Abstract
Purpose. We have demonstrated hue-selective mechanisms for encoding luminance information at suprathreshold contrast levels (Hardy & De Valois, ARVO 1998). The purpose of these experiments is to investigate the hue-selectivity of luminance encoding at the contrast detection threshold. Methods. Subjects adapted to stimuli of different colors alternating at 1Hz. In one frame of the adapting stimulus, gabor patches were presented to the left and right of fixation (s = .87deg; center-to-center separation = 5.2deg). The phases of the gabor patches were shifted psuedo-randomly at 6Hz. In the other frame of the adapting stimulus, a blank field of the complementary color was presented. Initial adaptation lasted 150sec and top-up adaptation between test trials lasted 6sec. A spatial 2AFC procedure was used. Contrast thresholds were determined using the method of constant stimuli. Post-pattern-adaptation thresholds were compared to thresholds when subjects were adapted to blank, colored fields. Two spatial frequencies of adapt and test patterns were used (2 and 8c/deg). Four colors were used in this set of experiments. These colors correspond to the maximum excursions possible on our monitor in the 0°–180° (L–M axis) and the 90°–270° (tritan axis) directions of MBDKL color space. Mean luminance was set to 20cd/m2 and the white point was set to illuminant C. Results. While significant inter-subject variability was found in post-adaptation threshold elevation, the general pattern of results seems to indicate some hue selectivity of threshold elevation. No significant difference was found between tritan axis and L–M axis color pairs or between 2 and 8c/deg patterns. Conclusions. Hue-selective luminance-encoding mechanisms seem to be operating at luminance contrast threshold. These mechanisms seem to operate equally for colors along both chromatic axes and at both low and high spatial frequencies.
Supported by NIH EY00014, NSF IBN-01111059 and an NTT grant.