes (Fig. 123.8a), however, because genes downstream of the first two at the 5' end of the array are usually not expressed, the additional genes do not bear on the color vision phenotype.
TRITAN COLOR VISION DEFICIENCY
Color vision defects caused by abnormalities of the S cones are denoted with the prefix tritan, and exhibit autosomal dominant inheritance. In addition, tritan defects show incomplete penetrance, meaning that there is variability in the degree to which color vision is impaired among individuals with the same underlying gene defect, even within a family. That is, even among members of the same family, some individuals might exhibit a complete loss of S-cone function, whereas other members may exhibit a milder, incomplete loss.[66-69] Thus, tritan defects do not parallel red-green color vision defects and cannot be dichotomized into analogous dichromatic and anomalous trichromatic forms.
Tritan deficiencies have been associated with mutations of the S-opsin gene resulting in four different amino acid substitutions.[70-72] In one study,[70] it was concluded that unlike mutations in the rod pigment rhodopsin that cause autosomal dominant retinitis pigmentosa (adRP), amino acid substitutions in the S opsin do not cause retinal degeneration. Given that only ?5% of the cone photoreceptors in humans are S cones,[73-74] this is not surprising. However, the absence of retinal degeneration does not imply that the S cones do not degenerate. An explanation for the low penetrance of tritan defects has yet to be found, but one interesting possibility is that the S cones degenerate over time, analogous to the degeneration of rods in adRP. If so, the tritan phenotype would be a function of age, reflecting a progressive loss of S cones over time. The low penetrance aspect of the disorder may simply reflect that younger observers have not yet lost enough S cones to manifest symptoms. The prevalence of inherited tritan defects have been reported to be quite low, but they may be grossly underestimated for a variety of reasons including the fact that standard color vision tests do not test for tritan defects, they are extremely difficult to test for, and the phenotype may be age-dependent. The true incidence of inherited tritan defects and the ultimate fate of the S cones in affected individuals must await further experimentation. A particularly exciting prospect is the application of cutting edge imaging technologies using adaptive optics to the study of the retinal architecture in tritan subjects.
ACHROMATOPSIA
Also extremely rare are the monochromatic color vision defects known as achromatopsias. These disorders are associated with reduced or absent cone function, denoted as incomplete and complete achromotopsia, respectively. Blue cone monochromacy is a form of incomplete achromatopsia in which affected individuals base their vision on S cones and rods, and thus have diminished capacity for all aspects of vision mediated by cones including color vision and acuity. Rod monochromacy is a form of complete achromatopsia in which vision is mediated only by rods. Affected individuals are completely colorblind and have very poor acuity. Achromatopsia has been reported to affect fewer than one in 30000 individuals.[29]
BLUE CONE MONOCHROMACY
Genetically, blue cone monochromacy is a heterogeneous disorder, but in all cases the underlying cause is loss of function of both L and M cones. One major cause of blue cone monochromacy is the deletion of a critical DNA element known as an enhancer or LCR responsible for facilitating the expression of the L and M opsin genes. In the absence of the LCR, none of the X-chromosome opsin genes are expressed normally, and thus functional L and M cones are not produced.[35-37] The second major cause of blue cone monochromacy is the deletion of all except one of the X-chromosome opsin genes and the presence of an inactivating mutation in the remaining gene.[35,37] The most common mutation is the C203R mutation that has been found in conjunction with red-green color vision defects. Blue cone monochromats who in psychophysical tests appear to have more than one class of functional cone have been reported.[75-76] A complete understanding of this disorder must await further experimentation.
COMPLETE ACHROMATOPSIA AND FORMS OF INCOMPLETE ACHROMATOPSIA OTHER THAN BLUE CONE MONOCHROMACY
Although achromatopsias are extremely rare, specific human populations have been identified that exhibit an extraordinarily high incidence of the disorder. One example is autosomal recessive incomplete achromatopsia which has a prevalence of 5% among the Pingelapese islanders in Micronesia.[77] The underlying genetic cause of the disorder among the Pingelapses is an amino acid substitution in the beta subunit of the cyclic-GMP gated ion channel. Phototransduction in all three cone types relies on the function of the same cyclic-GMP gated ion channel, which has two subunits, the alpha subunit encoded by the CNGA3 gene on chromosome 2, and the beta subunit encoded by the CNGB3 gene on chromosome 8. Mutations in the gene encoding the alpha subunit have also been found in families with rod monochromacy, and in patients with incomplete achromatopsia.[78] Patients with incomplete forms of achromatopsia appear to have residual cone function whereas patients with the complete forms do not, implying that not all of the mutations identified completely abolish channel function.
