Color-normal Observers Research Articles

Pigment tests evaluated by a model of chromatic discrimination.

Clinical color-vision tests are evaluated within the framework of a model of chromatic discrimination in terms of cone excitation. The motivation for this study was to derive a method for evaluation of test design, test sensitivity, and observer performance. The discrimination model is based on the assumption that chromatic discrimination is mediated in two independent channels, one for short-wavelength cones and one for long- and middle-wavelength cones. Luminance-dependent templates are derived for each channel, and they describe chromatic-discrimination behavior of the young color-normal observer. The templates incorporate receptor- and opponent-level gain controls. We show how the chromaticities of clinical tests can be calculated in cone-excitation units and how discrimination behavior on the tests can be plotted on the templates. The tests include the Farnsworth-Munsell 100-hue, the Farnsworth Panel D-15, the Farnsworth Panel D-15 desaturated, the American Optical Hardy-Rand-Rittler, the Farnsworth F2 plate, the Standard Pseudoisochromatic Plates, Part II, the Ishihara, and the Minimalist tests. Clinical-test data collected on young color-normal observers at different illumination levels show the validity of the techniques.

Red-green and yellow-blue opponent-color responses as a function of retinal eccentricity

Using the hue cancelation technique, I measured the spectral sensitivities of red-green ( r-g) and yellow-blue ( y-b) opponent-color responses of two color-normal observers at 0 (fovea), 3.3, 10, 21, and 31° along the horizontal meridian of the temporal retina. The stimulus size as a function of retinal eccentricity was varied according to the cortical magnification factor. I found that the net spectral sensitivity (i.e. corrected for macular pigment) of the r-g channel was approximately constant as a function of retinal eccentricity. On the other hand, the net spectral sensitivity of the y-b channel was not constant. For the y-b channel I found a constant spectral sensitivity as a function of eccentricity when no macular correction was made. This means that the y-b system changes its yellow and blue responses with eccentricity to counterbalance the effects of macular pigment.

Sawtooth contrast sensitivity: Effects of mean illuminance and low temporal frequencies

Temporal contrast sensitivity was measured for mirror-image sawtooth (rapid-on and rapid-off) and sine waveforms for a 1.8 deg foveal target. In one experiment, contrast sensitivity was measured for 2–26 Hz stimuli at target mean illuminance levels of 5–1260 td. At 5 td, contrast sensitivity functions for sawtooth and sine waveforms, expressed in terms of the Fourier fundamental amplitude, are equivalent. At higher light levels, sawtooth sensitivity increasingly exceeds sine sensitivity and rapid-off (decremental) sawtooths show progressively greater sensitivity than rapid-on (incremental) sawtooths. This pattern of results was obtained for two color-normal observers and for a deuteranopic observer. In a second experiment, sawtooth and sine sensitivity was tested at 500 td with an extended low-frequency range, to 0.5 Hz. Rapid-off and rapid-on sensitivities declined only slightly at low temporal frequencies in contrast with sine sensitivity. To interpret our data, we evaluate two single-pathway models (last-stage asymmetric detector and compressive response-intensity nonlinearity) and a dual-pathway model in which incremental and decremental waveforms are detected by separate ON and OFF visual mechanisms.

A clinical test to estimate chromatic thresholdsa

A quick and relatively inexpensive colour vision test is described which estimates chromatic thresholds for representative hues of the Munsell colour space. Results suggest that the test is not as sensitive as several pseudo-isochromatic plate tests for colour vision screening. However, the test results correlate well with the anomaloscope range which indicates that the test is useful in estimating the severity of a red-green defect. Results from the screening trial also showed that there is a substantial learning or attention artefact for colour-normal observers. Approximately 36 per cent of the colour-normal subjects made at least one mistake on the test, whereas only nine per cent repeated at least one of their errors. Because these errors made by colour-normals were age-related and were clustered near the blue-yellow axis, detection of subtle blue-yellow defects in older patients will be difficult.

