Abstract
Purpose: :
The chromaticities distinguished by the normal eye all fall within a two-dimensional space that can be defined in terms of the ratios of photon absorptions in the three cones. In such a diagram, chromaticity varies continuously, but human perception imposes discontinuous categories of hue (e.g. Witzel & Gegenfurtner, 2011, Is there a lateralized effect for color? Journal of Vision, doi:10.1167/11.12.16). We ask whether discrimination is enhanced at the boundaries between categories, in the way that auditory discrimination is enhanced at the boundaries between phonemes (Liberman et al, 1957, The discrimination of speech sounds within and across phoneme boundaries, J. Exp. Psychol., 54, 358).
Methods: :
In interleaved experimental sessions, we obtained both (a) phenomenological measurements of the boundaries between reddish and greenish colors and between yellowish and bluish colors, and (b) forced-choice performance measurements of discrimination along lines orthogonal to the phenomenological boundaries. The stimuli were foveal, subtended 2 degrees, had a duration of 150 ms, and were presented on a steady background of 10 cd/m2. In the performance experiments, the field was vertically divided by a thin line and the difference in chromaticity of the two halves was adaptively varied on the basis of the observer's accuracy in discrimination. The chromaticity of the background was varied in different experiments.
Results: :
(i) Near the subjective boundary between reddish and greenish hues, there are clear minima in the forced-choice thresholds when thresholds are expressed in terms of the change in ratio of cone absorptions. (ii) When the chromaticity of the background is changed, there is a concomitant shift both in the red-green subjective boundary and in the minimal thresholds. (iii) No enhancement is seen at the boundary between yellowish and bluish hues, i.e. at chromaticities that appear unique green.
Conclusions: :
To explain results (i) and (ii), it is plausible to postulate a neural channel that draws synergistic inputs from short-wave and long-wave cones and an opponent input from middle-wave cones. This channel would be in equilibrium, and at the most sensitive point of its operating range, at the chromaticity of the background. Result (iii), however, implies that an analogous explanation cannot be given for all category boundaries.
Keywords: color vision • perception • color appearance/constancy