FFA and OFA Encode Distinct Types of Face Identity Information

Faces of different people elicit distinct fMRI patterns in several face-selective regions the human brain. Here we used representational similarity analysis to investigate what type identity-distinguishing information is encoded three regions: fusiform face area (FFA), occipital (OFA), and posterior superior temporal sulcus (pSTS). In a sample 30 participants (22 females, 8 males), measure brain activity elicited by naturalistic videos famous identities, compared their distances each region with models differences between identities. We built diverse candidate models, ranging from low-level image-computable properties (pixel-wise, GIST, Gabor-Jet dissimilarities), through higher-level descriptions (OpenFace deep neural network, trained cluster faces identity), complex human-rated (perceived similarity, social traits, gender). found marked represented FFA OFA. Dissimilarities identities were accounted for perceived Social Traits, Gender, OpenFace network. contrast, OFA mainly driven image-based (pixel-wise dissimilarities). Our results suggest that, although can both discriminate representation further removed image, encoding perceptual information.SIGNIFICANCE STATEMENT Recent studies using have shown that face-responsive distinguish It however unclear whether these similar or ways. computations within response naturalistically varying revealed two regions, area, encode identity about faces. Although be decoded representations primarily contained gender, high-level visual features, whereas lower-level image features.