Pre-adaptation to isoluminant pattern generates sensitivity losses selective for chromatic polarity. R.D. Beer , D.I.A. MacLeod , E.M. Tan. Psychology, U.C. San Diego, La Jolla, CA.

Purpose: Pattern-specific losses in perceived contrast can be to some degree selective for luminance contrast polarity (ARVO 1998), and this remains true even when the preadapting and test patterns have chromatic as well as luminance contrast (ARVO 1999). Here we ask whether similar polarity selectivity occurs for isoluminant patterns, and whether the color shifts that test patterns undergo as a result of preadaptation to chromatic patterns are consistent with simple models of chromatic processing in which color is represented by the excitation of mechanisms encoding just two axes of color space. Methods: Three observers pre-adapated to a pattern of drifting, high-contrast dots on a gray background. They were then asked to match, by means of a trackball that allowed free movement in the isoluminant plane, the color appearance of similar dots on the pre-adapted side of their visual field to the appearance of dots on the other side. The dots were unidirectionally offset in color, but not in luminance, from the gray background and the difference from the background was either along the principal chromatic axes, or along intermediate directions. Results: 1)Adaptation caused the greatest change in appearance of test patterns in the same half-plane as the adapting dots. 2)The pattern of color shifts was different with principal-axis adaptation vs. off-axis adaptation. Shifts tended to be parallel to the adapting direction for principal-axis adaptation, but showed more complex behavior for off-axis adaptation. Conclusions: Color space in these experiments is not isotropic. Some aspects of the results are consistent with a simple model where adaptation effects occur more or less independently for four primary half-axes, but the results for adapting patterns that lie off the primary axes are more complex than predicted by such a model.
CR: None      Support: NIH grant EY01711