Tsaiyao Yeh

Chromatic Discrimination with Variation in Chromaticity and Luminance: Data and Theory

Boynton and Kambe developed a model of chromatic discrimination in which thresholds are mediated by two independent mechanisms: the short-wavelength sensitive (S-) cones (S-cone axis), and the middle-wavelength sensitive (M-) and long-wavelength sensitive (L-) cones (M/L-cone axis). In this study, we used a Maxwellian view optical system to investigate fundamental properties of the model as a function of chromaticity and luminance. We confirmed that discriminations along the S-cone axis were dependent on S-cone excitation level. However, changes in chromaticity and changes in mean luminance were not described by a single threshold-vs-radiance (TVR) template. We developed a model to account for the different effects of changing S-cone excitation by varying mean chromaticity and by varying mean luminance. M/L-cone discriminations showed a minimum at the L-cone excitation to white, indicating strong opponency. The thresholds increased with luminance approaching a Weber region and showing parallel functions for differing chromaticities. These data are fit by a model allowing retinal gain controls and spectral opponency.

Mechanisms subserving temporal modulation sensitivity in silent-cone substitution

Temporal contrast sensitivity data were collected with sine-wave-modulated lights for achromatic, chromatic, and silent-cone-substitution stimuli. Achromatic (556- and 642-nm lights in phase) and chromatic (556- and 642-nm lights in counterphase) modulation sensitivities were measured at a constant time-average retinal illuminance of 1256 trolands (Td) and chromaticity of 595 nm. These data were considered to represent isolated temporal responses of luminance and red-green chromatic channels, respectively. Silent cone substitution was achieved with counterphase modulation of the 556- and the 642-nm lights and by suitable adjustment of the modulations or the radiances of the two lights. (1) The peak modulation depth of the 642-nm light was reduced to silence the long-wavelength-sensitive (LWS) cone, and the peak modulation depth of the 556-nm light was reduced to silence the middle-wavelength-sensitive (MWS) cone. These protocols maintained the time-average retinal illuminance and chromaticity as for the control conditions. (2) The luminance of the 642-nm light was decreased to silence the LWS cone and was increased to silence the MWS cone. In this procedure the time-average retinal illuminance and chromaticity differ for the silenced-LWS-cone (1047 Td and 589.5 nm) and the silenced-MWS-cone (4358 Td and 622 nm) conditions. The response modulation of the achromatic and the chromatic channels was calculated for the silent-substitution conditions. The chromatic channel is more sensitive at low frequencies, with a transition to greater achromatic channel sensitivity near 13 Hz for the silenced-LWS-cone condition and near 6 Hz for the silenced-MWS-cone condition.(ABSTRACT TRUNCATED AT 250 WORDS)

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.

Colorimetric purity discrimination: Data and theory

Colorimetric purity, measured as the first step from white toward the spectrum has a V-shaped function. Purity discrimination is best near 400 nm, least at 570 nm and intermediate at mid-spectrum and long wavelengths. A much flatter function occurs when colorimetric purity is measured as the first step from the spectrum toward white. In this study, we applied the formulation of chromatic discrimination thresholds measured along the S-cone and M/L-cone axis to account for chromatic discrimination in the equiluminant plane. The modeling results show that the purity step from white has a 1.6 log unit calculated range, similar to the classical data. The purity step from the spectrum is much flatter. The predicted range is dependent on the individual variance in chromatic discrimination thresholds and the luminance level. We then used psychophysical procedures to test the models predictions. The resulting purity discrimination functions were generally in agreement with the model. Our modeling indicates that discrepant data of colorimetric purity can be explained with the context of discrimination models.

S-cone discrimination sensitivity and performance on arrangement tests

Previous psychophysical data from our laboratory have revealed that tritan discriminations are characterized by a set of luminance-dependent threshold-vs-radiance templates. The templates were expressed in units of S-cone trolands. In this communication, we used the templates to evaluate performance on four arrangement tests: the Farnsworth-Munsell 100-hue test, the Farnsworth Panel D-15 test, the Lanthony desaturated Panel D-15 test and the Lanthony New Color Test. We calculated S-cone excitation in S cone trolands for the test caps and derived a value of AS for given cap errors at four illuminance levels. Our calculations showed that tritan errors were expected at reduced illuminations for all tests. The level of illumination at which tritan errors were predicted was consistent with published data.

The Farnsworth-Munsell 100-hue test in cone excitation space

The Farnsworth-Munsell 100-hue test (FM 100-hue) has been subject to a variety of sophisticated analyses in recent years including conversion to uniform color space and Fourier analysis of error patterns. The majority of acquired color defects seen by clinicians, often assessed by the FM 100-hue or by arrangement tests, arise from retinal or optic nerve disorders. Chromatic discrimination at the retinal level is well described within cone excitation space. Retinal and optic nerve disorders should show chromatic discrimination loss along ‘critical axes’ in cone excitation space. We therefore calculated cap locations in the MacLeod-Boynton cone excitation space, where S/(L + M) cone excitation is plotted vs L/(L + M) cone excitation in a constant luminance plane. The FM 100-hue cap locations form an ellipse tilted from the S and L/(L + M) axes. Excitation on one or the other axis gives an approximately sinusoidal error pattern as a function of cap number, but the separation is not 90°. Error patterns for congenital and acquired defects can be predicted and compared with data. Normal FM 100-hue test error scores can be plotted on a Δl/l template. A similar approach can be taken for arrangement tests and confusion errors can be plotted on the template. The results are consistent with chromatic discrimination data as a function of test illuminance.