Semantic memory deficits have been reported in both Alzheimer's disease (AD) and amnestic mild cognitive impairment (aMCI). However, the nature of this decline is still a matter of debate. The aim of this study was to explore the patterns of semantic memory impairment in aMCI by examining performance on naming tasks, and on tests assessing both general and specific semantic knowledge.

Participants were divided in two groups matched for age and education, one comprising 33 aMCI individuals and the other 39 healthy controls. Three experimental tests assessing naming and semantic knowledge of unique items of famous persons (FACE) and places (PLACE), logos recognition (LOGO: brands and pictograms), and non-unique entities (Boston Naming Test: BNT) were administered, and the performance of the two groups was compared.

Lower scores were observed on all naming tests (PLACE, FACE, LOGO and BNT) in the aMCI group compared to controls. On the PLACE test, the general knowledge mean score (M=84.5, SD=12.9) was significantly higher than the specific knowledge mean score (M=54.2, SD=18.5) in aMCI participants (t(31)=11.9, p<.001), but not in controls (general: M=92.2, SD=11.1; specific: M=73.7, SD=15.8), and there was a significant Group X Type of knowledge interaction (F(1,1)=15.13, p <.001, n2 = 18). On the FACE test, in addition to significant group and condition (naming, semantic questions) main effects, a significant interaction was found (F(1,1)=7.19, p = .009, n2 = .09). On the LOGO task, controls were significantly better on brand items (M= 94.4, SD=10.5) than on pictograms (M=83.3, SD=12.2), while no significant difference was noted in aMCI (brands: M=81.5, SD=22.6; pictograms: M=77.5, SD=14.1). Lastly, on the BNT, aMCI participants benefited more from phonemic cues than controls (F(1,1)=16.56, p<.001, n2=19), suggesting a lexical access deficit, in addition to their semantic memory impairment.

This study adds to the growing evidence confirming the presence of semantic memory deficits in aMCI. Specific semantic knowledge seems to be more affected than general semantic knowledge, a finding reported in previous studies. Lexical access deficits, in addition to semantic decline, were also observed in the aMCI group. These results allow for a better understanding of the pattern of semantic memory deficits in the prodromal stage of AD and could potentially facilitate diagnosis of aMCI.