POLARITY-SPECIFIC CONTRAST ADAPTATION
R. Beer and D. MacLeod University of California San Diego La Jolla, California

Purpose:How specific are the effects of contrast adaptation to regions of color space? Last year we showed that pre-exposure to contrast can asymmetrically compress the achromatic half-axes of color space (Beer&MacLeod, ARVO'98). This year, we investigated the transformations in brightness and color that result from pre-exposure to combined luminance and color contrast: 1) Are the red and green half-axes separately compressible? 2) Does compression along the red-green axis occur for only one luminance polarity at a time? 3) Does compression along the bright-dark axis occur in one chromatic half-plane at a time? Methods: 3 subjects were shown a pattern of short pinstripes drifting on a gray background in one part of their visual field (pre-exposure). Afterward, they adjusted the color and luminance of test stripes in this area to match the appearance of standard stripes presented in a non- pre-exposed area. Pre-exposure was to high contrast bright or dark red, or bright or dark green stripes, and caused a substantial reduction in contrast of test stripes in the pre-exposed region. The increase in physical contrast required for a perceptual match was taken to indicate the degree of perceived contrast loss. Results: As previously reported, pre-exposure to one achromatic polarity (e.g. dark stripes) caused greater loss of achromatic contrast for tests of the same polarity than for those of opposite polarity (bright stripes). Moreover, 1) Pre-exposure to one chromatic polarity caused a greater loss of chromatic contrast for tests of the same chromatic polarity: pre-exposure to red stripes on a gray background reduced contrast of red test stripes more than contrast of green ones. 2) Chromatic contrast reduction was greatest when pre-exposure and test stripes are of the same luminance polarity (e.g. bright red pre- exposing stripes reduced redness of bright red test stripes more than redness of dark red ones), but 3) the color-selectivity of luminance compression was relatively weak. Conclusions: These results indicate separately-adapting contrast mechanisms for the red and green, as well as bright and dark half-axes of color space. Chromatic mechanisms are specific to luminance polarity, and each has its own contrast gain control.