Are faces special in Alzheimer's disease? Cognitive conceptualisation, neural correlates, and diagnostic relevance of impaired memory for faces and names.

Memory for faces and names has increasingly become a focus of cognitive assessment and research in Alzheimer's disease (AD). This paper reviews evidence from cognitive and clinical neuroscience regarding the question of whether AD is associated with a specific deficit in face recognition, face-name association, and retrieval of semantic information and names. Cognitive approaches conceptualizing face recognition and face-name association have revealed that, compared to other types of visual stimuli, faces are "special" because of their complexity and high intraclass similarity, and because their association with proper names is arbitrary and unique. Neuroimaging has revealed that due to this particular status, face perception requires a complex interplay of highly specialized secondary visual areas located in the occipitotemporal cortex with a widely distributed system of cortical areas subserving further task-dependent processing. Our review of clinical research suggests that AD-related deficits in face recognition are primarily due to mnestic rather than perceptual deficits. Memory for previously studied or famous faces is closely related to mediotemporal and temporocortical brain regions subserving episodic and semantic memory in general, suggesting that AD-related impairments in this domain are due to neural degeneration in these areas. Despite limited specificity due to the apparent absence of a "genuine" domain-specific deficit of face memory in AD, testing memory for faces and names is useful in clinical contexts, as it provides highly sensitive indices of episodic and semantic memory performance. Therefore, clinical assessment of face memory can usefully contribute to early detection of memory deficits in prodromal and initial stages of AD, and represents a basis for further attempts at rehabilitation. Further advantages, such as ecological validity, high task comprehensibility and, in the case of novel face learning, independence from premorbid intelligence level, render measures of face recognition valuable for clinical assessment in early AD.