Timothy Vickery¹, Kallie Sweetman¹, Mary Peterson²; ¹University of Delaware, ²University of Arizona

The presence of an object in a 2-D display distorts perceived space. For example, two dots placed on a region perceived to be an object appear farther apart than two equally spaced dots placed on region that appears to be its background ("object-based warping"). The extent to which perceived objecthood per se generates this distortion, as opposed to image factors, is still debated. We asked whether cues that lead to border assignment in classic figure-ground displays could generate distortions of space when other factors are held constant. We employed horizontally elongated rectangular images divided into eight alternating black and white convex and concave regions, presented on a gray background. (The black/white color of convex and concave regions was balanced and they were equal in area). In such displays, the convex regions are highly likely to be perceived as figures/objects and the concave regions appear to form a background. We placed two fixed-spacing "reference" dots on either a convex or concave region, and two "adjustment" dots on the other region type. Subjects adjusted dot spacing to match the reference. In a control condition, all dots were placed on a uniform rectangle of white or black. Subjects systematically estimated the dots as farther apart when the reference dots were on a figure region versus a ground region (p < .001). Furthermore, ground-region estimates were significantly underestimated compared to control estimates (p < .001) and figure-region estimates were significantly overestimated relative to control estimates (p < .001). Thus, even on displays matched for local boundedness of the dots, assignment of relative objecthood distorted space, such that perceived spatial distance was expanded in object regions, and shrunk in ground regions. The use of such figure-ground displays allows for the matching of many important low-level properties in the study of object-based distortion effects.

Acknowledgements: Supported by NIH NIGMS grant 5P30GM145765-03