Humans and animals encode representations of the large-scale spatial structure of the environment that allow them to understand the spatial relationships between real-world locations and to navigate accurately between them. We used a continuous carry-over fMRI design to probe the neural concomitants of these hypothesized cognitive “maps” in humans. Subjects were scanned with fMRI while viewing images of 16 locations from a familiar college campus. Four images were acquired for each location; these images depicted the views one would see facing Northeast, Northwest, Southeast, or Southwest. On each of many trials, subjects were presented with one image and a reference direction (North, South, East, or West), and reported whether one would rotate left or right to face the reference direction. This task encouraged subjects to recover geographic place information, but did not require them to explicitly manipulate information about allocentric (map) position. We determined how (1) trial-by-trial fMRI adaptation and (2) the distributed pattern of fMRI response to each place varied as a function of distance between locations. Distance was defined both in physical space, and also psychologically by querying subjects about the perceived distances between locations in a separate behavioral session. Preliminary results indicate that the differences in the spatially distributed patterns of fMRI response to different locations varied significantly as a function of the physical and psychological distances between these locations. These results indicate that fMRI can be used as a tool to query the cortical representations of large-scale space.
This study was supported by NIH R01 EY-016464 and Whitehall Foundation 2004-05-99-APL