Spectral Sensitivity and Color Discrimination Changes in Glaucoma and Glaucoma-Suspect Patients

Spectral sensitivity and color discrimination changes in glaucoma and glaucoma-suspect patients

Anthony J. Adams, Rosemary Rodic, Roger Husted,* and Robert Stamper

Color vision changes may occur early in the course of glaucoma and may precede visual field loss. Glaucoma suspects, having raised intraocular pressure and no diagnostic optic nerve head or visual field changes, may also have color vision loss. Unfortunately, the instruments used in the studies that have demonstrated these color vision changes were not feasible for routine clinical use; likewise, the studies did not carefully control for the effects of small pupil size and age or did not point to the underlying mechanisms responsible. We studied 19 glaucoma patients, 19 glaucoma suspects, and age-matched controls for each group by means of the Farnsworth D-15 panel test, a desaturated version of the D-15 test, and by measures of spectral increment threshold. Minor modifications of the Farnsworth D-15 panel test produce highly significant differentiation of glaucoma and glaucoma-suspect patients from age-matched normal groups. Further, spectral increment thresholds, with a two-degree spectral target flashed at either 1 or 25 Hz on a bright white background, show that both achromatic and chromatic sensitivity are significantly reduced when compared with their age-matched normals. Pupil size does not seem to be a significant factor. These results suggest that the function of two different ganglion cell populations is affected in glaucoma and that glaucoma may produce functional loss in the central foveal area earlier in the disease process than previously believed. (INVEST OPHTHALMOL VIS SCI 23:516-524, 1982.)

Key words: color vision, glaucoma, glaucoma suspects, spectral sensitivity, luminosity, color vision testing, color discrimination

I t has been known for some time that color glaucoma. These changes are usually de- vision changes occur early in the course of scribed as "blue-yellow" defects, with some studies showing as much as 80% of glaucoma patients having disturbed color discrimina- From the School of Optometry, University of California, tion.1 Color vision defects may be apparent Berkeley, the Ophthalmology Clinic, Silas Hayes 2 3 Army Community Hospital, Fort Ord, and Depart- before visual field defects. ' Lakowski and 4 ment of Ophthalmology, Pacific Medical Center, San Drance have also shown that about 20% of Francisco, Calif. ocular hypertensive patients show severe Supported by NIH grant EY02271 to A. J. A. and the color vision defects; they consider losses in Dorothy Perry Research Fund and Research to Pre- color vision to be an important risk factor vent Blindness Grants to R. L. S. Submitted for publication Aug. 17, 1981. in the transition of ocular hypertension to Reprint requests: Anthony J. Adams, School of Op- chronic simple glaucoma. tometry, University of California, Berkeley, Berkeley, Unfortunately, the color vision tests that Calif. 94720. have been most sensitive in detecting these •Lt. Col., MC, USA. blue-yellow color vision losses are not feasi- The views expressed in this article are those of the au- thors and do not reflect official United States Army or ble for routine clinical use (e.g., FM 100 Hue Department of Defense policy. test, Pickford-Nicolson anomaloscope), are

516

Downloaded from iovs.arvojournals.org on 09/30/2021 Volume 23 Number 4 Color vision changes in glaucoma 517

