Abstract
The face processing network consists of a core system, involved in the processing of visual appearance of faces, and an extended system, involved in the retrieval of person-knowledge and other nonvisual information associated with faces. Behavioral evidence shows that personally familiar faces are processed in a more robust and efficient way than unfamiliar faces. Personally familiar faces evoke a stronger activation in the extended system as compared to faces of strangers. This increased activation of the extended system may reflect aspects of familiarity that are shared by all familiar identities or enhanced representations that distinguish individual familiar identities. Here we distinguished between these two aspects of familiarity, and investigated the organization of the core and extended systems derived from similarities in neural representations. We applied multivariate pattern classification and network-level representational similarity analysis to fMRI data collected while 33 subjects viewed personally familiar faces and unfamiliar faces. Both face identity and identity-independent familiarity information could be decoded from patterns of response to faces in extended system areas. Analysis of the representational similarities in the face processing system revealed a clear distinction between core and extended systems, and a subdivision of the core system into ventral, dorsal and anterior components. This study provides evidence that patterns of activity in the extended system carry information about both individual identities and personal familiarity, while clarifying and extending the organization of the core system for face perception.
