Current models suggest that face processing is accomplished by distinct visual pathways. Each pathway specializes in a particular aspect of facial information, such as emotion or identity, and has neural substrates that are at least partially separable from those of other pathways. Our research examines the pathway subserving face identification, which is traversed to identify a face by name. The face identification pathway is believed to occur in sequential stages, consisting of at least five stages, including face detection, face recognition, and face name identification.
We integrated magnetoencephalographic (MEG) and behavioral research to examine the staged nature of the face identification pathway. Specifically, we used Gaussian blur on celebrity face, non-celebrity face, and building images to produce three main findings. First, we gauged where a known face-processing MEG signal, the M170, is positioned within the face identification pathway. Second, we examined the effect of familiarity on the M170, and used this result to propose a functional description of the M170 and other stages that may lie between face detection and recognition. Finally, we gained insight into the relationship between face identification and the visual processing of non-face objects. To explain these findings, we incorporate a pragmatic view of what MEG can tell us about the face identification pathway and other neural processes, and we present the findings within the cohesive framework of the staged model of face identification.