Studies on the Physiological Color Blindness of the Human Fovea with the Polarization Method

Home , Color anomaly, The Red and the Blue

STUDIES ON THE PHYSIOLOGICAL COLOR BLINDNESS OF THE HUMAN FOVEA WITH THE POLARIZATION METHOD

KOITI MOTOKAWA, MITSURU EBE, YASUHIRO ARAKAWA AND TOSHIHIKO OIKAWA * Department of Physiology, Tohoku University,Sendai

Using a small test field viewed centrally, Konig (1) could match any radi- ation by a suitable mixture of two radiations such as 650mμ and 475mμ, and ar- rived at the conclusion that the center of the fovea was blue-blind. Hartridge

(2) observed the confusion of blue with dark grey or dark, and yellow with pale grey or white, when the visual angle of a test object was reduced. Willmer (3) studied the same phenomenon, experimenting with small colored patches painted on cards and using suitable fixation spots and viewing technique, and came to a conclusion similar to Konig's namely that the central fovea was tritanopic. The phenomenon was studied more extensively and quantitatively by Wright (4, 5) in conjunction with Willmer with modern color matching devices. Ebe, Isobe and Motokawa (6) analyzed the retinal processes of color-blind subjects by means of Motokawa's method (7) and found that the data thus obtained could not be accounted for on the basis of the three-color theory. In the present experiment, we analyzed, by means of the same technique, the color processes at the fovea of the retina evoked by microillumination, in order to elucidate the mechanism of the physiological color blindness under such experimental conditions.

METHOD

The eye shows a supernormal sensitivity to an electric stimulus after ex-

posure to light. This property was utilized by Motokawa (7, 8) for analyzing retinal processes of color vision. In the present experiment, measurements were carried out with a test patch of reduced visual angle (2minutes of arc), instead of a greater one of 2° used in the experiments by Motokawa. The apparatus for microstimulation was a reducing lens system similar to that used by Hartridge (9, 10). As illustrated schematically in fig. 1, a real image of a small patch of ground glass illuminated from behind, was focused at R by means of a microscope, M, and viewed by the subject from a distance of 90cm. The microscope was placed in such a manner that the ocular (Oc) faced toward the patch. The size of the real image was measured accurately in order to know the reduction rate. A minute light spot was placed at a distance of 30minutes in visual angle from the image, R, so as to serve as a

Received for publication June 26, 1951. *本川弘一 ,江部充,荒川安廣,及川俊彦

50 PHYSIOLOGICAL COLOR BLINDNESS OF FOVEA 51

fixation mark. When it was necessary to stimulate the center of the fovea, two light spots were presented on both sides of the center of the fovea so as to make the subject gaze at the point half-way between the two fixation marks.

Fig. 1. Apparatus used for microstimulation. F fixation mark, M microscope, Ob objective, Oc ocular, P patch of ground glass, R real image of P, S spectroscope, Sc screen.

The intensity of the equal energy spectrum used in our experiment was such that only the middle range from about 640mμ to 590mμ was visible; the other parts of the spectrum evoked no light sensation. Never the less, the responses to these subthreshold lights could be determined by our method with the same degree of accuracy as those to the superthreshold lights. The details of the apparatus and the procedure for electrical stimulation of the eye are described in a paper by Motokawa (7). After a preliminary dark adaptation of about 20minutes, the electrical excitability of the eye was measured by applying single constant current pulses of O.1sec. in duration to the eye, the index of excitation being the least perceptible electric phosphene. Stimulating voltages were reduced step by step until the subject could no more distinguish an electric phosphene from the background of intrinsic light of the retina. For raising accuracy of such determinations, it was found absolutely necessary to compare the effect of a weak electric stimulus and that of a con- trol stimulus far below the threshold. Similar measurements were done at vary- ing moments after 2sec. exposure of the eye to spectral light, in order to de- termine the time course of supernormal excitability following the light stimulus. Excitability increases were expressed in percentage of the excitability at the resting level. It is to be noted that in such measurements, more than one threshold values could be obtained for a certain phase of supernormality. This is due to con- current excitation of several receptors with different thresholds, as was eluci- dated by Motokawa (11). In such cases, we adopted the lowest threshold value for construction of an excitability curve which represents the time course of electrical excitability after an illumination.