Foveal cone detection statistics in color-normals and dichromats

We measured for six male observers, the psychometric functions for the detection of two simultaneously presented points of light. The test stimuli were two 1 min point sources separated by 17 min arc, and pulsed for 0.5 msec. The stimuli varied in wavelength from 500 to 620 nm. The psychometric functions were fit with a model that assumes ideal detection and the following properties: (1) Poisson-distributed quantal absorptions; (2) binomial sampling of foveal long-wavelength-sensitive (LWS) and middle-wavelength-sensitive (MWS) cones; (3) independent responses of the LWS and MWS cones; and (4) the Smith-Pokorny fundamentals for cone spectral sensitivies. Based on χ 2 fits to the psychometric functions for detecting neither, one or both of the two-point stimuli presented, estimates were derived for the minimum quantal catch by a single cone for detection ( C), the number of effective cones illuminated by a point stimulus at threshold ( N) and the proportion of central foveal cones of the LWS type ( P L ). Three observers were color-normal, two were protanopes and one was a deuteranope. A second deuteranope was included in the design but his data were too unreliable for an unambigous solution. The estimated quantal requirement C was consistently near 5 (4–6), and the effective number of illuminated cones always was 1 or 2. The plausible range of P L (98% confidence interval) for the color-normal observers was 0.48–0.68 (observer YY), 0.76–0.90 (observer MW) and 0.77–0.95 (observer DF). The best fitting P L solution for these observers were 0.61, 0.82 and 0.88, respectively. These were comparable to the values obtained from flicker photometric data. The best P L value for each of the protanopes was 0.00 and for the deuteranope the best P L value was 0.98.

A simple behavioral demonstration of blue-cone anisotropy: Distance-induced tritanopia on standard color vision tests

In the course of a larger research program examining popular color-vision tests, the very powerful role of varying viewing distance was demonstrated with color-normal observers. More importantly, the effect of increasing distance was shown to be particularly evident for items designed for de­ tection of blue-yellow dichromacy. This is attributed to distance-induced tritanopia, which in turn reflects the unusual distribution of the blue cones in the retina. The utility of this finding for the simple demonstration of a nonintuitive aspect of retinal anatomy is discussed.

Effect of Short Wavelength Absorbing Filters on Farnsworth-Munsell 100 Hue Test and Hue Identification Task Performance

The influence of four short wavelength absorbing filters (yellow filters) on performance of the Farnsworth-Munsell 100-Hue test and three hue identification tasks was assessed in 10 color-normal observers. The yellow filters included the NoIR 40% Amber, Vuarnet No. 4006, Corning CPF 550, and Blu Blocker. A neutral density filter and a no-filter condition served as comparisons. Results showed that the Vaurnet, CPF 550, and Blu Blockers significantly lowered performance on the Farnsworth-Munsell 100-Hue test and the color-dependent tasks below levels obtained in either the neutral-density filter or no-filter conditions.

The role of small‐field tritanopia in two measures of colour vision

The present work extends the findings of previous efforts examining the comparability of current colour-screening tests. Several popular tests are shown to differ greatly in the performance exhibited by colour-normal observers as well as in their differential sensitivity to experimental manipulations of viewing duration and viewing distance. Those tests designed to identify yellow-blue dichromacy are especially sensitive to the manipulation of viewing distance, which is interpreted as reflecting 'small-field tritanopia' and the asymmetry in retinal density of the three cone types. These findings are discussed in terms of factors that influence the comparability of current colour-screening devices and the particular need for close adherence to standardized conditions with such instruments.

Temporal induction of blackness—I. Color appearance

Three color-normal observers described the appearance of colors under conditions of temporal induction. One of four inducing fields (unique blue, unique green, unique yellow, unique white) was foveally viewed for 5 sec followed immediately by a 400 msec reference stimulus of the same size (0.75 deg) and spatial location as the inducing field. The reference stimulus was 3 or 5 td and appeared achromatic to the observer when viewed without the inducing field. After presentation of this temporal sequence, the observers described the reference stimulus by assigning percentages to the terms red, green, yellow, blue, white and black. The range of inducing field intensities was −0.4−3.2 log td. Both chromatic and achromatic induction occurred at low illuminance levels of the inducing fields. As the illuminance of the inducing field increased, the reference stimulus gradually became blacker and the chromatic components less prominent. The minimum illuminance level at which the reference appeared 100% black was the same for the four different inducing fields.

