Variability of chromatic sensitivity : fundamental studies and clinical applications
This investigation involved a number of related studies with the principal aim of assessing the variability in chromatic sensitivity in "normal" trichromats and colour deficient observers. An important outcome was the development of a new method for accurate and efficient measurement of chromatic sensitivity and the establishment of reliable statistical limits that describe the distribution of redgreen(RG) and yellow-blue (YB) chromatic sensitivity in normal trichromats. Chromatic sensitivity was assessed using a computer-based psychophysical procedure that employs spatiotemporal luminance contrast (LC) masking techniques to isolate the use of colour signals. The colour-defined stimuli were buried in dynamic LC noise and moved diagonally across a square. The subject's task is to indicate one of the four possible directions of stimulus motion by pressing the corresponding response button. The Colour Assessment and Diagnosis (CAD) test was optimised for stimulus size and background luminance level to provide an efficient measure RG and YB chromatic sensitivity. Colour detection thresholds are assessed along 16 directions in chromaticity space, selected to yield maximum information on RG and YB chromatic sensitivity loss and to distinguish between deutan and protan deficiencies. The CAD test was used to measure chromatic discrimination thresholds in 472 observers. The results made it possible to evaluate the screening and diagnostic efficiency of the CAD test in comparison with a number of established clinical colour vision tests, e.g., Nagel anomaloscope and Ishihara pseudoisochromatic plates. The specificity and sensitivity of the CAD test versus the Nagel anomaloscope (223 subjects) was found to be 100% and 97.5%, respectively. The diagnostic agreement with respect to the Nagel anomaloscope was 0.99; only two subjects showed inconsistent classification deficiency. Chromatic thresholds measured in normal trichromats were also examined and the variability amongst normal colour vision was investigated. The data obtained were used to establish the statistical limits for the standard normal observer on the CAD test. A template based on these limits was generated and provides an efficient way of separating accurately colour deficient from normal trichromats. The high specificity found is consistent with the correct detection of minimal colour deficiencies that sometimes go undetected in other colour vision tests. Molecular genetic analysis was also carried out in a small group of subjects with unusual colour deficiencies in an attempt to understand the genotype-phenotype relationship between predicted and observed colour vision losses. A comparison of genetic results from DNA sequencing and chromatic detection thresholds measured on the CAD test suggests that the separation of 30 nm between peak sensitivity of L and M normal photoreceptor pigments is not an absolute requirement for an observer to exhibit normal colour vision as assessed with conventional colour vision tests. The macular pigment (MP) optical density was also measured in 23 subjects selected for their higher than normal levels of pigment as a result of their participation in a carotenoid supplementation trial. This pre-receptoral filter absorbs light preferentially in the short wavelength region of the visible spectrum and has been shown to affect colour matches. Contrary to expectations, the findings of this investigation showed no correlation between YB thresholds and MP optical density levels. A model that explains this finding was produced confirming the observed experimental findings. In addition, the model also predicts a small improvement in RG chromatic discrimination sensitivity with increased MP optical density in the eye.