COLOR APPEARANCE
The human eye is popularly described as being capable of discriminating as many as 10 million 'colors'. This is offered as a reason that computer displays are made to be capable of displaying 256 intensities for each of the red, green, and blue channels which makes the total number of possible 'colors' for each pixel equal to 16777216 (often approximated as 16 million). In this context, differences in 'color' include differences in brightness, hue, and saturation. It has long been understood that the 'millions' of colors humans can discriminate represent subtle gradations of a much smaller set of basic sensations. For example, most people agree that all color experience can be described using eleven basic color terms which in English are white, black, red, green, yellow, blue, brown, gray, orange, purple, and pink. Among these 'basic' colors, theorists agree that some are more fundamental than others. For example, pink might be described as a very pale red, brown - a very dark orange, and purple - a reddish-blue. Seven of the basic colors-red, green, blue, yellow, black, white, and gray-seem to be truly fundamental in that each of these sensations seem to be unique and not describable as a combination of the others.
Thus, color experience can be reasonably explained as the combination of six unique sensations. Accordingly, navy blue is the simultaneous sensation of blue and black. Pale violet is the combined sensation of white, red, and blue. The seventh 'color', gray, in this scheme is the absence of all color sensation. Unique hues were first described by Hering.[79] He also noted that red and green are opposite hues because they cannot be elicited simultaneously by a single color stimulus. There is no reddish-green color nor is there a bluish-yellow. Blue and yellow form a second opponent pair. The same is true of black and white if one accepts that gray is not the simultaneous sensation of black and white but rather the absence of either. Hering's concept of opponency consequently organizes the six fundamental sensations into the activity of three pairs of opponent sensations, black and white, red and green, and blue and yellow. In systems designed to represent all the possible hues on a continuous surface such as Figure 123.11, the four fundamental hues standout as unique while all other colors are seen as blends of the unique hues.
|
|
|
|
FIGURE 123.11 Fan of colors in three-dimensional pigment space. The fan represents the signal that various colored lights evoke in the S, M, and L cones. |
The facts of our vision are that three types of cone photoreceptor are responsible for three pairs of sensations. The question is: how are the signals from the three types of receptors combined in the nervous system to yield the three opponent pathways? A great deal is known about the anatomy and physiology of the visual system that can be brought to bear on that question, but many aspects of the neural circuits for coding color remain puzzling. In the task of understanding the neural operations responsible for transforming the cone signals into perception, a simplification is that of considering reduced color vision systems with fewer cone types and fewer fundamental sensations.
THE CIRCUITRY FOR CODING BLACK-WHITE, RED-GREEN, AND BLUE-YELLOW
Information from the retina is carried to higher centers via the axons of ganglion cells. At the ganglion cell stage, the responses of receptors has already been combined by postreceptoral circuitry to produce neurons with specialized response properties. As introduced in Chapter 111, this processing begins at the output terminal of the cone itself, the cone pedicle. Every cone pedicle receives a lateral inhibitory input from surrounding cones via horizontal cells (Fig. 123.12). Thus, for a cone near the fovea, the response communicated to postsynaptic bipolar cells is the result of a cone's own electrical response to light and the opposing responses of its neighboring cones that arrive via horizontal cell input. Thus, at its synaptic terminal (pedicle), every cone compares the number of quanta it absorbs as opposed to the average number of quanta absorbed by its neighbors. Assuming that the potential change at the pedicle produced by a cone absorbing light is evenly balanced with the opposing input from the average of its neighbors, then, if the number of quanta absorbed by a cone is greater than the average absorbed by its neighbors, it hyperpolarizes. If a cone absorbs fewer quanta than its average neighbor it depolarizes. Most importantly, when the number of quanta absorbed by a cone is somewhere near equal to its average neighbor, no signal is transmitted, i.e., the two opposing inputs null. Thus, a cone does not signal information about its photon catch, rather it signals information about the number of photons it has absorbed relative to the average number it's neighbors have absorbed. In the central retina a cone's neighbors would include an average of six cones that immediately surround it and cones more distant tiers away; however, the strength of the signal falls off exponentially with distance making the nearest neighbors the most important.
