Abstract
BACKGROUND & PURPOSE: Background patterns composed of concentric rings alternating in s=S/(L+M) chromaticity can induce larger shifts in appearance than a uniform background at either chromaticity (Monnier & Shevell, OSA/UCI 2001). These appearance shifts are along the s direction. We now consider l=L/(L+M) variation within similar backgrounds to test whether l and s variations in the background independently shift appearance in the l and s directions, respectively.
METHODS: Using an asymmetric matching paradigm, observers set the hue, saturation and brightness of a comparison ring within a uniform white-appearing surround to match the appearance of a test ring presented within a chromatic pattern. Stimuli were presented on a calibrated CRT. RESULTS Uniform backgrounds produced weak shifts, usually consistent with contrast. On the other hand, patterned backgrounds composed of rings alternating in only s chromaticity (l-constant/s-contrast patterns) caused large appearance shifts in the s direction, corroborating previous work. Adding l variation to these s varying patterns (l-contrast/s-contrast patterns), while holding constant local contrast at the edge of the test, shifted the matches along the l direction. Furthermore, the added l contrast altered the magnitude of the appearance shifts in the s direction. The direction of the s-shifts (i.e., enhancement or reduction) depended systematically on the l chromaticity of the test.
CONCLUSIONS: The appearance of the test within the l-constant/s-contrast patterns can be understood by independent induction within l and s mechanisms. Appearance of the test within the l-contrast/s-contrast patterns, however, shows an interaction between the l and s mechanisms. These results are consistent with a model in which the gain of the s system is regulated by l contrast within the background.