RESULTS 1. Excitability curves obtained from the fovea by means of microstimulation Retinal responses to microstimulation are expressed in fig. 2 in form of excitability curves, in which excitability increases are plotted as ordinates against 52 K. MOTOKAWA ET AL.

time after termination of the light stimulus as abscissas. The curves generally consist of two elevations, one of which has a maximum at about 1 second and the other at about 2.5 or 3 seconds. These humps are, however, not so con- spicuous in the curve for white and yellow pre-illuminating lights as in the other curves. According to the analysis by Motokawa (7, 8), the excitabilitycurves for red, yellow, green and blue receptors of normal trichromats show a maximum at 1, 1.5, 2.25 and 2.75 or 3 seconds respectively. If the result of Motokawa's

Fig. 2. Excitability curves obtained at 20minutes from foveal center. Ordinates: percentage increases of electrical excitability above resting level. Abscissas: time in sec. after end of pre- illumination. Figure by each curve indicates wave-length of light used for pre- illumination. Zero-level of each curve is given on left side.

analysisis applicableto the data obtained with smaller test fields,then from the excitabilitycurves shown in fig.2, it can be said how the test patch appeared to the subject in this experimemt. Since the firstmaximum of these curves liesat 1 second, it is certain that the red receptorwas excited by micro- stimuiation. The second maximum located at about 2.5sec. can be considered as representingan envelope of green and blue processes which have a maximum at 2.25sec. and at 2.75 or 3sec. respectively.Hence, this hump must subserve a sensation of blue-green,and according to whether the firstor the second hump is predominant, the test patch must have appeared red or blue-green. As a matter of fact,Hartridge (9, 10) reported that a small testobject appeared blue-green if it was green, blue-green and indigo, and that red, orange and crimson appeared always red. In case the two humps are obscure as in the curves for white and yellow, sensations must have been neutral,because the two systems, red and blue- PHYSIOLOGICAL COLOR BLINDNESS OF FOVEA 53

green, which are complementary to each other, are considered to have been excited to the same extent in such cases. In fact, Hartridge reported that

yellow appeared colorless and that white looked always white. These obser- vations were confirmed by the subjects in our experiments. In the excitability curves for the lights from the violet end of the spectrum, the maximum of the second hump lies at 3 seconds, instead of 2.5 seconds (see the curve for 440mμ), and this finding suggests that the light from this part of the spectrum would appear blue or violet, instead of blue-green. However, the excitability curve for blue or violet light is generally so low that the sensation caused by it must be very slight; we have confirmed in a series of ex-

periments that the subject could not perceive the test patch unless the intensity of pre-illumination was such that the maximum height of the excitability curve obtained under these conditions surpassed a certain critical level. In most cases, the critical level was found about 7 in terms of percentage increase of electrical excitability. As Hartridge observed, a blue test object appears dark

grey or dark when it is sufficiently small in visual angle, and this fact seems to correspond to our finding that the excitability curve for blue or violet is lower than those for lights from the other parts of the spectrum.

It is apparent from this experiment that the red and the blue-green are the main processes involved. However, our excitability curves as such give no

more information about the exact nature of the processes involved than do any other sensory experiments, because an excitability curve represents an envelope of a train of processes developing at different velocities. Therefore, the

phenomenon was subjected to further analysis in the manner as described in Motokawa's papers (7, 8). 2. Spectral response curves A series of measurements was carried out, with a fixed interval of 1 second intervening between the end of the pre-illumination and the electric stimulus. By plotting the values so obtained as ordinates against the wave-lengths of the pre-illuminating lights as abscissas,we obtained a curve representing the spectral distribution of the red receptor. The curve R in fig.3 is an example. In a similar manner, the response curves for the yellow, green and blue receptors were obtained and denoted by Y, G and B respectively,where the interval between the light stimulus and the electric stimulus was fixed at 1.5, 2.25 and 3

Fig. 3. Spectral response curves at 20minutes from foveal center. R, Y, G and B represent spectral loci of electrical excitability measured 1, 1.5, 2.25 and 3 seconds after termination of light stimulus. 54 K. MOTOKAWA ET AL.

