Abstract
Functional imaging studies have consistently documented face-selective activation in a core set of regions in human occipitotemporal cortex, particularly in the fusiform gyrus (FFA). However, the role of the FFA in face processing is still controversial and evidence supporting its exact contribution to face identification is not clear-cut. Here, we examine this issue by exploring whether there is a differential sensitivity for familiar vs. unfamiliar faces in the FFA and other cortical regions. We then determine whether face representation within regions showing such sensitivity is based on precise facial geometry or whether it is more abstract and based on facial identity irrespective of the exact physical appearance. To do so, we used a rapid event-related fMRI adaptation paradigm in which trials contained two consecutive faces for same/different identity judgment. ‘Same’ trials included two identical pictures or two different pictures of the same individual whereas ‘different’ trials included two pictures of two different individuals. Better performance (RT and accuracy) was found for familiar than unfamiliar faces, with poorest responses for different pictures of the same unfamiliar individual. fMRI findings revealed a differential response for familiar vs. unfamiliar faces in right (but not left) FFA, reflected in an overall elevated signal for familiar faces and a stronger adaptation effect when the exact same face was repeated. Interestingly, this region showed no adaptation for the identity of familiar faces but exhibited a signal enhancement when the identity of an unfamiliar face was repeated. Additional foci of activation in the left posterior cingulate and medial prefrontal gyrus also showed preferential response for familiar faces but each exhibited a different adaptation pattern. Thus, our results show that a network of cortical regions, each contributing a unique pattern of activation, is involved in familiarity and identity coding of faces.