Abstract
Recent work has suggested that visual scene processing is composed of two distinct systems: one for navigation, including the occipital place area (OPA) and retrosplenial complex (RSC), and the other for scene categorization (e.g., recognizing a scene as a kitchen), including the parahippocampal place area (PPA). If scene processing is indeed composed of these two systems, then it should be possible to selectively modulate the two systems by changing task demands to reflect their functions. Thus, we used fMRI to test our prediction that PPA would respond more to an image of a scene when a viewer was performing a 'categorization' task on an image than when the same viewer was performing a 'navigation' task on the exact same image, while OPA and RSC would show the opposite pattern. Crucially, both tasks were completed on the exact same stimuli, behavioral performance was matched between tasks, and participants were required to fixate during both tasks, ensuring that any neural differences between tasks were not due to low-level visual stimuli differences, differences in task difficulty, or differences in eye movements. As predicted, we found that PPA responded significantly more during the categorization than navigation task, while OPA showed the opposite pattern. Interestingly, however, RSC responded similarly to both tasks. In a second experiment conducted outside of the scanner, we used eye tracking in a separate group of participants to further test our two-systems-for-scene-processing hypothesis. The design of this experiment was similar to the first experiment in every way, except now the participants' eyes were free to wander. Consistent with our hypothesis, we found a different pattern of eye movements when participants performed a navigation task compared to a categorization task. Taken together, these results suggest distinct neural and cognitive systems selectively involved in navigation and categorization of scenes.
Meeting abstract presented at VSS 2016