Motion minima for different directions in color space

We have used the minimum-motion stimulus of Cavanagh, MacLeod & Anstis [(1987) Journal of the Optical Society of America A, 4, 1428-1438] to examine how signals along different directions in color space interact in motion perception. Stimuli were pairs of counterphasing gratings combined 90 deg out of phase in both space and time and modulated along different color-luminance axes. The axis for one of the gratings was fixed, while the axis for the second was varied so as to null perceived motion in the stimulus. The motion nulls show that observers are sensitive to motion signals carried by each of the cardinal directions of color space [an achromatic axis and L-M and S-(L+M) chromatic axes], but that signals along different cardinal axes are not combined to yield a net direction of motion. Pairing an achromatic and chromatic grating resulted in a motion null regardless of the relative or overall contrast of the two gratings, while the null directions for intermediate axes shifted depending on contrast. This result points to the special status of the luminance and chromatic axes. However, our results do not reveal a special pair of axes within the equiluminant plane. When contrasts along the cardinal axes are scaled for equal multiples of their respective detection thresholds, the L-M and S chromatic contrasts contribute roughly equally to the perceived motion, but are many times weaker than luminance contrast. Moreover, sensitivity to luminance motion is little affected by the presence of chromatic contrast, whereas sensitivity to chromatic motion is strongly masked by either luminance or chromatic contrast. These asymmetric interactions suggest that the motion of the luminance and chromatic components is encoded in qualitatively different ways.