Physiological mechanisms underlying psychophysical sensitivity to combined luminance and chromatic modulation.

If psychophysical detection thresholds are plotted in a middle-wavelength-sensitive (M) and long-wavelength-sensitive (L) cone coordinate system, the shape of the contour can be used to infer underlying detection mechanisms. We measured responses of macaque ganglion cells to combine chromatic and luminance modulation and expressed our results in such an M,L-cone space. Our aim was to test whether, with the use of this space, readily separable luminance and chromatic psychophysical mechanisms might be expected from physiological data. For parvocellular pathway cells, detection contours approximated elongated ellipses with maximum responsivity to chromatic modulation. The degree of elongation decreased as temporal frequency increased. Responses could be well described by linear subtraction of M- and L-cone signals, with a phase delay of 1-3 deg/Hz. For cells of the magnocellular pathway, detection contours were more complex. Orientation was variable between cells and temporal frequency dependent, and a frequency-doubled component was evoked by chromatic modulation. In relation to psychophysical detection thresholds plotted in such a space, the properties of parvocellular-pathway cells were sufficiently linear and homogeneous to make it plausible that this pathway might form the substrate for a linear chromatic mechanism. The properties of magnocellular-pathway cells, however, indicate that, insofar as a psychophysical luminance mechanism is based on their activity, its signature in the M,L-cone contrast space would be more difficult to identify.