Abstract
Crowding is the failure to recognize an object due to surrounding clutter. We quantify crowding by the crowding distance, the minimum spacing between a target object and flankers needed for recognition. To characterize the statistics of crowding, we measured crowding distance for both radial and tangential flankers in 50 observers at 12 locations: three eccentricities (0, 5, and 10 deg) along the four cardinal meridians. Each threshold was measured twice. We fit the well-known Bouma law — crowding distance grows linearly with radial eccentricity — to log crowding distance of the 50 participants, explaining 82% of the variance, cross-validated. We then fit an enhanced Bouma model, with factors for meridian, flanking direction, target kind, and observer, explaining 94% of the variance, again cross-validated. The enhanced model improves the fit in several ways. At a fixed eccentricity, crowding distance varies two-fold across meridians, observers and flanking direction. We also present a peeking model that allows the same Bouma law to fit two sets of spacing thresholds, one measured with unmonitored fixation and one with accurate fixation. The associations of crowding with auditory informational masking, reading difficulty and size of hV4 suggest that it might be a useful biomarker for development and cortical health. Favoring that use, it can be measured in two minutes with a standard deviation across observers that is three times larger than the standard deviation of test-retest.