|
|
|
|
FIGURE 123.12 The wiring diagram of midget ganglion cells in the human fovea. In the central retina, each midget bipolar cell receives input from a single middle-(M) or long-wavelength sensitive (L) cone, and contacts a single midget ganglion cell. This one-to-one organization allows the signal from each M or L cone to be transmitted to higher brain regions. However, at the cone terminal prior to the transmission to the bipolar cell, a cone's response is compared in an opponent fashion to the responses of the neighboring cones. This comparison is achieved via horizontal cells that provide lateral reciprocally inhibitory interconnections between all cones. Every cone is served by two midget ganglion cells, one ON-center and one OFF-center. Neither will respond to stimuli, such as diffuse uniform white light, that produce activity in a cone that is equal to the average activity of its neighbors. |
The central retina near the fovea, the region that serves highest visual acuity, more than 90% of ganglion cells are of the 'midget' variety, so called because of their small dendritic arbor, which in the central 7-10° connects to a single cone via a 'midget' bipolar cell. Each cone provides input to many ganglion cells each presumably specialized to carry specific information about the visual stimulus. These include one of each of two midget ganglion cell subtypes for each cone (Fig. 123.12), one ON-center and one OFF-center which receive input through a corresponding pair of ON- and OFF-midget bipolar cells that have opposite responses to the neurotransmitter, glutamate (Chapter 112), released by cones. While some types of ganglion cells are specialized to carry information collected from many cones, each midget ganglion cell of the central retina is specialized to transmit information from a single cone.
By virtue of its horizontal cell interconnections, a cone compares the number of quanta it absorbs opposed to the average number of quanta absorbed by its neighbors. The ON-center midget ganglion cell is specialized to signal with an increased rate of action potentials when its cone absorbs a greater number of photons than the average absorbed by its neighbors as would happen when a relatively lighter region of an image falls on a cone while its neighboring cones 'see' adjacent darker parts of the image. The OFF-center ganglion cell is specialized to signal the reverse, i.e., when a cone absorbs fewer quanta than its average neighbor as when a relatively darker region of the scene falls on the cone. When the number of quanta absorbed by a cone is equal to its average neighbor, no signal is transmitted, i.e., the two opposing inputs null and both ON- and OFF-center ganglion cells fire at their spontaneous rates.
The midget ON and OFF midget gangion cells have all the qualities to serve as the biological substrate for the opponent percepts of black and white as proposed by Hering. This is particularly evident for individuals with a red-green color vision defect who have a reduced number of cone types from three to two and they have a reduced number of fundamental color sensations. As explained above, individuals with only two spectrally different types of cones are refered to as dichromats. The prefix 'di-' in dichromacy does not refer to two types of cone. Dichromacy means, literally, two hues and derives from the fact that dichromats can match any color using mixtures of just two 'primary' hues. However, the term dichromat is also appropriate because dichromats see only two hues. To them, objects are black, white, shades of gray, or one of two hues. The image in Figure 123.13 has been digitally altered to simulate the appearance for a dichromat. For the dichromat there is only black and white and blue and yellow. In contrast, people with normal color vision see more than 100 different hues in addition to black, white, and gray. Dichromats confuse red with green, and they confuse, with red and green, all colors in the spectrum that fall between them, including yellow, orange, and brown. They see blue and violet as the same color, and blue-green is indistinguishable from white or gray. Magenta and its pastel counterpart pink also appear white or gray.