The relative numbers of long-wavelength-sensitive to middle-wavelength-sensitive cones in the human fovea centralis

The determination of the relative numbers of different cone types in the human retina is fundamental to our understanding of visual sensitivity and color vision; yet direct measurements which provide this basic information have not previously been made for all cone types. Here we present a model which links the detection of a test light of small dimension to the number of cones contributing to detection of the light. We selectively isolated either the long-wavelength-sensitive ( L) or the middle-wavelength-sensitive ( M) cones, by choosing combinations of wavelengths of adapting backgrounds and tests to favor detection by the cone class of interest. Our model was applied to the detection functions measured for six color normal observers to obtain estimates of the relative numbers of L to M cones. Our estimates ranged between 1.46 and 2.36 for our observers with a mean value near two L cones for every M cone in human fovea centralis.

Perception of electronic display colours as a function of retinal illuminance

Colour-naming data were objected from four colour-normal observers to determine the nature of the changes in colour appearance that can occur with changes in illuminance for spectral and nonspectral stimuli comparable to those used in electronic displays. Test stimuli were presented as one degree, foveal flashes of one second duration in Maxwellian view. The retinal illuminance of the nine test stimuli varied from −1.0 to 4.0 log trolands (0.01-1 147fL). Shifts in perceived colour were observed with changes in stimulus intensity. At low illuminance levels, the stimuli appeared desaturated, with black and white responses dominating chromatic responses. At higher retinal illuminances, the stimuli were more saturated, although perceived colour was not constant with changing light levels. In general, blue and yellow responses increased at a faster rate relative to red and green responses with increasing illuminance. The data have important implications for the use of colour on electronic displays.

DISTANCE, DURATION AND BLUR EFFECTS ON THE PERCEPTION OF PSEUDOISOCHROMATIC STIMULI

Two experiments are reported that examine the effects of viewing distance, viewing duration, and stimulus clarity on performance with three popular versions of the pseudoisochromatic tests of colour deficiency. With a large sample of colour-normal observers, highly significant effects of these viewing conditions are found. In addition, large differences in sensitivity to these conditions are obtained among the three tests employed (The American Optical Corporation test, the Dvorine test, and the Ishihara test). Results are discussed in terms of practical implications for test administration and in terms of likely processes, in addition to colour discrimination, assessed by these tests.

Distance, duration and blur effects on the perception of pseudoisochromatic stimuli

Abstract— Two experiments are reported that examine the effects of viewing distance, viewing duration, and stimulus clarity on performance with three popular versions of the pseudoisochromatic tests of colour deficiency. With a large sample of colour-normal observers, highly significant effects of these viewing conditions are found. In addition, large differences in sensitivity to these conditions are obtained among the three tests employed (The American Optical Corporation test, the Dvorine test, and the Ishihara test). Results are discussed in terms of practical implications for test administration and in terms of likely processes, in addition to colour discrimination, assessed by these tests.

Further investigation of viewing conditions on standard pseudoisochromatic tests

Two experiments are reported that examine the effects of variations in viewing duration, viewing distance, and illumination level on performance with three popular pseudoisochromatic tests of color blindness. Highly significant effects of these conditions are obtained with a sample of color-normal observers, but the three tests also differ markedly in their susceptibility to variations in these viewing conditions. Results are discussed both in terms of practical implications of test administration under nonstandardized conditions and in terms of other likely visual processes assessed by these tests.

Temporal sensitivities related to color theory.