second respectively. As can be seen in fig. 3, the curve R shows a maximum at about 610mμ and another in the violet part of the spectrum. The maxima of the curves Y, G and B lie at 575mμ, 520mμ and 460mμ respectively. The curves Y and B are lower and narrower than the curves R and G. These data were obtained from an area about 30minutes from the center of the fovea Similar measurements were performed at various parts of the fovea with the result that the dominance of R and G over Y and B was found more pronounced as the measuring area moved from the periphery of the fovea towards the center. In fig. 4 the values of the 4 processes R, Y, G and B are expressed as a function of the distance from the center of the fovea. This figure indicates that the values of R and G in the fovea are decidedly higher than those in the

Fig. 4. Magnitudes of responses to microstimulation at various parts of retina. R, Y, G and B represent responses of red, yellow, green and blue receptors respectively.

parafovea, whereas the values of Y and B in the fovea are much lower than those in the parafovea. The magnitudes of Y and B increase hand in hand from the center towards the parafovea, while no such definite tendency to increase or decrease is found with the processes R and G at least within the fovea. In particular, it deserves attention that there is no measurable Y at the center of the fovea, and that the value of B, though smaller than in any other part of the fovea, is not so small as that of Y. In the polychromatic theory of Hartridge (10), it is assumed that there are two complementary pairs of receptors: the crimson and the blue-green; the yellow and the blue, and that when one member of a pair fallsout, owing to some alteration in the conditions of vision, the other member of the pair does so. So far as sensory data are concerned, this scheme is very convenient to account for various phenomena observed with a reduced intensity of illumination, with reduced visual angle subtended by a test object at the eye, peripheral color vision,etc., but it dose not PHYSIOLOGICAL COLOR BLINDNESS OF FOVEA 55

strictly apply to the data obtained in the above experiment, 1. because the

yellow receptor Y falls out of action at the center of the fovea, while the blue receptor B does not; and 2. because we have not yet succeeded in finding such a special blue-green receptor having a sensitivity maximum at 495mμ as assumed in Hartridge's theory. If the second maximum at 2.5seconds of the excitability curves were due to a special blue-green receptor, the response curve obtained by fixing the interval between the end of pre-illumination and the e- lectric stimulus at 2.5sec. would show a maximum at the blue-green part of the spectrum. In reality, however, the response curve determined under such

experimental conditions showed a maximum at the same part of the spectrum as the response curve G. Therefore, Hartridge's hypothetical blue-green receptor must at the present time be regarded as composed of tow kinds of receptors, green and blue. The postulate of the complementary pair, red and blue-green, seems to hold

good without any special blue-green receptor, for the blue receptor, which re-

mains active in the foveal center, makes, together with the green receptor , a complementary counterpart to the red receptor. The experiment of color-mixing by Willmer and Wright also suggests that the blue receptor plays some role in the so-called dichromatic vision of the foveal center; they could match any spectral light by a suitable mixture of red

(650mμ) and blue (460mμ) lights. This fact would be difficult to understand if the blue receptor were completely out of action at the foveal center like the

yellow receptor.

DISCUSSION

There is no doubt that there should be at least three kinds of receptors concerning color vision, but the question as to how many kinds of receptors exist in reality, will not be settled without further experiments. By means of his micro-electrode technique, Granit (12, 13, 14) could isolate from the retinas of various animals 8 kinds of modulators each with a narrow band of sensitivity at 8 different parts of the spectrum. This experiment hardly provides any evi- dence that as many kinds of receptors does exist in the retina of an animal. By means of selective adaptation and polarization of cat's retina, Granit showed that there were three different groups of modulators, whose maxima of spectral sensitivity were found at about 600,530 and 460mμ. Motokawa could distinguish three kinds of processes at the human fovea, using a test patch of 2° in visual angle. The spectral response curves for these three processes were found to have a maximum at about 585,550 and 460mμ. The maxima of the fundamental sensation curves based on color-matching experiments lie at 580, 540 and 440mμ (Wright); at 590, 540 and 450mμ (Konig); at 580, 540 and 460 mμ (Pitt). Thus, satisfactory agreement can be seen between the maxima of the fundamental sensation curves and the corresponding maxima of the response curves as determined by Motokawa's method. With the same method, Motokawa (8) obtained from outside the human fovea 4 kinds of processes having sensitivity maxima at 600-610, 580-570, 530-520 and 460mμ, and denoted these processes by R, Y, G and B respectively. The 56 K. MOTOKAWA ET AL.