|
|
|
|
FIGURE 123.13 (a) A digital image of fruit that has been digitally processed in subsequent panels to represent the activity of different subsystems for coding color in the human visual system. (b) The image processed to appear as it would to a dichromat having color vision based on just S and L cones. The dichromat has only four color sensations, blue, yellow, black, and white. Because blue and yellow are processed in an opponent fashion, for a dichromat, there are no intermediate hues that represent the simultaneous sensation of both blue and yellow. (c) Digitally resampling the image at a much lower spatial resolution removes the fine details from the image and significantly degrades its quality. (d) The black and white components are separated from (e) the hue components of the image. Here the hue components have been resampled with the same reduced resolution as in panel (c). (f) The image is reconstituted combining the high resolution black and white components of (d) with the low resolution hue components of (e). The spatial resolution of the hue components (b) and (f) are very different but on casual inspection they appear nearly identical. If you look closely you can see that the yellow color is blurred outside the black lines. This is largely ignored by the visual system that uses the black and white edges to define detailed boundaries in the image. The hue information is used to define qualities of the objects other than spatial detail. |
During evolution, the midget ganglion cell system is believed to have arisen in an ancestor to modern primates prior to the emergence of trichromatic color vision. Like most other mammals, that ancestor presumably had two cone types, S cones and a second type of cone sensitive in the middle-to-long wavelengths (Fig. 123.7b). The midget ganglion cells do not receive direct center input from S cones. Thus, in a dichromatic ancestor to humans, the midget ganglion cell system's major function was to compare absorptions of middle-to-long wavelength cones with their neighbors. The neighbors of an L/M cone would have been predominately of the identical spectral type because S cones are many times fewer in number (?5% of the total in humans). Assuming a balanced center versus surrounds, this system signals the presence of light/dark boundaries in a visual scene. If a dark area in an image falls on a cone, but adjacent lighter areas illuminate neighboring cones, the OFF-center midget ganglion will signal with increased firing. Conversely, a light area bounded by darker regions will be signaled by increased firing of the ON-center ganglion cell. These are exactly the stimulus conditions we associate with the percepts of black and white.
In addition to serving the percepts of black and white, this system presumably evolved to serve high acuity spatial vision, signaling the presence of dark/light edges with great detail and precision. Representing more than 90% of the ganglion cells that serve the central retina, the midget ganglion cells are the only output neurons that have sufficient numbers to transmit information about fine detail in the image. Accordingly, the midget ganglion cells appear to simultaneously serve two purposes in vision for a dichromat. The ON- and OFF-center midget ganglion cells form the basis for the percepts of black and white, respectively. In turn, the percepts of black and white in the dichromatic image are responsible for coding information about fine detail in the image. This is evident when the black and white of an image (Fig. 123.13c) is separated digitally from its color, all the detail of the image is preserved.
We use information about hue to find out about the internal qualities of objects. For example, a trichromat can tell that a banana is ripe by its yellow color. However, to determine its ripeness we do not need information about the fine detail of the spatial distribution of the yellow color. This is true of all the information we extract from the hue of objects. We know that it is evening because of the red of the sunset or that a colleague is embarrassed because of the redness in his face. However, we do not need to know details of the spatial distribution of the colored areas. Even when we use hue to locate objects in visual search, the color can 'catch our eye' in the absence of the spatial details of the object. The S cones are used in color vision but they do not participate in providing high acuity. They can participate in providing hue information at a much lower sampling density than the L and M cones that provide information about spatial detail. The unimportance of extracting color vision with high spatial resolution is illustrated in Fig. 123.13. In Fig 123.13e, the hue components of the image were separated and resampled at a much lower resolution. These low resolution hue components provide almost no information that allows us to recognize the objects in the scene. However, when the low quality hue components are recombined with (d), the black and white components of the image, (Fig. 123.13f) the result is almost imperceptibly different than the original image (b).
It is not completely clear which ganglion cell subtypes are specialized for carrying information corresponding to Herring's blue-yellow opponent channel. Small bistratified ganglion cells make up less than 10% of the total output of the retina and they have opponent blue-yellow responses. Thus, they are candidates to have some role in blue-yellow color vision. However, they all respond to blue and inhibit to yellow, leaving us without an explanation of the physiological basis for the yellow half of the blue-yellow opponent system. It is possible that a subpopulation of midget ganglion cells is also involved in blue-yellow color vision and they could be important for providing the required two opponent parts of the system.
The story would be relatively simple if humans were all dichromats. If this were true, details of the specialized functions of each of three major ganglion cell populations carrying information to the brain would be easily understood. The midget ganglion cells provide the biological substrate for percepts of black and white and are responsible for carrying information about high acuity spatial detail. The second major ganglion cell population, the parasol cells, are responsible for the perception of motion. These make up only ?5% of ganglion cells in the fovea but are a much larger proportion of ganglion cells in the peripheral retina where our motion perception is most acute. Although details are left to be worked out, the third major ganglion cell population, the small bistratified ganglion cells must participate in blue-yellow color vision. However, the introduction of red-green color vision complicates the story because adding a randomly distributed third cone type makes almost every midget ganglion cell respond to either diffuse red or diffuse green light in addition to dark and light edges. A guiding principle in our understanding the biological basis of sensation for more than 150 years is Johannes Müller's (1833) 'Law of Specific Nerve Energies' which states "Qualitatively different sensations must derive from different organs." Accordingly, even though the midget ganglion cells in trichomats respond to both diffuse red and green lights and black and white edges, ultimately the two percepts must be separated from each other at a higher level of visual processing, into two perceptually opponent processes of black and white and red and green. How this is accomplished is one of the fascinating remaining mysteries of our visual system.
