More than meets the eye: Neural evidence for scene grammar representations during individual object processing.

Published: April 22, 2024

Objects in real-world scenes typically follow regular, hierarchically structured arrangements, where anchor objects (e.g., a sink) guide the placement and identification of associated local toothbrush), forming clusters referred to as "phrases." Depending on actions performed within scene, it can consist multiple such phrases. According scene grammar framework, these hierarchical relationships enable brain efficiently process complex behavior by leveraging statistical regularities object arrangements. This study investigates whether shared neural representations reflect this organization. Using EEG, we explored temporal dynamics phrase-specific through cross-classification analysis. Specifically, classifiers were trained data from one type (either or its object) tested for their ability generalize other between phrases, well across different scenes. Our findings reveal an early time window (128–164 ms) that supports existence representations. Additionally, representational similarity analysis (RSA) revealed predominantly rely high-level visual semantic features, implied actions, rather than low-level similarities. Notably, "upward" generalization (local anchor) is driven primarily while "downward" (anchor local) influenced features actions. These provide evidence processing mirrors behaviorally relevant structure

Language: Английский

Object Representations Reflect Hierarchical Scene Structure and Depend on High-Level Visual, Semantic, and Action Information

Published: Jan. 1, 2024

Language: Английский