Abstract
Psychophysical differences between the detection of S-cone increments and decrements (S+ and S-) were studied using bipolar, dynamic noise masks of several contrast power levels, with a forced-choice method. Noise chromaticities were L, M, and S cone, as well as L-M, L+M, and achromatic (L+M+S). We previously reported (Wang, Guilianini, & Eskew, 2002) that the threshold Energy vs. Noise (EvN) functions differed dramatically for S+ and S- tests. With identical noises, the EvN's for S+ tests are much steeper than those for S- tests (except for achromatic noise) even though the noises consisted of two symmetric chromatic polarities of equal contrast power. The S+/S- differences suggest possible contrast gain differences in S-ON and S-OFF pathways. Here we model the entire set of data (for each of two observers) by combining test energies across all noise chromaticities and powers, separately for S+ and S-. A model in which cone signals are linearly combined to produce mechanism noise power (Giulianini & Eskew, 2007) accounts for the S+ and S- thresholds surprisingly well (r2≥0.89). The fitted relative cone contrast weights are nearly identical in three of the four cases (S+ and S- tests for both observers). In all four cases: (1) the L and M weights are of opposite sign; (2) the M-cone weight is largest in magnitude, larger than the S-cone weight; and (3) the S-cone weight is of the same sign as the L-cone weight. This pattern of relative weights is consistent with detection of these S+ and S- tests being mediated by a classical red-green hue mechanism, one with S cone inputs.
Meeting abstract presented at VSS 2014