REFERENCES
1. Green DG, Powers MK, Banks MS: Depth of focus, eye size and visual acuity. Vision Res 1980; 20:827.
2. VanMeeteren A: Calculations on the optical modulation transfer function of the human eye for white light. Opt Acta 1974; 15:47.
3. Campbell FW, Gregory AH: Effect of size of pupil on visual acuity. Nature 1960; 187:1121.
4. Campbell FW, Green DG: Optical and retinal factors affecting visual resolution. J Physiol 1965; 181:576.
5. Green DG: Visual acuity: The influence of refraction and diffraction and the use of interference fringes. Int Ophthalmol Clin 1978; 18:21.
6. Green DG: Regional variations in the visual acuity for interference fringes on the retina. J Physiol 1970; 207:351.
7. Le Grand Y: Form and space vision, Bloomington, Indiana University Press, 1967.
8. Thibos LN, Cheney FE, Walsh DJ: Retinal limits to the detection and resolution of gratings. J Opt Soc Am A 1987; 4:1524.
9. Lovegrove WJ, Bowling A, Badcock D, Blackwood M: specific reading disability: differences in contrast sensitivity as a function of spatial frequency. Science 1980; 210:439.
10. Hyvarinen L, Rovamo J, Laurinen P, Peltomaa A: Contrast sensitivity function in evaluation of visual impairment due to retinitis pigmentosa. Acta Ophthalmol 1981; 59:763.
11. Hyvarinen L: Contrast sensitivity in visually impaired children. Acta Ophthalmol 1983; 157(Suppl):58.
12. Hyvarinen L, Rovamo J, Laurinen P, et al: Contrast sensitivity in monocular glaucoma. Acta Ophthalmol 1983; 61:742.
13. Hyvarinen L, Laurinen P, Rovamo J: Contrast sensitivity in evaluation of visual impairment due to macular degeneration and optic nerve lesions. Acta Ophthalmol 1983; 61:161.
14. Hyvarinen L, Laurinen P, Rovamo J: Contrast sensitivity in evaluation of visual impairment due to diabetes. Acta Ophthalmol 1983; 61:94.
15. Koskela PU, Hyvarinen L: Contrast sensitivity in amblyopia. IV. Assessment of vision using vertical and horizontal gratings and optotypes at different contrast levels. Acta Ophthalmol 1986; 64:570.
16. Gaskill JD: Linear systems, Fourier transforms, optics, New York, John Wiley, 1978.
17. Bodis-Wollner I: Visual acuity and contrast sensitivity in patients with cerebral lesions. Science 1972; 178:679.
18. Brown B: Reading performance in low vision patients: relation to contrast and contrast sensitivity. Am J Optom Physiol Opt 1981; 58:218.
19. Rubin GS, Legge GE: Psychophysics of reading. VI. The role of contrast in low vision. Vision Res 1989; 29:79.
20. Pirenne MH, Denton EJ: Accuracy and sensitivity of the human eye. Nature 1952; 170:1039.
21. Shlaer S, Smith E, Chase A: Visual acuity and illumination in different spectral regions. J Gen Physiol 1942; 25:553.
22. Mendelbaum J, Sloan LL: Peripheral visual acuity. Am J Ophthalmol 1947; 30:581.
23. Yarbus AL: The perception of an image -xed with respect to the retina. Biophysics 1957; 2:683.
24. Riggs LA, Tulunay SU: Visual effects of varying the extent of compensation for eye movements. J Opt Soc Am 1959; 49:741.
25. Barlow HB: Slippage of contact lenses and other artefacts in relation to fading and regeneration of supposedly stable retinal images. Q J Exp Psychol 1963; 15:36.
26. Tulunay-Keesey U, Jones RM: The effect of micromovements of the eye and exposure duration on contrast sensitivity. Vision Res 1976; 16:481.
27. Aquilar M, Stiles WS: Saturation of the rod mechanisms of the retina at high levels of stimulation. Opt Acta 1954; 1:59.
