Abstract
The visual system is able to maintain a degree of constancy of the color appearance of objects in the presence of variations in the spectral content of the illumination. It has been suggested that the physiological basis for color constancy involves processing at the level of the cortex in which specialized double-opponent neurons compute local ratios of cone activity. There is evidence that the neural components responsible for color constancy arise in an independent neural pathway from those required for detection and wavelength discrimination because color constancy can be selectively impaired after circumscribed lesions in cortex. Similarly, we found a striking selective impairment in color constancy in a patient with mutations in the GRM6 gene that normally encodes the glutamate receptor mGluR6. This neurotransmitter receptor is present only in the synapses of the ON bipolar cell dendrites. The patient showed normal color discrimination when tested on a battery of color vision tests including the Farnsworth Panel D-15, 100-Hue test, the Ishihara and HRR color plates, the Nagel anomaloscope and the Cambridge colour test. However, he responded abnormally to color samples with reflectances along the tritan axis when they were shown under an illuminant that had a preponderance of short wavelength light. Under this illuminant, normal subjects see samples with bluish reflectances as less blue than would be predicted from the wavelength content of the light reaching the eye from the sample alone in the absence of its surroundings. Similarly, under a bluish illuminant, they see samples with yellowish reflectances as more yellow than would be predicted from the wavelength content of light reaching the eye from the sample alone. These results are consistent with a normal ability to partially discount the extra short wavelength light in the illuminant. In contrast, the patient reported blue color samples to be dramatically more bluish than normal and yellow samples to be more yellowish than normal under the short wavelength light-biased illuminant. Thus, compared to normal, his system appeared to grossly under correct blue samples for the illuminant while it overcorrected the yellow samples. These results have implications for understanding the physiological mechanisms underlying color constancy.