Sensitivities of color-normal observers to temporal variations in stimulus luminance and chromaticity were measured for sine-wave stimuli between 1.5 and 20 Hz. Clear differences were found in observers' sensitivities to isochromatic luminance variations and to isoluminous chromaticity variations for wavelength pairs selected to test temporal discriminability along the red-green and yellow-blue dimensions, respectively. Despite interobserver differences in individual red-green functions, a given observer's sensitivity could be described by a single curve shape specific to that observer. Overall sensitivity for yellow-blue was less than that for red-green for all observers. Differences in curve shape between red-green and yellow-blue functions are found for individual observers, but group averages reveal that the differences are not systematic. Red-green temporal sensitivity is largely unaffected by adapting backgrounds in red-green equilibrium but is attenuated at low frequencies by nonequilibrium backgrounds of the same luminance. Isochromatic luminance sensitivity is largely independent of our adapting backgrounds, but heterochromatic luminance modulation functions undergo expected changes in form.

Densitometric measurement of human cone photopigment kinetics.

We measured the radiance response function for steady state bleaching lights and the regeneration of the cone visual photopigments using the continuous recording densitometer described by v. Norren and v.d. Kraats. Measurements made on 5 deuteranopes, 1 protanope and 2 color-normal observers were similar. The radiance response function was steeper than the function predicted by a simple first-order kinetic equation. For a measured density (ca 0.3) we evaluated whether high stray light (ca 47.5%) and high two-way optical density (ca 1.3) could account for the deviation from the prediction of a first-order equation. Such a model was rejected because these parameters predicted different estimates of the time course of regeneration for different test wavelengths (554 and 605 nm). Statistical analysis of the regeneration data revealed a highly significant non-linearity. A model in which the rate of regeneration increases as the proportion of bleached photopigment decreases is required to explain both the radiance function and the regeneration data.

Color sensation of normal and anomalous trichromats measured by magnitude estimation

Colors perceived by normal, protanomalous, and deuteranomalous trichromats were scaled by magnitude estimation. As a result, opponent-colors functions were obtained which show the hue sensations produced by spectral lights. For color-normal observers, the loci of maximal hue sensations were different in some aspects from those expected from current opponent-colors theories.

Chromatic opponent-response functions of anomalous trichromats

Chromatic valence functions for the red-green and blue-yellow opponent-response systems of normal and anomalous trichromats were measured psychophysically using the Jameson and Hurvich hue cancellation technique. Four protanomalous and two deuteranomalous observers with varying degrees of color discrimination loss, and two color-normal observers, served in the experiment. The color-normal data were in good agreement with the results reported by Jameson and Hurvich. Protanomalous and deuteranomalous observers differed with respect to the relative strengths of the two chromatic opponent systems: decreased red-green strength was correlated with poorer color discrimination. Theoretical fits to the opponent-response data were calculated by assuming linear combinations of cone inputs for both chromatic systems and different sets of photopigments for the three classes of observer.

Multidimensional scaling of subjective colors by color-blind observers

Temporal coding theories of color vision suggest explanations of flicker-induced subjective colors such as those that appear on Benham’s disk. If color blindness were due simply to photopigment anomalies, then subjective colors might be elicited by central patterns of neural activity in color-blind observers that mimic those which the cones normally produce in colornormal observers. We had color-normal and color-blind observers scale subjective colors like those on Benham’s disk for similarity. The inferred color spaces for six normal observers resembled the familiar hue circle, but the spaces for five red-green-deficient observers were compressed along the red-green axis. This is consistent with the position that flicker colors are due to retinal processes, and suggests that color blindness may involve variations of the central nervous system in addition to photopigment anomalies.

Variables of color, duration, frequency, presentation order, and sex in the estimation of dot frequency.

The estimation of number depends on the information as dot frequency, exposure conditions, order of presentation, and sensitivity of the observer to color. Ninety-six color-normal observers (age 17-27) equally divided by sex were assigned so that each one viewed 42 cards, each of which contained dots of one of three colors on one of two grounds at one of seven frequency levels. For any observer, cards were exposed for either .5 second or 1.0 second intervals and in either ascending or descending order. The ANOVA indicated a significant effect for the following variables: dot frequency, dot color, and exposure time. Estimates increased with dot number on the several displays. Yellow-red (YR) dots were slightly overestimated. Yellow-green (YG) and gray dots were not. The briefer exposure yielded an underestimation and the longer an overestimation. Interpolated intervals are given for correct estimates.