Table I. Data for glaucoma subjects and age-matched normals

Diag- Pupil Age- Sub- Age nostic size matched Age ject Sex (w) Acuity results Field defect Treatment (mm) normal Sex (yr) D.J. M 17 20/15 Con- BS enlarge- Timolol 0.5% 4.0 T. G. M 20 genital ment C. F. M 20 20/15 OAG Bjerrum Epinephrine 1% 4.0 B. N. M 22 B. S. F 28 20/15 ACG Nasal step Iridectomy, epineph- 3.0 J. R. M 26 rine 1% S. B. M 40 20/15 OAG Bjerrum Timolol 0.5%, Ocusert 40, 4.0 I. B. M 40 epinephrine 1% J. C. M 42 20/20 Pigmen- Bjerrum Timolol 0.5% 3.5 S. G. F 42 tary S. K. F 44 20/15 OAG BS enlarge- Timolol 0.5% 4.0 S. P. F 44 ment E. W. M 48 20/20" OAG Bjerrum Pilocarpine 4%, timolol 2.0 R. M. M 47 0.5%, epinephrine 1% JJ- M 53 20/20 OAG Bjerrum Ocusert 40, epineph- 2.0 s. s. M 53 rine 1% S. H. F 57 20/20 OAG Early nasal Pilocarpine 2% 3.0 M. S. M 57 step P. H. M 58 20/15 OAG Early Bjerrum Ocusert 40, epinephrine 1% 3.0 C. F. F 58 C. C. M 59 20/20 OAG BS enlarge- Pilocarpine 4%, timolol 2.0 K. K. M 56 ment 0.5%, epinephrine 1% F. S. M 59 20/25= OAG Nasal step Timolol 0.5% 4.0 K. M. F 59 B. P. F 63 20/20 ACG BS enlarge- Timolol 0.5% 3.5 E. H. M 62 ment R. S. M 63 20/40+ OAG BS enlarge- Epinephrine 2%, timolol 4.0 K. H. F 62 ment 0.5% R. B. F 64 20/25+ ACG Bjerrum Pilocarpine 1% 2.0 R. P. M 64 F. S. M 64 20/40+2 OAG Bjerrum Timolol 0.5% 3.0 E. W. M 65 M. K. F 65 20/25+ OAG Bjerrum Acetazolamide 3.0 B. N. M 64 J. D. M 67 20/20 OAG Bjerrum Timolol 0.5%, pilocarpine 1.5 M. M. M 68 2% C. E. M 68 20/25 OAG Bjerrum Epinephrine 1%, timolol 3.0 R. W. M 72 0.5%

OAG = open-angle glaucoma; ACG = angle-closure glaucoma; BS = blind spot.

not designed for detecting blue-yellow de- large diameter (fast) and small diameter fects (e.g., Dvorine, Ishihara, AO pseudo- (slow) nerve fibers, respectively, of the gan- isochromatic plates), or appear to be rela- glion cell layer. Such fast and slow neural tively insensitive in detecting these defects systems have been described in a number of (e.g., AO-HRR, F-2 plate, TMC plates). Fur- species and have been linked to achromatic thermore, most studies of color vision with and chromatic activity in the rhesus mon- glaucoma patients deal with patients having a key.10 Selective loss of chromatic activity has wide range of acuities and are often compli- been shown in diabetics11 and in a case of cated by the color vision changes that accom- unilateral color vision loss.9 pany age and small pupils. Both of these lat- The primary purpose of this study was to ter factors can produce blue-yellow errors on examine the spectral increment threshold color vision tests.5"7 In fact, the underlying sensitivity of glaucoma and glaucoma-suspect mechanisms for the reported color discrimi- patients; increment threshold measures may nation changes have received little attention. be used to isolate chromatic from achromatic Fortunately there is a growing body of evi- signals in the visual pathways.8 A secondary dence that psychophysical vision tests can be goal was to use an established simple clinical fashioned to selectively test achromatic (lu- test of color vision on the same subjects and minance) and chromatic (color) sensitivity8' 9 to use modified scoring and testing proce- and that these two functions are mediated by dures that might be anticipated to reveal

Downloaded from iovs.arvojournals.org on 09/30/2021 Invest. Ophthalmol. Vis. Sci. 518 Adams et al. October 1982

' 3 ! i- RETERENCE o*

a. \ o

I 15

o 13 12 11

Fig. 1. Farnsworth dichotomous D-15 panel test, a, Typical test failure involving two crossings across the color circle on the left; the line joins the cap numbers in the order in which this patient arranged them. A test with no errors is indicated by a line joining the reference cap to 1 through 15, as shown on the right, b, Examples of error scores that reflect test failure in this study but which, by conventional scoring, do not indicate test failure. The score to the right shows two one-place errors, while that on the left shows a single two-place error. Both meet the failure criterion used for this study (see Methods).

