The illusion of a neural common factor for illusions

Abstract

The source of variability in the susceptibility to visual illusions has been the subject of long-standing debates. Studies have explained variability in susceptibility to size illusions with idiosyncrasies in the surface area of early visual areas or the population receptive field (pRF) sizes in these areas. However, recent research has revealed weak between-illusion correlations, raising questions about the reproducibility or generalizability of these neural findings. Here, we tested 30 participants on a battery of 13 visual illusions. The illusion battery included three illusions that were previously shown to correlate with either V1 surface area or pRF size (the Ponzo “hallway”, Ebbinghaus and Delboeuf illusions), four other size illusions, and six control illusions for which size is irrelevant (contrast, uniform texture, and perceived orientation illusions). The same participants underwent a 3T fMRI experiment that included pRF mapping. For each participant, we estimated the surface areas and pRF sizes in V1 to V4. At the behavioral level, we confirmed our previous results of weak between-illusion correlations and high test-retest reliability. Contrary to previous results, we found no significant correlations between illusion magnitude and visual surface areas or pRF size. Our results fail to support the existence of both a behavioral and a neural common factor underlying visual illusions.