Abstract
Crowding distance and size of cortical visual areas vary several-fold across individuals. Multiplying perceptual crowding distance (deg of visual angle) by cortical magnification (mm of cortex per deg) yields cortical crowding distance in mm. Cortical crowding distance is conserved across eccentricity in V1, V2, V3 and hV4 (Pelli, 2008). It is approximately conserved across orientations of the axis of flankers in hV4 but not V1, V2, and V3 (Zhou et al., VSS, 2018). Here we assess conservation across individuals. For each visual area we test the hypothesis that cortical crowding distance (mm) is conserved across individuals, despite individual variations in map size (mm2) and crowding distance (deg). With fMRI, we measured retinotopic maps in 26 observers and manually defined the boundaries of V1, V2, V3 and hV4. For each of these regions we calculated surface area in mm2. In the same individuals we measured crowding distance at 12 locations (4 meridians, 3 eccentricities) of the visual field. We use these psychophysical data to build a model of crowding distance from which we derive the Bouma factor, B, (slope of crowding vs eccentricity) for each individual subject. Conservation predicts that the Bouma factor scales inversely with the square root of area. Combining fMRI with psychophysics revealed that cortical crowding distance is approximately conserved in hV4 but not in V1, V2, or V3. Specifically, in hV4 the Bouma factor scaled nearly with the inverse square root of surface area predicted by conservation (exponent of –0.55, 95% CI [-0.2 -0.9]): Observers with smaller hV4 tend to have a larger Bouma factor and thus need a larger target-flanker spacing to identify the crowded target. The hV4 surface area explained ~30% of the variance in Bouma factor across observers. More work is needed to understand the sources of the remaining variance.