Abstract
The human visual system shows a division of labor between object-processing pathways and scene-processing pathways. However, in continuous visual experience, singleton objects and full-scale scenes represent only the extremes—intermediate scales of space, e.g. the view of a desk-top or a kitchen counter, are also frequently experienced. How do these intermediate "reachspaces" drive visual cortex? On one hand, reachspaces are like scenes in that they have spatial structure and contain multiple objects, and thus may drive neural responses similar to full-scale scenes. However, in visual search behavior, reachspaces perceptually dissociate from full-scale scenes (Josephs & Konkle, VSS 2017), raising the possibility that there may be neural regions tuned to reachspace features. To test this, in Experiment 1, observers (N=10) underwent whole-brain functional neuroimaging while viewing pictures of objects, scenes, and reachspaces in a standard blocked design (N=50 images per view type). A conjunction analysis revealed a bilateral region in posterior lingual gyrus with a stronger response to reachspaces over both singleton objects and full-scale scenes. Further, in all classic scene and object regions, reachspaces dissociated from both objects and scenes with an intermediate response magnitude. In Experiment 2 (N=12), we replicated the existence of the lingual reachspace-preferring region (10/12 subjects), and found some evidence for additional reachspace-preferring regions in inferior parietal sulcus (6/12 subjects) and superior parietal lobule (8/12 subjects). Taken together, these results establish that reachspaces have a distinct representational signature from both scenes and objects. Broadly, they provide initial evidence that intermediate-scale spaces engage a distinct network of regions, in addition to object and scene networks, comprising a new large-scale division of neural responses.
Meeting abstract presented at VSS 2018