more subtle disturbances of color discrimi- pared with their age-matched normal con- nation. To this end we used the Farnsworth trols. Furthermore, the measurement of dichotomous panel D-15 test and a desatu- spectral increment thresholds for the same rated version of it, as well as a modified scor- groups also demonstrated loss of both ach- ing procedure for both tests. romatic and chromatic sensitivity compared Our results suggest that simple mod- with the age-matched controls. ifications of a standard clinical test of color vision allows demonstration of a significant Methods loss of color discrimination of both glaucoma Subjects. The 19 glaucoma subjects (14 primary and glaucoma-suspect groups when com- open angle) were each individually age matched to

Downloaded from iovs.arvojournals.org on 09/30/2021 Volume 23 Number 4 Color vision changes in glaucoma 519

Table II. Clinical color vision test failures

AOHRR AOHRR Desaturated screening diagnostic D-15* D-15* plates plates

Glaucoma (n = 19, mean age 51.5 yr) 53% (10/19) 78% (14/18) 37% (7/19) 11% (2/19) Normals (age-matched to glaucoma subjects, n = 19, mean —(0/19) 11% (2/19) 5.3% (1/19) —(0/19) age 51.6 yr) Glaucoma suspects (n = 19, mean age 53.6 yr) 32% (6/19) 58% (11/19) 21% (4/19) — (0/19) Normals (age-matched to glaucoma suspects (n = 19, mean — (0/19) 11% (2/19) — (0/19) — (0/19) age 52.8 yr) Normals (n = 78, mean age 34.9 yr) 1.3% (1/78) 5.1% (4/78) Not done Not done

*D-15 and desaturated D-15 test failures are explained in Fig. 1, b. For the AO-HRR plates, missing any plate or part of a plate constitutes test failure.

within 4 years with a normal subject. Similarly, 19 not meet this "pass-fail" criterion, whereas nor- glaucoma suspects were each age matched to normals rarely made more than one single-place er- mal subjects. Glaucoma subjects were defined as ror. Consequently, for this study we adopted the having applanation of intraocular pressures (IOPs) pass-fail criterion as follows: test failure results greater than 22 mm Hg on two or more occa- when a subject makes more than one single-place sions, with diagnostic optic nerve head and/or vi- error or any error greater than single-place (Fig. 1, sual field defects. (Clinical details of each subject h). In this study we also used a desaturated version are shown in Table I.) Glaucoma suspects were of the test (each color cap was of the same hue and defined as having IOPs of greater than 22 mm Hg lightness but was less saturated by two steps in the on two or more occasions, without any diagnostic Munsell notation) and applied the same pass-fail optic nerve head or field changes and without hav- criterion; we anticipated that the desaturated ver- ing received any treatment for the raised IOP. sion would result in an increased error score. Both Normal subjects were recruited on the basis of tests were administered monocularly under a normal retinal and optic nerve head findings, IOP MacBeth Easel lamp. Test order was reversed for less than 22 mm Hg, no crystalline lens changes each successive subject to avoid favoring one ver- greater than grade two nuclear sclerosis, and cor- sion of the D-15 test. Each subject was also tested rected acuity better than 20/40 (glaucoma and with the AO-HRR plates13 under the MacBeth glaucoma-suspect patients were included on the Easel lamp. same lens, retinal, and acuity criteria). No aphakic Achromatic and chromatic sensitivity test. The or intraocular lens subjects were included in the detection of flicker increments of spectral light on study. In addition to the control subjects used to a white background is mediated by chromatic age match the glaucoma and glaucoma-suspect pathways for low flicker rates and by achromatic patients, normal subjects were also studied on the pathways for high alternation rates.9 In our exper- modified clinical Farnsworth D-15 panel color vi- iments a 2 deg, centrally fixated spectral test spot sion test. was viewed against a large (10 degree), 1270 tro- Tests. land, white (color temperature ~3200° K) back- Farnsworth dichotomous panel D-15 test. This ground. After adapting for 3 min to the white test, originally designed to identify observers with background, the observer decreased the radiance moderate and severe color vision loss,12 contains of the test spot until it no longer flickered. Three 15 color caps taken from the Munsell color circle such settings were averaged at each spectral wave- with approximately equal hue steps; each has the length tested. The spectral sensitivity measures same saturation (chroma) and lightness (value). allow an estimate of absolute sensitivity loss. In The subject is asked to "arrange the caps in order our experiments, the 1 Hz flicker was used at 460, according to color" and typically does this in less 500, 550, and 600 nm for chromatic sensitivity than 1 min. The standard scoring practice to iden- measures and the 25 Hz flicker was used at 460, tify "test failures" requires that two "major" errors 550, and 650 nm for achromatic sensitivity mea- be made, each resulting in colors across the color sures. All measures were made on a portable two- circle being placed side by side (Fig. 1, a). We had channel, Maxwellian-view optical system with the noted in pilot studies that patients with retinal single light source (tungsten halogen ophthalmo- disorders frequently made multiple errors that did scope bulb) filament focused to a 1 mm image in

