Abstract
Color vision investigations often employ the concept of a color space based on a standard observer with standard cone fundamentals and cone ratios to calculate the relative activations of the cones and opponent mechanisms. However, since there are large variations in individuals’ cone fundamentals and luminosity function, each observer has their own unique color space. It would often be beneficial to account for these individual differences, particularly when attempting to study isolated cone or opponent mechanisms. While techniques for measuring individual cone fundamentals and establishing unique color spaces already exist, they are impractical and time-consuming for most applications. We have previously modeled how easily-measured luminance settings of an observer might be used to predict individual differences in color space due to their shared sources of variability. Here we report results of an empirical test of the ability of luminance differences to predict individual cone opponent axes. To estimate the tilt of the equiluminant plane we employed a minimum motion task to stimuli modulated along the assumed opponent axes. To locate the individual opponent axes, we used an established adaptation/contrast-matching paradigm (Webster et al, 2000). For both tasks we included measurements from both the fovea and at 4-deg in the periphery wherein the values should vary significantly within an observer due to macular pigment and other retinal inhomogeneities. As predicted by our previous model, we found a strong correlation (r = .72, p=.003) between measures of luminance evidenced as a tilt in the equiluminant plane and estimates of the location of the SvsLM opponent axis revealed as a rotation within the equiluminant plane. These results are consistent with our previous modeling in suggesting that standard and relatively accessible measures of equiluminance can be employed to also refine the estimate of the chromatic directions that isolate the cone-opponent mechanisms.