Contour interaction, spatial extent, and size contrast in a visual size illusion

Previous research has documented multiple contextual influences on perceived size in classic visual illusions (e.g., Ebbinghaus and Delboeuf), including contour interactions (e.g., Todorović and Jovanović, 2018, Acta Psychologia), spatial extent (Kirsch and Kunde, 2021, Vision Research), and size contrast (e.g., Roberts et al. 2005, Perception). Such studies typically focus on one factor, as multiple factors are difficult to disentangle (but see Roberts et al., 2005). However, the postulated contextual mechanisms are not mutually exclusive. We sought to quantify the relative contributions and potential interactions of size contrast, contour interactions, and spatial extent on illusory changes in perceived size. Participants reported which of two squares (“standard” or “target”) appeared larger. The standard (1.5°) was either isolated (control) or surrounded by four inducers, each composed of a single parallel line and a single perpendicular line (forming the shape of plus or a tack). Across five experiments, we systematically manipulated the separation between the central square and (1) the perpendicular line, (2) the innermost point of the parallel line, and (3) the outermost point of the parallel line (1.5°, 3.25°, of 5° in all cases). This combination of stimulus configurations allowed us systematically explore changes in contour interactions (between parallel edges of the target and inducers), spatial extent (defined as the outermost extent of the inducers), and size contrast (based on the overall size/length of the inducers). The point of subjective equality (PSE) was extracted from psychometric curves (target square size range: 1.2° to 1.8°). We found two main factors influencing perceived size: (1) contour interactions for perpendicular contours near the target made the central target appear larger and (2) larger spatial extents made the central target appear smaller. Overall, our results are consistent with a contour interaction and spatial extent influencing perceived size, with no specific contribution from size contrast.