Downloaded from iovs.arvojournals.org on 09/30/2021 Invest. Ophthalmol. Vis. Sci. 520 Adams et at. October 1982

CO LU CO CD O I -

400 500 600 700 400 500 600 700 WAVELENGTH (nm) Fig. 2. Spectral sensitivity of glaucoma subjects (n = 19, mean age 51.5 ± 15.8 years) and their age-matched normals (n = 19, mean age 51.6 ± 15.6 years). Solid lines (through the normal data), From a separate study of 11 young normals (mean age 24 years) under identical test conditions, except thresholds were determined at 20 nm intervals across the spectrum. The log difference between glaucoma and normal subjects is shown in the lower panels. A, Chromatic increment thresholds for a 2 deg circular target flashed at 1 Hz on a 1270 photopic troland white background and foveally fixated. Open symbols, Means forth e glaucoma subjects shown in Table I; filled symbols, means for the age-matched normals. Error bars indicate ±1 standard deviation for the normals (the standard deviation was always less than 0.3 log units at each wavelength). B, Achromatic increment thresholds for a 25 Hz test target. All other conditions are as described in A. Standard deviations never exceed 0.23 log units for the normals. At 460 nm, seven of the glaucoma subjects were unable to see 25 Hz flickeri n the test target at the maximum intensity available. Consequently, comparisons with normal values at this wavelength are less meaningful.

the plane of the subject's pupil. Intensity was con- glaucoma subjects and one of the glaucoma trolled by a 4 neutral density Inconel wedge; for suspects failed the test. However, Table II wavelength control the half-bandwidth of the in- shows that 10 of 19 glaucoma subjects failed terference wedge was approximately 12 nm. Field with the modified scoring method and six of stops at the focal point of the Maxwellian lens 19 glaucoma suspects also failed. No age- provided the test and background stimuli. matched normals failed with this procedure. The difference between glaucoma and Results normal subjects is highly significant (p < Farnsworth panel D-15 tests. When the 0.0001, Fisher exact probability test); the dif- conventional scoring method (see Methods) ference between glaucoma suspects and nor- was applied to the D-15 test, only four of the mals is significant at better than the 0.01

Downloaded from iovs.arvojournals.org on 09/30/2021 Volume 23 Number 4 Color vision changes in glaucoma 521

