Thresholds to chromatic spots of cells in the macaque geniculate nucleus as compared to detection sensitivity in man

1. The relation between wavelength and psychophysical threshold for chromatic spots on a white background provides evidence for the existence of chromatic channels in the primate visual system. To find the physiological substrate of this task, we compared increment thresholds of different cell types in the macaque lateral geniculate nucleus with human psychophysical thresholds to the same stimuli, using two spot sizes, 4 and 0.4 deg. 2. At different wavelengths, different opponent cell classes in the parvocellular layers of the nucleus were most sensitive, so that at long wavelengths (greater than 600 nm) red on-centre cells were most sensitive, while at short wavelengths (less than 500 nm) S-cone, blue on-centre cells were most sensitive, from 500 to about 550 nm green on-centre cells being most sensitive. A rare cell type with inhibition from S-cones was most sensitive at about 570 nm, although its maximum contrast increment sensitivity was poor compared with that of other cell types. Variation in strength of cone opponency caused a considerable range in threshold in each of the opponent cell classes of the parvocellular layers. 3. On- and off-centre cells from the magnocellular layers were more sensitive than opponent cells to white and yellow spots (as is the case with achromatic gratings). 4. With different wavelengths and spot sizes, the most sensitive cells found approached (to within 0.1-0.3 log units) human psychophysical sensitivity, suggesting that the most sensitive cells available may underlie detection. 5. Measurements of psychophysical chromatic discrimination thresholds, both with nearly monochromatic spots and with spots of differing saturation (purity), support this hypothesis. When magnocellular cell sensitivity corresponded to psychophysical threshold, a suprathreshold stimulus, capable of activating opponent cells, was required for chromatic discrimination.