Why Is It Just Faces I Can't Remember? A Look Inside the Brain

In everyday life, it might feel like recognizing a chair or a road sign uses the same visual machinery as recognizing a person's face. Research suggests otherwise, however: face processing is thought to involve a highly specialized set of brain mechanisms that differ, at least in part, from the processing of objects in general.

This is thought to be connected to how critical faces are as a source of information in daily life. Whose face it is, what expression they're wearing, where their gaze is directed — this kind of information is essential for building relationships with other people. One view holds that this is why the brain may have developed a network specialized for processing faces.

Among the regions involved in face processing, one that comes up especially often is the "fusiform gyrus." Located on the underside of the temporal lobe, roughly at the level below the ear, a portion of it known as the "fusiform face area" has been shown in numerous studies to become strongly active when a face is seen.

The fusiform gyrus is thought to be involved not in processing individual facial features — eyes, nose, mouth — separately, but in processing their arrangement and overall configuration all at once. When this process of "perceiving a face as a single whole" doesn't work well, the individual parts may still be visible, but working out whose face it is can become difficult.

Facial information isn't thought to be processed in just one location — rather, several regions spanning from the occipital lobe to the temporal lobe are believed to work together. The proposed flow is that face-related regions in the occipital lobe handle relatively early visual features, such as the arrangement of the eyes and nose, and that information is then passed on to areas like the fusiform gyrus in the temporal lobe, where it feeds into the more advanced task of working out whose face it is.

Information about changing expressions and shifts in gaze, meanwhile, is thought to be processed through a separate pathway, and some research points to the possibility that recognizing "who someone is" and reading "what expression they're making" proceed, to some extent, along separate tracks.

Face processing is thought to involve the brain's right hemisphere more strongly than the left. Case studies of acquired prosopagnosia have relatively often reported symptoms arising from damage centered on the right temporal and occipital lobes, and this is one piece of evidence cited in support of the right-hemisphere-dominance view.

That said, some reports describe involvement of both hemispheres, so it can't be said with certainty that this is purely a right-hemisphere issue. Exactly how the face-processing network is divided between the two hemispheres, and to what degree, remains an active area of research.

In developmental prosopagnosia, there's often no clear brain damage of the kind seen in acquired cases, leading to the idea that there may be some difference — present from birth, or arising over the course of development — in how the face-processing network, including the fusiform gyrus, functions or connects. Some brain-imaging studies have reported differences in how active the fusiform gyrus becomes when viewing faces, or in how strongly different regions are connected to one another, but not all of the causes or mechanisms have been worked out.