Abstract
Scene understanding is a vital part of human's ability to function in the real world. A prominent question in scene perception research asks which aspect of scene information is being represented in different scene-selective regions. One prominent view holds that the parahippocampal place area (PPA) represents a scene's spatial layout or how open or closed a scene is. Another recent view suggests that this area encodes the low-level spatial frequency of scene images. Here we examined a third possibility: the actual location of an observer relative to a scene. This last possibility draws on an observation that a scene with an open spatial layout such as a beach when viewed from inside a room may indeed be considered closed, due to the closedness of an interior space that surrounds the observer. In this experiment, we attempted to tease apart the three possibilities by studying patterns of neural responses to different scene types varying in their level of image energy, spatial layout, and location of observers. In an fMRI block-designed experiment with one-back repetition task, participants passively viewed three types of scene images: 'inside-open' (i.e., open spatial layout scenes viewed from an inside), 'inside-closed' (i.e., closed spatial layout scenes viewed from inside), and 'outside-open' (i.e., open spatial layout scenes viewed from outside). We carefully chose all image stimuli such that the difference in the level of image energy reflected the distinction between the scene's open vs. closed layout but not inside vs outside location. Thus, under the image energy and the spatial layout hypothesis, scene-selective regions are expected to be more sensitive to the layout distinction than the observer's location distinction. The reverse pattern is true under the observer's location hypothesis. To anticipate, our preliminary MVPA results support the latter, suggesting the significance of the observer's location information in scene processing.
Meeting abstract presented at VSS 2017