Does exposure within an experiment affect the influence of familiar parts versus wholes on figure assignment?

Recurring questions in visual perception concern the role of experience and the relationship between parts and wholes. We investigated these questions in an experiment in which observers reported where they perceived a figure relative to a central border in bipartite displays in which the region on one side depicted either: (1) an identifiable configuration of a familiar basic-level object in its typical upright orientation (a "whole" familiar configuration); (2) an inverted version of (1); (3) a part rearranged (PR) version of (1) comprising the same parts as the upright familiar configuration; (4) an inverted version of (3). Previous research showed that regions are perceived as figures more often when they depict upright familiar configurations rather than inverted or PR configurations, supporting claims that familiar configurations are figural priors, but familiar parts are not. Here, displays depicting the four configuration types were counterbalanced over four trial blocks to elucidate the role of parts versus wholes in figure assignment and to examine whether previous experience with one configuration altered performance with another. Results showed that inverted familiar configurations were weaker figural priors than upright, but were stronger than PR configurations, revealing the importance of inter-part connectivity. Moreover, a reliable effect of block was observed for all configurations, revealing a role for familiar parts in figure assignment. Finally, no influence of the type of configuration viewed in block 1 was observed for any configuration in block 2. We attribute the pattern of results to neural populations representing parts and wholes of well-known objects.