Time course of rod influences on hue perception.

Stimulation of dark-adapted rods can shift the hues associated with specific wavelengths throughout the spectrum: Rods exert a green bias (strengthen green relative to red) at longer wavelengths and a blue bias (strengthen blue relative to yellow) at short-to-middle wavelengths. A third rod influence at shorter wavelengths is more complicated because it has been shown to reverse direction with change of stimulus duration. Thus, for 30-ms stimuli, rods exert a green bias like that observed at longer wavelengths. However, for 1-s stimuli, rods exert a red bias that is observed nowhere else in the spectrum. We examined the latency (time course) of rod hue biases by measuring the shifts of the three spectral unique hues under dark-adapted versus bleached (cone plateau) conditions. The rod green bias at unique yellow (mean 10 nm) and, in contrast to some prior studies, the rod blue bias at unique green (mean 21 nm) were not systematically affected by test stimulus duration. A quick rod green bias (mean 5 nm) was shown at unique blue for two of three observers but was dominated by a slower rod red bias (mean 11 nm) after 30-50 ms of rod stimulation. These opposing rod influences may reflect competing effects of rod signals on ML-cone and S-cone pathways.

Foveal and extra-foveal influences on rod hue biases.

Green, blue and short-wavelength-red rod hue biases are strongest and most reliable with large, dimly-mesopic, extra-foveal stimuli but tend to diminish when stimuli are confined to a small area of the central fovea. This study explores how the stimulation of foveal and extra-foveal areas interact in determining rod hue biases, and whether large stimuli are as effective for revealing rod hue biases when foveally centered as when eccentrically centered. We assessed rod influence by measuring wavelengths of unique green and unique yellow (with 1-s duration, 1 log scot td stimuli and a staircase procedure) under bleached and dark-adapted conditions. We measured unique hues with foveally centered 2 degrees - and 7.4 degrees -diameter disks, a 7.4 degrees (outer) x 2 degrees (inner) diameter annulus, and a 7 degrees -eccentric, 7.4 degrees -diameter disk. The rod green bias (shift of unique yellow locus) was typically <10 nm and remained fairly constant across spatial configurations, indicating no special foveal influence. The rod blue bias (shift of unique green) varied more among observers and spatial configurations, reaching up to 47 nm. However, stimuli covering the fovea typically produced no rod blue bias. Thus, the present results add differences in spatial dependence (i.e., foveal/extra-foveal interaction) between green and blue rod biases to previously demonstrated differences (e.g., differences in amount of light level dependence, in time course and in the spectral range influenced by each bias).

Do rods influence the hue of foveal stimuli?

To understand the generality and mechanisms of previously reported rod hue biases, we examined whether they are present for small foveal stimuli by comparing the wavelengths of the three spectral unique hues under dark-adapted and flash-bleached conditions. Rod green bias (shift of unique yellow) and rod blue bias (shift of unique green) were found for some observers with 1 degrees -diameter foveal stimuli, the size most likely to stimulate rods. Smaller stimuli (0.2 degrees and 0.6 degrees diameter), which were least likely to stimulate rods, produced no large or consistent differences between dark-adapted and bleached conditions. This suggests that rod hue biases result from the local stimulation of rods by light, not from remote suppression by dark-adapted, unstimulated rods, and not from bleaching light artifacts.

Generality of rod hue biases with smaller, brighter, and photopically specified stimuli.

This study tests the generality of previously demonstrated rod hue biases (red and blue biases at shorter wavelengths and a green bias at longer wavelengths) that cause the loci of the three spectral unique hues to shift to longer wavelengths. We found rod hue biases for 2-deg targets to be generally similar in magnitude and light-level dependence to those observed for 7.4-deg targets (the size most often studied) when measured at 7-deg eccentricity. The largest effects for both test sizes occurred at the lowest light levels tested, 1 log scotopic troland. All three rod hue biases were found with 0.6-deg targets, but were not reliably measurable at the lowest light levels and were reduced in magnitude and consistency across observers. The largest rod hue biases all occurred at the same scotopic light level, which corresponds to different photopic light levels for the three hue biases, because of differences in photopic and scotopic spectral sensitivity. This suggests that no single photopic light level will produce such large effects for all three rod hue biases. Finally, when the rod influence on a specific unique-hue locus was measured using photopically (rather than scotopically) constant stimuli, rod hue biases were still found but were more variable in magnitude and incidence across observers. We conclude that the rod hue biases we have previously described can be found with smaller stimuli, at somewhat higher light levels, and under photopically constant conditions, although our prior conditions tend to produce larger, more reliable rod hue biases.