Separated and overlapping neural coding of face and body identity

Area of Science:

• Neuroscience
• Cognitive Psychology
• Computer Vision

Background:

• Person recognition relies on face and body cues, with tolerance to visual variations like viewpoint.
• Previous research localized face identity processing to anterior face-responsive regions.
• The neural basis for body identity processing and its integration with face identity remains unclear.

Purpose of the Study:

• To investigate brain regions involved in disentangling body identity from viewpoint variations.
• To determine if face and body identities are encoded separately or integrated into a unified person identity representation.
• To explore the role of attention in the neural representation of person identity.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to record brain activity.
• Participants viewed face and body images of three distinct identities from various viewpoints.
• A task-based fMRI design involved participants attending to either identity or viewpoint.

Main Results:

• Body identity was consistently decoded across viewpoints in the fusiform body area, anterior temporal cortex, middle frontal gyrus, and insula.
• Identity information was decodable across both face and body stimuli in early visual cortex, occipital cortex, parahippocampal cortex, and parietal cortex.
• Attending to identity significantly improved identity decoding accuracy, enhancing its neural representation.

Conclusions:

• The fusiform and anterior temporal cortices play a role in extracting high-level identity information from bodies, similar to their role in face processing.
• A distributed neural network supports an abstract representation of person identity, integrating information from both faces and bodies.
• Attention modulates the neural coding of person identity, highlighting its importance in recognition.