Abstract

Face recognition confronts the same fundamental challenge as all object recognition: the identification of individuals belonging to a specific category despite a huge range in possible appearances. The specialized architecture of the face patch system—the concentration of face cells into modules and the spatial separation of modules—makes it possible to dissect the steps leading to invariant object recognition in unprecedented detail. It discusses both anatomical experiments revealing the connectivity of the six face patches and electrophysiological experiments revealing the response selectivity of single units within the patches. It reviews insights into this question obtained from fMRI and electrophysiological experiments in macaque monkeys. It presents evidence that macaques have specialized face regions and discusses their anatomical connectivity as well as functional properties measured with both fMRI and targeted electrophysiology.