Home >  Events

fMRI brown bag : February 28

Please join us a for a talk by Adam Steel, a post-doctoral researcher in the lab of Dr. Caroline Robertson, at Dartmouth College.

Distinct neural substrates for scene perception and imagery

Time: 12:00-1:00pm

Place: Moore 418

Abstract:

It is widely assumed that perception and mental imagery of high level visual stimuli engage the same neural circuitry. Here, we report surprising evidence to the contrary.

First, we report three regions of the human brain are selectively activated during memory of familiar places (Experiment 1). Fine-grained individual-subject fMRI analyses revealed that these regions partially overlapped with, but were distinct from the perceptual scene network (the parahippocampal place area, occipital place area, and medial place area/RSC). 

Next, we confirmed that these three areas constitute a “place memory network.” In Experiment 2, we showed that these areas form a functionally distinct network that is preferentially connected to the hippocampus during naturalistic movie viewing. In Experiment 3, we found that these areas are preferentially activated by visual presentation of a scene when the surrounding (out of view) context is known.

In Experiment 4, we directly compared activation during perception with mental imagery of places. We found that the scene perception network showed greater activation than the place memory network during perception. In contrast, during mental imagery, the place memory network was significantly more active than the scene perception network. Remarkably, the scene perception areas PPA and OPA showed significantly below baseline activity during mental imagery. 

Taken together, these results demonstrate that the place memory network, a set of brain regions adjacent to, but distinct from, the perceptual scene network support the mental imagery of scenes. Given the preferential selectivity of the place memory areas for memory and their enhanced responsiveness to familiar scenes, we hypothesize that these areas might provide remembered, contextual representations to support ongoing perception during navigation.