Behavioural and Visual Evoked Potential Measures of the Role of Disparity in Stereo Detection and Discrimination

Abstract

The initial experiment in this research used psychophysical responses to dynamic random-dot stereograms (DRDS) to examine disparity detection and disparity or depth discrimination in stereopsis. Using identical crossed and uncrossed disparity stimuli across conditions at 150 ms exposure, we found that while disparity could be detected and form discriminated, observers were unable to discriminate depth direction. These results suggest that disparity detection is a faster process than disparity discrimination and that form can be discriminated prior to the assignment of depth direction. In a second experiment we examined disparity (correlation) and noncorrelation in stereopsis and used visual evoked potentials (VEPs) in response to three DRDS displays: (a) a target containing disparity on a correlated background, (b) an uncorrelated target on a correlated background, and (c) a target containing disparity on an uncorrelated background. At 200 ms stimulus exposures, VEPs to regions of disparity and regions of noncorrelation on correlated backgrounds were largely undifferentiated. There was not, however, any discernible VEP to stimuli incorporating disparity on an uncorrelated background. These findings imply that while form may be discriminated on the basis of a region of disparity (correlation) in stereopsis, the uncorrelated background condition suggests this is a slow process, since there was no apparent response to form or depth in the recording epoch. Early processes may regard the entire DRDS as a region of noncorrelation with both target and background being mismatched, and hence initially undifferentiated. A faster process may discriminate form on the basis of a region of noncorrelation, as suggested by the correlated background conditions.