level by the same test. The desaturated ver- probability test, p = 0.37). Further, none of sion of the D-15 test caused 78% of the glau- the glaucoma suspects had miotic pupils and coma patients and 58% of the glaucoma sus- their color vision test results also differed pects to fail. By comparison, the age-matched significantly from those of normals. normal group had about 11% failures. Both Achromatic and chromatic sensitivity. To groups are significantly different from their what extent do the clinical results above have age-matched normals (glaucoma, p < 0.0001; a parallel in the sensitivity of achromatic and glaucoma suspects, p < 0.01, Fisher exact chromatic processing? One might anticipate probability test). In fact, better estimates of some loss of chromatic processing, but there the failure rate in the normal population for is little expectation of a loss in achromatic or both the D-15 and desaturated D-15 tests can luminosity sense. In fact, spectral luminosity be estimated from our larger study of normals is generally reported as normal for glaucoma (mean age 35 years), where we estimate ap- patients (see for example, Pokorny et al.,14 proximately 1% and 5% failures, respec- p. 292). Our results suggest a loss of both tively. We believe the failure rate in the achromatic and chromatic sensitivity for glau- normal population is likely to be less than this coma and glaucoma suspects compared with for the under-50 age group and greater than their age-matched normals. Fig. 2, A, shows this for those over 65. up to a four times (0.6 log unit) loss of One of the few clinically feasible alterna- chromatic sensitivity for the glaucoma sub- tives to the D-15 test in detecting retinal dis- jects, with the greatest loss in the short- ease is the AO-HRR plate test. From Table I wavelength (blue-violet) end of the spec- it can be seen that fewer glaucoma patients trum. When compared with that of the age- and glaucoma suspects fail the AO-HRR matched normals, a two to three times loss of screening test than the D-15 tests. Further- sensitivity (0.3 to 0.5 log units) was noted in more, most of those who fail go on to pass all achromatic sensitivity for the glaucoma sub- the diagnostic plates of this test, including jects (Fig. 2, B). The difference at all wave- all the glaucoma suspects. Although the AO- lengths is statistically significant (p = 0.0005 HRR screening test fails significantly more to 0.005 for chromatic and p < 0.01 for ach- glaucoma subjects than normals (p < 0.02, romatic, t test) when compared with that of Fisher exact probability test), the test does age-matched normals. not reveal significantly different performance Fig. 3 shows that glaucoma suspects are (at the 0.05 level) between glaucoma suspects also significantly different from their age- and normals. matched normals (p < 0.05 at all wavelengths The above results suggest that a rapid and except 550 nm for achromatic sensitivity, simple clinical test, the panel D-15, is capa- where p < 0.10 by t test). Here the magni- ble of failing more than half of glaucoma tude of the difference from normals for chro- patients while failing very few normals. How- matic and achromatic sensitivity is reduced ever, it is known that miosis can produce in- when compared with that in the glaucoma creased errors of color discrimination by re- subjects. ducing the effective retinal illuminance. This The data presented in Figs. 2 and 3 do not has been shown on the FM 100 Hue test5' 6 answer the question of whether the chro- and on the Pickford-Nicolson anomaloscope.7 matic and achromatic sensitivity losses occur What evidence is there that reduced pupil in the same subjects. In fact, those subjects size is responsible for the results here? From who showed reduction in achromatic activity Table I it can be seen that most of the glau- also showed reduced chromatic sensitivity. coma subjects had normal pupil size; only There are a number of ways of showing this. five of the 19 had pupil diameters of 2 mm or For example, for the glaucoma subjects the less. In addition there is no significant differ- chromatic sensitivity at 550 nm (C550) corre- ence in the failure rate on the D-15 tests be- lates well with the achromatic sensitivity at tween those with small pupils (Fisher exact the same wavelength (r = 0.71), and of the

Downloaded from iovs.arvojournals.org on 09/30/2021 Invest. Ophthalmol. Vis. Sci. 522 Adams et al. October 1982

400 500 600 700 400 500 600 700 WAVELENGTH (nm) Fig. 3. Spectral sensitivity of glaucoma-suspect subjects (n = 19, mean age 53.6 ± 14.8 years) and their age-matched normals (n = 19, mean age 52.8 ± 14.6 years). A, Chromatic increment thresholds for the same conditions as in Fig. 2, A. Open symbols, Means for the glaucoma suspects;^i//erf symbols, means for the age-matched normals. Error bars indicate ±1 standard deviation for the normals (standard deviation always less than 0.31 log units). B, Achromatic increment thresholds. Conditions as in Fig. 2, B. Standard deviations never exceed 0.23 log units for the normals. At 460 nm, five of the glaucoma suspects and two of the age-matched normals were unable to see 25 Hz flicker at the maximum intensity available. Consequently, comparisons at this wavelength are less meaningful. Solid lines as in Fig. 2. The log differences between glaucoma suspects and normal subjects are shown in the lower panels.

