Abstract
The parahippocampal place area (PPA) responds strongly to environmental stimuli such as landscapes, buildings, and urban scenes. Previous work suggests that the PPA encodes perceptual aspects of these stimuli, such as scene geometry and visual statistics. Whether the PPA also encodes the abstract identity of these items—the "place" as well as the "scene"—is unknown. Here we examined this question by testing whether two very different visual stimuli that correspond to the same place—images of the inside and outside of a building—elicited a common code in the PPA. Participants were scanned with fMRI while they viewed multiple photographs of the interiors and exteriors of ten landmarks from the University of Pennsylvania campus. The participants fell into two groups: Penn students, who were familiar with the buildings and thus knew the correspondences between the interior and exterior images, and Temple University students, who did not have this knowledge. In both subject groups, multivoxel analyses of activity patterns in the PPA revealed that the identity of interiors could be decoded from interior images, and exterior identities from exterior images—an unsurprising result that might be driven by visual or geometric similarities between the images. Notably, however, cross-decoding between interior and exterior images was only significant in Penn students—in Temple students it was at chance. A whole-brain searchlight analysis confirmed that this familiarity-dependent cross-decoding effect was only found in the right PPA. We hypothesize that the PPA constructs abstract "place" representations for familiar landmarks that develop through navigational experience.
Meeting abstract presented at VSS 2014