28. Rushton WAH: The sensitivity of rods under illumination. J Physiol 1965; 178:141.
29. Sharpe LT, Stockman A, Jägle H, Nathans J: Opsin genes, cone photopigments, color vision, and color blindness. In: Gegenfurtner KR, Sharpe LT, ed. Color vision: from genes to perception, New York: Cambridge University Press; 1999:3.
30. Dean M, Carrington M, O'Brien SJ: Balanced polymorphism selected by genetic versus infectious human disease. Ann Rev Genom Human Genet 2002; 3:263.
31. Nathans J, Piantanida TP, Eddy RL, et al: Molecular genetics of inherited variation in human color vision. Science 1986; 232:203.
32. Jacobs GH: The distribution and nature of colour vision among the mammals. Biol Rev 1993; 68:413.
33. Jacobs GH: A perspective on color vision in platyrrhine monkeys. Vision Res 1998; 38:3307.
34. Jacobs GH, Neitz J: Color vision in monkeys: sex related differences suggest the mode of inheritance. Vision Res 1985; 25:141.
35. Nathans J, Davenport CM, Maumenee IH, et al: Molecular genetics of blue cone monochromacy. Science 1989; 245:831.
36. Wang Y, Macke JP, Merbs SL, et al: A locus control region adjacent to the human red and green visual pigment genes. Neuron 1992; 9:429.
37. Nathans J, Maumenee IA, Zrenner E, et al: Genetic heterogeneity among blue-cone monochromats. Am J Hum Genet 1993; 53:987.
38. Smallwood PM, Wang Y, Nathans J: Role of a locus control region in the mutually exclusive expression of human red and green cone pigment genes. PNAS 2002; 99:1008.
39. Hagstrom SA, Neitz M, Neitz J: Cone pigment gene expression in individual photoreceptors and the chromatic topography of the retina. J Opt Soc Am A Opt Image Sci Vis 2000; 17:527.
40. Onishi A, Koike S, Ida M, et al: Dichromatism in macaque monkeys. Nature 1999; 402:139.
41. Hayashi T, Motulsky AG, Deeb SS: Position of a 'green-red' hybrid gene in the visual pigment array determines colour-vision phenotype. Nat Genet 1999; 22:90.
42. Bollinger K, Sjoberg S, Neitz M, Neitz J: Topographical cone photopigment gene expression in deutan-type red-green color vision defects. Vision Res 2004; 34:135.
43. Sjoberg SA, Neitz M, Balding SD, Neitz J: L-cone pigment genes expressed in normal colour vision. Vision Res 1998; 38:3213.
44. Shevell SK, He JC, Kainz PM, et al: Relating color discrimination to photopigment genes in deutan observers. Vision Res 1998; 38:3371.
45. Neitz M, Neitz J: A new mass screening test for color-vision de-ciencies in children. Color Res Appl 2001; 26:S239.
46. Jagla WM, Jägle H, Hayashi T, et al: The molecular basis of dichromatic color vision in males with multiple red and green visual pigment genes. Hum Mol Genet 2002; 11:23.
47. Neitz M, Carroll J, Renner A, et al: Variety of genotypes in males diagnosed as dichromatic on a conventional clinical anomaloscope. Vis Neurosci 2004; 21:205.
48. Kainz PM, Neitz M, Neitz J: Molecular genetic detection of female carriers of protan defects. Vision Res 1998; 38:3365.
49. Neitz J, Neitz M, He JC, Shevell SK: Trichromatic color vision with only two spectrally distinct photopigments. Nat Neurosci 1999; 2:884.
50. Deeb SS, Lindsey DT, Hibiya Y, et al: Genotype-phenotype relationships in human red/green color-vision defects: molecular and psychophysical studies. Am J Hum Genet 1992; 51:687.
51. Carroll J, Neitz M, Neitz J: Testing hypotheses about visual pigments underlying deutan color vision. Color Res Appl 2001; 26:S106.
52. Neitz M, Neitz J: Molecular genetics of color vision and color vision defects. Arch Ophthalmol 2000; 118:691.
53. Winderickx J, Sanocki E, Lindsey DT, et al: Defective colour vision associated with a missense mutation in the human green visual pigment gene. Nat Genet 1992; 1:251.
