Abstract
Purpose: :
Protanomaly and deuteranomaly describe colour vision deficiencies which rely on opsin gene arrays with the first two genes encoding either M- or L-like pigments that differ in peak spectral responsivity (Vis. Res. (2011), 51, 633-651). The differences between M- and L-class variant pigments arise largely because amino acid substitutions in M-class pigments contribute less to the corresponding changes in spectral responsivity. For each distinct spectral separation, other factors such as the relative numbers of L and M cones, their optical density and the midpoint between their spectral peaks can also contribute to the subject’s overall chromatic sensitivity. The purpose of this study was to examine whether the rank order of chromatic sensitivity loss measured in deutan- and protan-like deficiency can be predicted using samples from discrete Gaussian distributions selected to reflect the most common, wavelength separations within L- and M-class of pigments.
Methods: :
Red/green (RG) colour thresholds were measured in 269 deutans, 132 protans and 330 normal trichromats using the colour assessment and diagnosis (CAD test). The colour vision of every subject was also examined using the Nagel anomaloscope. Classification into normal, deutan and protan classes was based on the results obtained on CAD and anomaloscope tests.
Results: :
The RG CAD thresholds measured within each subject group were ranked in increasing order. A number of random samples equal to the number of subjects within each group was taken from a single Gaussian distribution, in the case of normal trichromats, or from several distributions in the case of deutan and protan groups. A program was written to optimise the mean values, the number of samples and the variances for the minimum number of distributions needed to predict the rank order of the measured RG thresholds in the deutan and protan groups.
Conclusions: :
The rank order of thresholds measured in normal trichromats can be predicted by a single Gaussian distribution with parameters computed from the measured data. Deutan subjects produced the most complex rank order which could only be predicted adequately with four or more Gaussian distributions. In contrast, the rank order for the protan group was much simpler and could be predicted well with only two or at most three Gaussian distributions.
Keywords: color pigments and opsins • color vision • computational modeling