nine glaucoma patients with the lowest C550 ceivable that the subjects with a miotic pupil sensitivity, all of whom are at least 2 standard may find the beams from the test and back- deviations from the mean of the age-matched ground stimuli partially obstructed during normals, seven are also in the group of nine part of the test. In fact, we find no significant who have the lowest achromatic sensitivity at difference in the results for the five subjects the same wavelength (L550). A similar corre- who had miotic pupils when compared with lation is seen for the glaucoma suspects for the other glaucoma subjects. For the chro- the same functions (r = 0.89); of the nine matic sensitivity the difference was less than subjects with the lowest C550, six are also in 0.1 log units at all wavelengths and reached a the group of nine who have the lowest L550. maximum of only 0.14 log units at 550 nm for With the instrumentation used to measure the achromatic sensitivity. spectral sensitivity, pupil size should not have influenced the results, since all the light Discussion passes through a 1 mm beam focused at the Both glaucoma patients and glaucoma sus- subject's pupil. Although unlikely, it is con- pects, as groups, show significantly reduced

Downloaded from iovs.arvojournals.org on 09/30/2021 Volume 23 Number 4 Color vision changes in glaucoma 523

chromatic and achromatic sensitivity, dem- est predicted saturation discrimination (low- onstrating that both color and luminosity est ratio). The four with the highest predicted functions are altered. Inasmuch as chromatic saturation discrimination all passed the D-15 and achromatic activity are carried by two test and the four with the lowest predicted different populations of ganglion cells, each saturation discrimination all failed the D-15 with different conduction velocities and firing test (p < 0.04, Fisher exact probability test). patterns (tonic or phasic), the results suggest A similar analysis for the glaucoma suspects that our glaucoma and glaucoma-suspect revealed that the four with the highest pre- groups both have changes in integrity of dicted saturation discrimination all passed these pathways. It is generally accepted that the D-15 test and three of the four with the foveal function is not affected until late in the lowest saturation discrimination failed the glaucomatous process. Since these testing D-15 test. The Fisher exact probability sta- conditions isolate the central two degrees of tistic (p < 0.07) fails to reach normal levels the fovea, retinal changes in this region may of significance (p < 0.05); nevertheless, the be an earlier component of visual loss associ- trend is the same as for the glaucoma sub- ated with glaucoma than previously thought. jects. It is interesting to note that if we simply The high percentage of glaucoma-suspect look at the chromatic thresholds or the achro- failures is surprising. It could simply be that matic thresholds in isolation, neither of them the raised IOP alone results in these changes. alone shows a significant correlation with the It may also be that these changes are a very performance on the D-15 test. early sign of glaucoma; if this is the case, we Longitudinal studies are necessary to de- could have been dealing with a higher per- termine whether these kinds of color vision centage of patients destined to develop glau- tests identify those glaucoma suspects at high 15 16 coma than expected from other studies. ' risk for subsequent glaucoma. A less likely explanation is that the decreased Previous studies4 have shown that glau- spectral sensitivity is simply associated with coma patients and glaucoma suspects may the gene for abnormal aqueous dynamics. have reduced color discrimination. Unfortu- The task for the D-15 and desaturated nately, these color vision changes are most D-15 tests is essentially a saturation discrimi- commonly identified by tests that are difficult nation task. This is particularly obvious with to incorporate into clinical practice (about 20 the desaturated test, which has pastel-col- to 30 min is required to test and score the ored caps. In the visual system there is evi- FM 100 Hue test, and the Pickford-Nicolson dence for separate pathways for chromatic or Anomaloscope is expensive and complicated color signals and achromatic or brightness enough to place it out of the realm of routine (whiteness) signals. The difference between office procedure). The Farnsworth dichot- the achromatic and chromatic signals in the omous D-15 panel test generally requires less visual system should provide an estimate of than a minute of actual testing time for each saturation discrimination. Consequently, pa- eye, excluding instructions and scoring. By tients whose signal for chromaticity is consid- conventional scoring, four of our^glaucoma erably different than the signal for achroma- subjects would have failed the test; one of the ticity would be expected to have good sat- glaucoma suspects would have failed. Modi- uration discrimination. Thus if we take a ratio fied scoring, in our study, improved test sen- of the achromatic sensitivity to the chromatic sitivity without significant loss of specificity. sensitivity (i.e., subtract the log sensitivities), Ten of 19 glaucoma patients, all with good the ratio ought to predict the relative dis- acuity, and six of 19 glaucoma suspects failed crimination ability of the patient. We did this with the modified scoring procedure. The for the 550 nm spectral thresholds by iden- desaturated version of the D-15 test with the tifying the four glaucoma patients who had modified scoring was even more sensitive the highest predicted saturation discrimina- (failing 14 of 18 glaucoma patients and 11 of tion (highest ratio) and the four with the low- 19 glaucoma suspects). We are unable to ac-

