Abstract
PURPOSE. For large, extrafoveal, 1-s duration targets presented at low mesopic light levels, we have previously shown three types of rod hue bias that can be sorted into two groups. A quick rod green bias (enhancement) apparently operates via LM-cone pathways. Blue and red rod hue biases build up more slowly and apparently operate via S-cone pathways.
METHODS. The present study addresses the generality of these rod hue biases to stimuli that are brighter, smaller, and foveally centered. We inferred rod hue biases from measurements of rod influence on the wavelengths of the three spectral unique hues (blue, green, yellow) by means of a double staircase procedure. Rod influence was assessed by comparing unique hue wavelengths for the same stimuli under bleached (minimized rod contributions) and dark-adapted (maximized rod contributions) conditions.
RESULTS. At lowest light levels (1 log scot td), largest target size (7.8-deg diameter), and greatest eccentricity (7-deg nasal retina), rods shifted all three spectral unique hues to longer wavelengths. Consistent with past results, rods produced the largest (∼35-nm) shifts of unique green (blue bias) and 5–15-nm shifts of unique blue (red bias) and unique yellow (green bias). As we increased light level over a 2-log-unit range or reduced stimulus size to 2-deg diameter, the rod green bias tended to be the most constant and the rod blue bias and red bias tended to disappear. All three rod hue biases tended to disappear when small stimuli were centered on the fovea. However, there was individual variation among observers in all of these trends.
CONCLUSIONS. Our prior stimulus conditions maximize the three rod hue biases. Observers vary in generalizability of these to brighter, smaller, and foveated stimuli. We found no condition, within our parameter ranges, that assured the absence of rod hue biases in all observers.