Neural basis of congenital face blindness

Prosopagnosia is a neurological condition characterised by an inability to recognize faces. In the most extreme cases, the prosopagnosic patient cannot even recognize their own face in the mirror or a photograph, and in his 1985 book The Man Who Mistook His Wife For A Hat, the neurologist Oliver Sacks describes the extraordinary case of a farmer who lost the ability to recognize the faces of his cows!

Also known as face blindness, prosopagnosia is associated with damage to specific parts of the temporal lobes. But there are also documented cases of patients who have the condition in the absence of brain damage. The form of face blindness is congenital: those who inherit a genetic mutation are born with an impaired ability to recognize faces.

A new study, published online in the journal Nature Neuroscience, now provides the first evidence of a neurobiological substrate for congenital prosopagnosia. The study, which was led by Cibu Thomas of Carnegie Mellon University, shows that the condition is associated with reduced connectivity between the regions of the brain involved in face processing.

Face perception is mediated by a diffuse network of interacting regions that are distributed widely throughout the brain.  Within this network is a "core"  system consisting of three discrete areas: the fusiform gyrus - also known as the fusiform face area - is selectively activated when viewing faces, and is thought by some to be a specialized face recognition module, although it is also activated by other categories of objects. Immediately adjacent to this is the inferior occipital sulcus; these two areas are together essential for recognition of individuals from their faces. The third  core component core is the superior temporal sulcus, which is involved in the perception of facial expressions and the direction of gaze.

This core system forms extensive interconnections with an extended network of regions which are involved in other aspects of the face recognition process, and which extract meaning from faces. The extended network includes the amygdala and the insula, both of which process the emotions conveyed in facial expressions, especially anger, disgust and fear; the hippocampus, which is involved in memory for faces; and the frontal cortex and nucleus accumbens, which are both believed to be involved in assessing the attractiveness of a face.

Thomas and his colleagues used an imaging technique called diffusion tensor imaging (DTI), which detects a signal generated by the movements of water molecules. Because water diffuses readily along the length of nerve fibres, DTI is particularly useful for "tractography" - visualizing the brain's white matter tracts, which contain bundles of axons projecting from one region of the brain to another.

Recently, DTI has been used to investigate connectivity in the brains of individuals with synaesthesia, in which the senses are merged, so that a stimulus in one sensory modality elicits sensations in another modality. For example, some synaesthetes experience colours when they hear or read words; in others, sounds may be associated with particular tastes or smells. DTI revealed that there is an increase in the connectivity between the sensory pathways. Thus, in synaesthetes, acitivty in one sensory pathway increases activity in the another, and this cross-talk between sensory modalities produces the synaesthetic experience.


Diffusion tensor imaging (DTI) tractography reveals a reduction in the volume of the inferior longitudinal fasciculus in the brains of 6 patients with congenital prosopagnosia (top).

(From Thomas et al 2008)

One possible explanation for the face perception deficits is abnormal functioning in the core areas. However, previous studies in which brain activity was analysed using functional neuroimaging, magnetoencephalography and event-related potentials failed to show any such abnormalities. Thomas and his colleagues therefore performed DTI tractography on 6 prosopagnosic patients and 17 controls, in order to investigate the connections between those parts of the brain involved in face perception.

They focused on two white matter tracts, called the inferior longitudinal fasciculus (ILF) and the inferior fronto-occipital fasciculus (IFOF), which connect the core face recognition system to the areas of the extended network in the temporal and frontal lobes, respectively. The data they collected revealed that there was indeed a significant difference between the two groups: compared to the controls, the prosopagnostics showed a marked reduction in the volume of both white matter tracts, on both sides of the brain.  

The researchers then investigated the consequences of these structural differences. All the participants were presented with a series of photographs, at random and for unlimited amounts of time. Half of the photographs consisted of familiar famous faces such as George Clooney and Elvis Presley, while the other half were of unfamiliar famous faces, such as famous actors or celebrities from other countries.

As expected, it was found that the prosopagnosics performed significantly worse on this task than the controls. Poor performance was strongly and negatively correlated with a reduction in the volume of ILF fibres in the right hemisphere, and somewhat less so with IFOF fibres - the greater the reduction in white matter tract volume, the greater was the number of errors on the face recognition task.

The results of this study are consistent with previous findings that the right hemisphere plays a more prominent role than the left in face processing. They are also in line with studies which show that prosopagnosia can occur following damage to the ILF as a result of multiple sclerosis. That there was so much variation in the performance of the prosopagnosics on the face recognition task lends further support to the idea that the condition varies in severity, such that some patients retain some ability to recognize faces, while others are completely unable to do so.

Finally, these findings point to a genetic basis for these structural abnormalities; it is possible that a single mutation is responsible for improper maturation of the white matter tracts during brain development. However, the findings should first be confirmed using more direct imaging methods on larger numbers of participants.


Thomas, C. et al (2008). Reduced structural connectivity in ventral visual cortex in congenital prosopagnosia. Nat. Neurosci. DOI: 10.1038/nn.2224

Ishai, A. et al (2005). Face perception is mediated by a distributed cortical network Brain Res. Bull. 67: 87-93 (2005). [PDF]

Haxby, J. V. et al (2000). The distributed human neural system for face perception. Trends Cog. Sci. 4: 223-233. [PDF]

More like this

One possible explanation for the face perception deficits is abnormal functioning in the core areas. However, previous studies in which brain activity was analysed using functional neuroimaging, magnetoencephalography and event-related potentials failed to show any such abnormalities.

Not true. For example, see:
Functional Plasticity in Ventral Temporal Cortex following Cognitive Rehabilitation of a Congenital Prosopagnosic
Joseph M. DeGutis, Shlomo Bentin, Lynn C. Robertson and Mark D'Esposito.
Journal of Cognitive Neuroscience, 19: 1790-1802 (2007).

They not only show neural differences and connectivity differences with both fMRI and EEG, but they also show a way to train a congenital prosopagnosic to recognize faces and show that the neural markers change with behavioral changes.

Thanks bsci. According to the paper:

...studies have failed to definitively implicate the core regions as the source of the impairment; face-selective electrophysiological markers, derived from evoked response potentials or magnetoencephalography, have been shown to be normal in many congenital prosopagnosic individuals, albeit not in all.

So perhaps I should have said that these studies in general have not shown abnormal activity in these areas. However, bear in mind that the study you point to involved just one participant.

Sure it's one participant, but it's a pretty cool case study. Also, saying nothing has never been observed is fairly easily contradicted by showing a single observation.