Downloaded from iovs.arvojournals.org on 09/30/2021 Invest. Ophthalmol. Vis. Sci. 524 Adams et al. October 1982

count for these results on the basis of pupil 8. King-Smith PE: Visual detection analysed in terms size or age. of luminance and chromatic signals. Nature 255:69, 1975. 9. King-Smith PE, Kranda K, and Wood ICJ: An ac- REFERENCES quired color defect of the opponent color system. 1. Verriest G: Les deficiences acquises de la discrimi- INVEST OPHTHALMOL 15:584, 1976. nation chromatique. Mem Acad R Med Belg Series 10. Gouras P: Identification of cone mechanisms in 2 4:35, 1964. monkey ganglion cells. J Physiol (Lond) 199:533, 2. Fishman GA, Krill A, and Fishman M: Acquired 1968. color defects in patients with open-angle glaucoma 11. Adams AJ, Zisman F, and Cavender JC: Sensitivity and ocular hypertension. Mod Probl Ophthalmol loss in chromatic processing channels of diabetics. 13:335, 1974. INVEST OPHTHALMOL VIS SCI 19(ARVO Suppl):169, 3. Lakowski R, Bryett J, and Drance SM: A study of 1980. colour vision in ocular hypertensives. Can J Oph- 12. Farnsworth D: The Farnsworth-Munsell 100-Hue thalmol 7:86, 1972. and dichotomous tests for color vision. J Opt Soc Am 4. Lakowski R and Drance SM: Acquired dyschroma- 38:568, 1943. topsias: the earliest functional losses in glaucoma. 13. Hardy LH, Rand G, and Rittler MC: H-R-R polyc- Doc Ophthalmol Proc Ser 19:159, 1979. hromatic plates. J Opt Soc Am 44:509, 1954. 5. Grutzner P and Schleicher S: Acquired color vision 14. Pokorny J, Smith VC, Verriest G, and Pinckers defects in glaucoma patients. Mod Probl Ophthal- AJLG, editors: Congenital and Acquired Color Vi- mol 11:136, 1972. sion Defects, Current Ophthalmology Monographs. 6. Marmion VJ: The colour vision deficiency in open New York, 1979, Grune & Stratton, Inc. angle glaucoma. Mod Probl Ophthalmol 19:305, 15. Banks JLK, Perkins ES, Tsolakis S, and Wright JE: 1978. Bedford glaucoma survey. Br J Ophthalmol 1:791, 7. Lakowski R and Oliver K: Effect of pupil diameter 1968. on colour vision test performance. Mod Probl Oph- 16. Armaly MF: Ocular pressure and visual fields: a ten thalmol 13:307, 1974. year follow up study. Arch Ophthalmol 81:25, 1969.

Downloaded from iovs.arvojournals.org on 09/30/2021