Selective noise masking of L and M cone stimuli: unipolar tests reveal theoretically significant asymmetries.

In an important study, Hansen & Gegenfurtner (2013) showed that chromatic masking noise placed near the corners of detection contours in the ΔL/L, ΔM/M plane of cone contrast color space can produce selective masking. Rather than being indicative of the existence of large numbers of mechanisms, as they suggested, we (Eskew & Shepard, 2013) demonstrated that this result could be accounted for by a model with a very limited number of mechanisms. Both studies used bipolar test stimuli that contain equal and opposite chromatic contrasts. The symmetry of such stimuli enforces symmetry on the threshold contour and the model fit that can obscure the presence of asymmetric mechanisms-ones that do not form pairs with equal and opposite cone weights. Unipolar mechanisms like this can act as if there were more mechanisms than actually exist, since different mechanisms can detect the two poles of bipolar stimuli independently (Eskew, 2009). Here, for the first time, we report that detection contours for unipolar stimuli in the LM plane may be asymmetric when masking noise is added near the corners of the detection contour. Thresholds for stimuli comprised of nearly equal L and M increments are, for almost all observers, higher than the thresholds for the corresponding stimuli comprised of nearly equal L and M decrements. There are also differences in the cone contrast weights for the R and G mechanisms (responsible for the long flanks of the detection contour), differences that predict selective masking, as we previously reported, without requiring large numbers of mechanisms. A plausible physiological reason for these asymmetries will be discussed. A very limited number (4 to 6) of unipolar, asymmetric mechanisms may be able to account for all the bipolar and unipolar test detection contours. Meeting abstract presented at VSS 2015.