FFA and OFA Encode Distinct Types of Face Identity Information

Area of Science:

• Neuroscience
• Cognitive Neuroscience
• Brain Imaging

Background:

• Distinct functional magnetic resonance imaging (fMRI) patterns are observed in face-selective brain regions for different identities.
• Understanding the specific identity-distinguishing information processed by these regions is crucial.

Purpose of the Study:

• To investigate the type of identity-distinguishing information encoded in the fusiform face area (FFA), occipital face area (OFA), and posterior superior temporal sulcus (pSTS) using representational similarity analysis.
• To compare representational distances within these regions with various models of identity differences.

Main Methods:

• fMRI data were collected from 30 participants viewing naturalistic videos of famous faces.
• Representational similarity analysis was employed to compare brain activity patterns with candidate models.
• Models included low-level image properties, a deep neural network (OpenFace), and human-rated social and perceptual features.

Main Results:

• The FFA represented identity differences based on perceived similarity, social traits, gender, and OpenFace network outputs.
• The OFA's representational distances were primarily driven by low-level image properties like pixel-wise and Gabor-Jet dissimilarities.
• Distinct encoding of identity information was observed between FFA and OFA.

Conclusions:

• Both FFA and OFA can discriminate between face identities, but they utilize different information types.
• FFA representations are more abstract, incorporating higher-level perceptual and social cues.
• OFA representations are more closely tied to low-level visual features of the faces.