Shape-independent object category responses revealed by MEG and fMRI decoding.

Area of Science:

• Cognitive Neuroscience
• Neuroimaging
• Visual Perception

Background:

• Neuroimaging studies show distinct neural patterns for object categories like faces and scenes.
• A key debate exists whether these patterns reflect true category selectivity or shape-associated properties.

Purpose of the Study:

• To investigate concurrent shape and category selectivity in the human visual cortex.
• To differentiate between category-specific and shape-specific neural responses using cross-classification.

Main Methods:

• Utilized functional MRI (fMRI) and magnetoencephalography (MEG) with a cross-classification analysis.
• Trained multivariate classifiers to distinguish shape within categories (bodies vs. clothing) and category within shapes (hands vs. gloves).
• Presented participants with body parts (hands, torsos) and shape-matched clothing items (gloves, shirts).

Main Results:

• Identified category-independent shape responses in visual cortex (fMRI) and from 90 ms (MEG).
• Detected shape-independent category responses in occipitotemporal cortex (fMRI) and from 130-200 ms (MEG).
• Demonstrated that category-level responses are not solely explained by 2D shape properties.

Conclusions:

• The human visual cortex exhibits concurrent selectivity for both shape and object category.
• Neural responses demonstrate that category information can be processed independently of specific shape features.
• Findings challenge explanations of category selectivity based purely on visual shape characteristics.