process B is obviously identical with the corresponding one found at the fovea, because the maxima of their sensitivity are located at the same part of the spectrum. There is, however, at the fovea no process which would correspond to the process Y found outside the fovea. While the maximum of the response curve R is located at about 610mμ, that of the corresponding curve obtaind from the fovea lies at 585mμ. A similar difference can be found between the response curve G and the corresponding one obtained from the fovea; the maximum of the former lies at 520mμ, while that of the latter is located at about 550mμ. The question now arises as to whether or not the kinds of photopic receptors existing outside the fovea are really different from those at fovea. The answer to this question will be given by the following facts. 1. From the foveas of dichromats, response curves showing the same spectral distribution as those found in the periphery of normal trichromats were obtained in the experiment by Ebe, Isobe and Motokawa (6). 2. The maxima of the response curves obtained with a small test patch in the present investigation have been found to be located at the same parts of the spectrum as those found outside the fovea of a normal trichromat or in the foveas of dichromats. In view of these facts it may be supposed that it is difficult to obtain response curves in pure form from the fovea when a large test patch is used. The curve having a maximum at 585mμ may be considered to be composed of R and Y, and the curve having a maximum at 550mμ to be composed of Y and G. In experiments with a large test patch, it is very difficult to isolate the process Y from the neighboring

processes R and G, probably because the yellow process develops to the same extent as the latter two processes at the fovea under such experimental conditions. The isolation is, however, much easier with a small test patch, because the development of the process Y is so weak that it can easily be distinguished from the more dominant ones, R and G. Outside the fovea, the process Y can most easily be isolated, because it is the most dominant color process in this region of the retina. Such being the case, we obtain only three response curves, two of which are of composite nature, when measurements are done with a great test field at the fovea of a normal trichromat. The same seems to be true to the fundamental sensation curves which, in many respects, resemble the response curves of the fovea as determind with a large test field. As stated above, marked development of the yellow process was demon- strated by Ebe et al. in all types of color blindness, and this fact explains well the situation that the sensation of yellow is well preserved in protanopia and deuteranopia. The yellow receptor is,however, completely inactive in the physiological color blindness under consideration. As stated above, Konig, Willmer and Wright called the color blindness in question blue-blindness or tritanopia. This terminology may be adequate if by this term it is meant that the function of the blue receptor is incomplete. But if it implies that the bule receptor is entirely lacking as is assumed in the trichromatic theory, it is obviously wrong, because it is not the blue process but the yellow one that is entirely lacking in the center of the fovea. In general, no component is entirely lacking in congeni- PHYSIOLOGICAL COLOR BLINDNESS OF FOVEA 57

tal color blindness, although the development of one component is incomplete, viz., R in protanopia and G in deuteranopia. It must, however, be mentioned that in some cases of color anomaly both G and R are not normal in size as well as in distribution of spectral sensitivity. For comparison, the relative amounts of 4 components in all types of color blindness are shown schematically in fig. 5. As to the mechanism of the physio- (a) logical color blindness of the foveal center, NORMAL Konig (1) advanced the hypothesis that it would be due to the absorption of blue rays by the macular pigment. However,

the pigmentation lies,according to Polyak, (b) PHYSIOLOGICAL within the nerve tissue of the retina, so TRITANOPIA that, where the retinal membrane is thinnest, as at the foveal center, the colo- ration should be least. Hence, the apparent tritanopia must have its origin in (c) the nature of the foveal receptors, as was PROTANOPIA pointed out by Wright (5). Moreover, the fact that the reduction of activity of the blue receptor is associated with the

loss of activity of the yellow receptor (d) cannot be accounted for from this point DEUTERANOPIA of view. Willmer tried to interpret the apparent tritanopia of the foveal center on the basis of Roaf's theory that the rods play the part of the blue receptor during Fig. 5. Graphs showing relative magnitudes of 4 color components (re- photopic vision, for it follows from this theory that the rod-free part of the fovea ctangles) and luminosity curves (broken curves) in various types of color must be blue-blind, too. However, the vision (schematical representation). blue receptor remains more or less active within the rod-free part of the fovea, and even at the center of the fovea, it is not entierly out of action, as has been emphasized above. Furthermore, with Motokawa's method it is possible to distinguish very clearly the blue process from the rod-process (15, 16). There- fore, Willmer's hypothesis is obviously untenable. Next, attention should be directed to the important role played by the yellow receptor in the problem of luminosity. While the maximum of the luminosity curve of a normal trichromat lies at about 555mμ, that of a protanope lies at about 540mμ, and that of a deuteranope at about 565mμ. In color blindness, the deviation concerning luminosity from normal is thus comparatively small. It is very difficult to understand this fact from the standpoint of the classical three-color theory; if only two components G and B were to contribute to the luminosity curve in protanopia, then the maximum of the luminosity curve would be expected to lie at the green or green-blue part of the spectrum. In 58 K. MOTOKAWA ET AL.

