Abstract
Crowding is a key limiting factor of form vision in the periphery. The neural origin of crowding is not known. Here we tried to establish an upper bound on the earliest neural stage of crowding by rendering the flankers invisible with interocular suppression.
We presented to the non-dominant eye of a subject a target Gabor at 5-degree eccentricity surrounded by 1–4 Gabor flankers on four sides in Experiment 1 and zero, 1, or 4 flankers in Experiment 2. We presented to the dominant eye four 100%-contrast dynamic concentric checkerboards (suppressors) at the same positions of the flankers. Subjects performed two tasks. The first task was to discriminate (clockwise vs. counter-clockwise) the orientation of the target Gabor relative to the reference direction of 45 degrees. The second task was to indicate the number of Gabors seen. A fixation and a pair of nonius lines were used to achieve fusion lock. Initiated by the subject, stimuli for both eyes were presented simultaneously for 100 ms. The contrast of the Gabors was 60% in Experiment 1 and 50% in Experiment 2.
Since there were always four suppressors, when the flankers were fully suppressed (82% of the trials in Experiment 1, 100% in Experiment 2), the subjective percept of the stimulus was unaffected by the actual number of flankers. Nevertheless, we found that for both experiments, discrimination accuracy decreased as the number of the flankers increased for the trials when the flankers were invisible. For the remaining 18% of the trials in Experiment 1, during which subjects reported seeing one or more flankers, the reported number of flankers accounted for no additional variance in discrimination accuracy. These results suggest that crowding starts at a visual-processing stage prior to interocular suppression, which is likely to be early, such as in V1.
National Institutes of Health Grant R01-EY017707.