Abstract
Much effort has been focused on understanding how scenes and landmarks are represented in the brain. However, a landmark or scene always has a set appearance; are there brain regions that also represent the more abstract idea of the "place" represented by a landmark or scene? Here we use two completely different visual stimuli that indicate the same place–images of the inside and outside of a building—to see if we can elicit a common code corresponding to that place. Students at the University of Pennsylvania were scanned with fMRI while they viewed photographs of the interiors and exteriors of ten familiar landmarks from the Penn campus. 22 interior views and 22 exterior views were shown for each landmark and no picture was repeated during the course of the experiment. Using multi-voxel pattern analysis, we were able to decode the identity of landmarks from their exteriors based on activity in scene-responsive regions, including parahippocampal place area (PPA) and retrosplenial complex (RSC), replicating previous results from our lab (Morgan et al., 2011). In addition, preliminary results suggest that RSC may contain sufficient information to decode the identity of a landmark’s exterior from images of its interior, and vice versa. This generalization across building interiors and exteriors suggests the existence of more abstract coding of place identity within RSC, conveying information that both interior and exterior views indicate the same ‘place.’ Such a representation might provide a mechanism for linking building exteriors to the spaces that lie within.
Meeting abstract presented at VSS 2013