Neural Patterns Reveal Lateral Occipital Complex Representation of Ensemble Mean Orientation

Area of Science:

• Neuroscience
• Cognitive Psychology
• Visual Perception

Background:

• Ensemble perception, the brain's ability to extract summary statistics from groups of elements, is well-studied behaviorally but its neural basis is unclear.
• Classical theories propose object features are stored in "object files," with recent work suggesting ventral stream representations form feature basis sets.

Purpose of the Study:

• To investigate whether average orientation is represented in early visual cortex or higher-level object-selective regions.
• To test the hypothesis that ensemble statistical characteristics are encoded in "ensemble files" within object-selective areas like the Lateral Occipital Complex (LOC).

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to study neural representations of ensemble average orientations from visual bars.
• Participants performed orientation estimation and two-alternative forced choice tasks while undergoing fMRI scans.

Main Results:

• Modest, broadly-tuned representations of average orientation were found in early visual area V3 and parietal regions.
• The strongest representation of ensemble average orientation was identified in the Lateral Occipital Complex (LOC).
• Multi-Voxel-Pattern Analysis (MVPA) successfully decoded ensemble orientation averages in LOC, with decoding strength correlating with participant performance.

Conclusions:

• The Lateral Occipital Complex (LOC) plays a major role in representing the average orientation of visual ensembles.
• These findings support the "object file" hypothesis and extend it to ensemble characteristics, suggesting high-level ventral stream involvement.
• The representation of ensemble statistics in LOC highlights its importance in understanding how the brain processes grouped visual information and defines objects.