Abstract
Purpose:
We attempted to find how stereo channels of different spatial scales interact when the observer detected the presence of disparity.
Methods:
We employed a 2afc method to measure disparity thresholds for compound gratings consisting of sinewaves of spatial frequency combinations F+3F, F+5F, and F+3F+5F. The fundamental frequency F was between 0.25c/d and 1c/d for different subjects. Each component was presented at a variety of contrasts to vary its detection threshold. We manipulated the disparities of the components by either yoking them in space (same spatial offset) or in phase (same phase offset) and compared thresholds for the compound gratings with thresholds for the components when tested alone.
Results:
When component disparities were yoked in space, thresholds were generally consistent with independent detection of the components' disparities. Thresholds were determined by the most sensitive channel, with the thresholds for compound gratings equaling the lowest of the component thresholds, in agreement with Heckmann and Schor (1989). However, when disparities were yoked in phase, the results were inconsistent with independent channels. Rather than equaling the lowest component threshold, compound thresholds were similar to or even higher than the highest of the component thresholds.
Conclusion:
To account for interactions among stereo channels in disparity detection, we hypothesize that disparity thresholds for the compound gratings are determined by a single channel among those activated by the stimulus. This channel is non-optimal, biased toward the high-frequency components of the compound, regardless of which components have non-zero disparities.
Supported by NEI Grant EY12286