Semantic and pre-semantic factors affecting face discrimination and recognition in Alzheimer's disease

Face recognition is subject to severe impairment in Alzheimer's disease (AD). According to Bruce and Young's model (1986), face recognition includes pre-semantic, semantic, and output stages. The present study used explicit memory tasks to establish lexical output abilities and an implicit memory task to assess semantic abilities, and it examined spatial frequency contrast sensitivity as a possible pre-semantic source of face discrimination deficits. Subjects included individuals diagnosed with probable AD (n = 27) and age-matched control participants (n = 21).; The new Implicit Face Recognition Test tested familiarity of famous faces with four conditions of relatedness (highly-associated, categorically-related, neutral, and unrelated) and examined the effects of cuing on face recognition in AD. Assessment of reaction time was the measure of the facilitation benefit of semantic relatedness (priming). Performance on the priming task was contrasted with that on measures of explicit memory (recall and recognition) of the same faces administered in a separate session.; AD patients performed poorly on the explicit memory tasks, as expected. By contrast, facilitation as a result of categorical relatedness was evidenced in both the AD and control groups, suggesting that their explicit face recognition deficits were the result of a retrieval problem rather than a breakdown of semantic knowledge. Additionally, poor performance on face recognition tasks without cues may indicate a breakdown in pre-semantic processing.; One pre-semantic structural process that is impaired in AD is contrast sensitivity, particularly at low spatial frequencies. The AD group showed face discrimination performance that was equal to that of the control group when sensitivity to low, but not higher, spatial frequency information was enhanced through manipulation of face size.; These results indicate that patients with AD are able to use their preserved semantic knowledge (stage 3 of Bruce and Young's model), presumably supported by lateral temporal structures, to benefit from associative cues and thereby enhance face recognition abilities. Additionally they demonstrate deficits in processing pre-semantic input (stage 2), specifically spatial frequency contrast sensitivity information contained in structural codes, and lexical output (stage 4).