Abstract
Purpose. High spatial-frequency, high contrast monochromatic gratings appear desaturated, even at frequencies above the resolution limit. Similar desaturation is observed for high-frequency flicker. Here we measure the spatial and temporal frequency-dependence of this desaturation. Methods. Two normal subjects adjusted the modulation depth (or contrast) of a standard stimulus in order to match its subjective saturation to that of a test, presented in alternating temporal intervals. The stimulus source was either a red LED or a HeNe (red) laser. Exp. 1: a 30 Hz flicker was matched to test frequencies of 5 to 50 Hz. Exp. 2: a 30 cpd interference fringe grating was matched to test frequencies of 5 to 90 cpd. Temporal and spatial contrast sensitivity functions (TCSF & SCSF) were also obtained. Results. The saturation mechanism exhibited broader high-frequency tuning than that measured by the TCSF and SCSF. While sensitivity for pattern-detection dropped twenty-fold between 10 and 40 cpd, sensitivity of the desaturation mechanism was virtually unchanged out to 60 cpd, beyond the resolution limit; at 80 cpd, sensitivity was down only 0.5 log unit from the peak. From 20 to 30 Hz, the TCSF tracked a roughly tenfold loss in sensitivity; the effectiveness of flicker in desaturating color, however, was reduced by only 0.1 log unit over the same range. Conclusions. Desaturation at high spatial and temporal frequencies can be explained by an early (e.g. retinal) contrast-response nonlinearity. If such a nonlinearity demodulates the contrast signal, the resulting dc output would be similar for L- and M- driven cells, thus producing a loss of saturation. The frequency response for this stage would reflect spatiotemporal filtering prior to the nonlinearity; its dc output would be unaffected by later high-frequency attenuation that limits pattern and flicker detection.
Supported by NIH grant EY-01711.