Abstract
BACKGROUND & PURPOSE Chromatic backgrounds alter color appearance. Background patterns that isolate S cones have been shown to cause particularly strong color shifts, which are larger than the shift from a uniform background at any chromaticity within the pattern (Shevell & Monnier, 2001). The S-cone patterns used in previous work were represented at both the receptoral and postreceptoral levels. The question addressed here is whether the color shifts induced by S-cone-isolating chromatic patterns reflect activity at the receptoral level (S-cone signals) or in a postreceptoral pathway (S/(L+M)). The results establish that patterns represented by postreceptoral signals, not S-cone excitation, mediate these large color shifts. METHODS Color appearance of a test field was measured using several background patterns composed of concentric circles alternating between two chromaticities. Pattern 1 had both S-cone and postreceptoral S/(L+M) variation, as in previous work. Pattern 2 had postreceptoral variation (as in Pattern 1) but no S-cone variation. Pattern 3 had S-cone variation (as in Pattern 1) but no postreceptoral S/(L+M) variation. The properties of Patterns 2 and 3 were achieved by adjusting the luminances of the two chromaticities. Color shifts induced by these patterns were measured using asymmetric matching. RESULTS Pattern 1 produced large shifts in color appearance, corroborating previous results. Similar shifts were produced by Pattern 2 (postreceptoral variation only) but not by Pattern 3 (S-cone variation only). CONCLUSION The large shifts in color appearance induced by S-cone patterns are mediated by signals in a postreceptoral S pathway. These results are consistent with a cortical neural mechanism with +s/−s spatial antagonism, as found in V1 (Conway, 2001) and which accounts for previously reported perceptual color shifts.