Abstract
This article examines the functions performed by each of the “core” face processing regions: the fusiform face area, occipital face area, and superior temporal sulcus. It reviews the data from two complementary sources: functional imaging in healthy subjects and behavioral data from neurological subjects with damage to these regions. Data from functional neuroimaging allows for the determination of which regions are specifically engaged by which stimuli and which tasks; functional magnetic resonance imaging adaptation and pattern classification methods even allow for the characterisation of the representations extracted in each. However, fMRI can only tell that a given region is sensitive to a particular stimulus property. fMRI cannot depict if the processing requires perceiving that property occurs in that area. In contrast, lesion data can answer the crucial question of whether a given brain region is necessary for a given computation.