Color Vision Panel Tests: A Metric for Interpreting Numeric Analytic Indices

Abstract

Various systems have been proposed for quantifying arrangements of colored caps in panel tests of color vision. We examine Vingrys and King-Smith's (Vingrys and King-Smith. Invest Ophthalmol Vis Sci 1988;29:50-63) computed values for their different numerical indices derived and propose a metric for interpreting these index values. Our approach is based on performance of color-normals when angular subtenses of the panel caps are reduced by changing the test distance. Seven participants with normal color vision performed the panel tests at each of five viewing distances from 0.5 to 7.56 m (chromatic portions of the caps subtending 75 to 5 minarc). To obtain cutoff values for the numeric indices at the standard viewing distance under our viewing conditions, we also tested 69 unselected, volunteers with the Farnsworth D-15 and Lanthony desaturated D-15 panels. From each participant one eye, usually the worse, was included in analyses to find the 90th percentiles for each index, which we used as cutoff values. Many of the participants also made Rayleigh anomaloscope matches. For normal observers, values for the C-index (confusion) and for the S-index (polarity of an individual's pattern of cap reversals) begin to worsen when viewing distance increases beyond 2 m (corresponding to a subtense of 15 minarc). Thereafter, increasing values of the indices (deteriorations in color vision) fall exponentially with distance. We find that the functions for the Farnsworth and Lanthony panels are so highly correlated that the same function can be used for both. The cutoff index values for our conditions, values that differ little from those reported by others, correspond to a viewing distance of about 2.5 to 3.0 m. An individual's color vision performance can be interpreted by relating it to performance of color-normals viewing the test caps at some non-standard distance. This is similar to Snellen notation for acuity.