54. Karnik SS, Khorana HG: Assembly functional rhodopsin requires a disul-de bond between cysteine residues 110 and 187. J Biol Chem 1990; 265:17520.
55. Kazmi MA, Sakmar TP, Ostrer H: Mutation of a conserved cysteine in the X-linked cone opsins causes color vision de-ciencies by disrupting protein folding and stablilty. Invest Ophthalmol Vis Sci 1997; 38:1074.
56. Nathans J, Thomas D, Hogness DS: Molecular genetics of human color vision: the genes encoding blue, green, and red pigments. Science 1986; 232:193.
57. Winderickx J, Battisti L, Hibibya Y, et al: Haplotype diversity in the human red and green opsin genes: evidence for frequent sequence exchange in exon 3. Hum Mol Genet 1993; 2:1413.
58. Neitz M, Neitz J, Grishok A: Polymorphism in the number of genes encoding long-wavelength sensitive cone pigments among males with normal color vision. Vision Res 1995; 35:2395.
59. Carroll J, McMahon C, Neitz M, Neitz J: Flicker-photometric electroretinogram estimates of L : M cone photoreceptor ratio in men with photopigment spectra derived from genetics. J Opt Soc Am A Opt Image Sci Vis 2000; 17:499.
60. Carroll J, Neitz M, Hofer H, et al: Functional photoreceptor loss revealed with adaptive optics: an alternate cause of color blindness. Proc Natl Acad Sci USA 2004; 101:8461.
61. Neitz M, Neitz J, Jacobs GH: Spectral tuning of pigments underlying red-green color vision. Science 1991; 252:971.
62. Asenjo AB, Rim J, Oprian DD: Molecular determinants of human red/green color discrimination. Neuron 1994; 12:1131.
63. Merbs SL, Nathans J: Absorption spectra of the hybrid pigments responsible for anomalous color vision. Science 1992; 258:464.
64. Neitz J, Neitz M, Kainz PM: Visual pigment gene structure and the severity of human color vision defects. Science 1996; 274:801.
65. Regan BC, Ref-n JP, Mollon JD: Luminance noise and the rapid determination of discrimination ellipses in colour de-ciency. Vision Res 1994; 34:1279.
66. Wright WD: The characteristics of tritanopia. J Opt Soc Am 1952; 42:509.
67. Kalmus H: The familial distribution of congenital tritanopia. Ann Hum Genet 1955; 20:39.
68. Pokorny J, Smith VC, Went LN: Color matching in autosomal dominant tritan defect. J Opt Soc Am 1981; 71:1374.
69. Went LN, Pronk N: The genetics of tritan disturbances. Hum Genet 1985; 69:255.
70. Weitz CJ, Miyake Y, Shinzato K, et al: Human tritanopia associated with two amino acid substitutions in the blue sensitive opsin. Am J Hum Genet 1992; 50:498.
71. Weitz CJ, Went LN, Nathans J: Human tritanopia associated with a third amino acid substitution in the blue sensitive visual pigment. Am J Hum Genet 1992; 51:444.
72. Gunther KL, Neitz J, Neitz M: A novel mutation in the short-wavelength sensitive cone pigment gene associtated with a tritan color vision defect. Vis Neurosci 2006; 23:403.
73. Curcio CA, Allen KA, Sloan KR, et al: Distribution and morphology of human cone photoreceptors stained with anti-blue opsin. J Comp Neurol 1991; 312:610.
74. Hofer H, Carroll J, Neitz J, et al: Organization of the human trichromatic cone mosaic. J Neurosci 2006; 26:722.
75. Smith VC, Pokorny J, Delleman JW, et al: X-linked incomplete achromatopsia with more than one class of functional cones. Invest Ophthalmol Vis Sci 1983; 24:451.
76. Crognale MA, Fry M, Highsmith J, et al: Characterization of a novel form of X-linked incomplete achromatopsia. Vis Neurosci 2004; 21:197.
77. Sundin OH, Yang JM, Li Y, et al: Genetic basis of total colourblindness among the Pingelapese islanders. Nat Genet 2000; 25:289.
78. Wissinger B, Gamer D, Jägle H, et al: CNGA3 mutations in hereditary cone photoreceptor disorders. Am J Hum Genet 2001; 69:722.
79. Hering E: In: Hurvich LM, Jameson D, ed. Outlines of a theory of the light sense, Cambridge: Harvard University Press; 1964.
