Abstract
Modularity of face processing is still a controversial issue. Congenital prosopagnosia (cPA), a selective and lifelong impairment in familiar face recognition without evidence of an acquired cerebral lesion, offers a unique opportunity to support this fundamental hypothesis. However, in spite of the pronounced behavioural impairment, identification of a functionally relevant neural alteration in congenital prosopagnosia by electrophysiogical methods has not been achieved so far. Here we show that persons with congenital prosopagnosia can be distinguished as a group from unimpaired persons using magnetoencephalography. Early face-selective MEG-responses in the range of 140 to 200ms (the M170) showed prolonged latency and decreased amplitude whereas responses to another category (houses) were indistinguishable between subjects with congenital prosopagnosia and unimpaired controls. Latency and amplitude of face-selective EEG responses (the N170) which were simultaneously recorded were statistically indistinguishable between subjects with cPA and healthy controls which resolves heterogeneous and partly conflicting results from existing studies. The complementary analysis of categorical differences (evoked activity to faces minus evoked activity to houses) revealed that the early part of the 170ms response to faces is altered in subjects with cPA. This finding can be adequately explained in a common framework of holistic and part-based face processing. Whereas a significant brain-behaviour correlation of face recognition performance and the size of the M170 amplitude is found in controls a corresponding correlation is not seen in subjects with cPA. This indicates functional relevance of the alteration found for the 170ms response to faces in cPA and pinpoints the impairment of face processing to early perceptual stages.
Highlights
We spend a considerable amount of our lives looking at and scrutinizing human faces
Evidence from brain imaging with selective activations in occipitotemporal cortex, namely the fusiform gyrus [1] and various case reports of acquired prosopagnosia, a deficit of familiar face recognition, e.g. caused by ischemic infarction are taken as evidence for a face selective module [2, 3]
We explored face processing in 16 controls and 13 persons with congenital prosopagnosia (cPA)
Summary
We spend a considerable amount of our lives looking at and scrutinizing human faces. This is not surprising because at least in vision these stimuli are of outstanding importance for successful social communication. For major cognitive functions like face processing it is of prime importance to identify and explicitly describe those computational steps which are realized by specialized neuronal ensembles (modules) This will deliver deep insight into cerebral function and biologically inspire technical applications which are still far off the efficiency of real brains. Evidence from brain imaging with selective activations in occipitotemporal cortex, namely the fusiform gyrus [1] and various case reports of acquired prosopagnosia (aPA), a deficit of familiar face recognition, e.g. caused by ischemic infarction are taken as evidence for a face selective module [2, 3] Opponents of this view claim that faces are exceptional due to massive overlearning because this leads to activations in the same locations for artificial objects [4]. In many instances acquired prosopagnosia is accompanied by general deficits of object recognition [5]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