reality, however, it lies at the yellow-green part of the spectrum. Similarly it is to be expected from the theory that the maximum of the luminosity curve lies at the orange part of the spectrum in deuteranopia, but in reality, it lies at yellow-green part of the spectrum. This behavior of luminosity curves will, however, be understood, if the fact is taken into account that the yellow receptor is well-developed at the foveas of dichromats. The type of luminosity curve for each kind of color blindness is indicated schematically by a broken curve in fig. 5. Such curves represent summation

of 4 components, but it was assumed in their construction that the component B is less effective in contribution to luminosity than the other components. As can be seen in (b) of fig. 5, it is to be expected that a notch is found in the luminosity curve obtained from the center of the fovea, owing to the lack of the

yellow receptor. As a matter of fact, Sloan found a notch in a luminosity curve determined with spectral lights of reduced intensity, and this observation was confirmed by Walters and Wright (17). The luminosity curve obtained by these authors showed two peaks, one with its maximum at about 600mμ and the other at about 540mμ, there being an intervening trough between them. A similar notch was found by Wright in the luminosity curve determined with a small test field (18). As to the mechanism of this phenomenon, Hartridge

(10) advanced a hypothesis that the notch in question is due to the dropping out of action of a specific yellow receptor with a reduction of intensity or visual angle, and in reality his hypothesis has received strong support from our present experiment.

SUMMARY It was shown by Konig, Willmer and Wright that the center of the fovea of normal man is blue-blind, when a test object of reduced visual angle is used. In order to elucidate the mechanism of this physiological color blindness, retinal processes were analyzed by means of Motokawa's method. 1. It was found that 4 kinds of color processes, red, yellow, green and blue exist at the fovea. 2. As the test patch was reduced in visual angle, the red and green processes became predominant over the yellow and blue processes. 3. At the center of the fovea there was no measurable Y, and B was found also small. 4. These findings account for Hartridge's observation that sensations of red and blue-green are predominant over those of yellow and blue, when small test fields are used. 5. The notch found in the luminosity curve, which was obtaind by Walters and Wright from the foveal center, is really due to the lack of the yellow receptor, as Hartridge surmised.

REFERENCES 1. KONIG, A. Sitz.B. Akad. Wiss. Berlin 1897: 559. 2. HARTRIDGE, H. J. Physiol.52: 175, 1918. 3. WILLMER, E. N. Nature 153: 774,1944. PHYSIOLOGICAL COLOR BLINDNESS OF FOVEA 59

4. WILLMER, E. N. AND WRIGHT, W. D. Nature 156: 119, 1945. 5. WRIGHT, W. D. Researches on normal and defective colour vision. London: Henry Kimpton, 88-95, 1946. 6. EBE, M., ISOBE, K. AND MOTOKAWA, K. Science 113: 353, 1951. 7. MOTOKAWA, K. J. Neurophysiol. 12: 291, 1949. 8. MOTOKAWA, K. J. Neurophysiol. 12: 465, 1949. 9. HARTRIDGE, H. Phil. Trans. B. 232: 519, 1947. 10. HARTRIDGE, H. Recent advances in the physiology of vision.London: Churchill, 1950. 11. MOTOKAWA, K. Tohoku J. Exp. Med. 55: in press, 1951. 12. GRANIT, R. Nature 151: 11, 1943. 13. GRANIT, R. J. Neurophysiol. 8: 195, 1945. 14. GRANIT, R. Ergeb. Physiol. 46: 31, 1950. 15. MOTOKAWA, K. AND EBE, M. Tohoku J. Exp. Med. 54: 215, 1951. 16. MOTOKAWA, K., EBE, M., ARAKAWA, Y. AND OIKAWA, T. J. Opt. Soc. Am. June, 1951. 17. WALTERS, H. V. AND WRIGHT, W. D. Proc. Roy. Soc. B. 131: 340, 1943. 18. WRIGHT, W. D. Nature 151: 726, 1943.