TITLE:
Semantic Intrusion Errors Overcome the Impact of Literacy and Educational Attainment in Black/African Americans with and without Mild Cognitive Impairment
AUTHORS:
Alexandra Ortega, Kirsten Horne Crenshaw, Triana Abel, Carlos Valentín-Camuñas, Stephanie Remedios, Brooke Bosworth, Sofia Ramirez, Dylan Hinton, Aleksandra Nedza, Elizabeth A. Crocco, Rosie E. Curiel-Cid, David Loewenstein
KEYWORDS:
Mild Cognitive Impairment, Semantic Intrusions, Proactive Interference, Literacy, African Americans
JOURNAL NAME:
Advances in Alzheimer's Disease,
Vol.14 No.4,
November
10,
2025
ABSTRACT: The Loewenstein-Acevedo Scale for Semantic Interference and Learning (LASSI-L) is a novel test of learning and inability to recover from proactive interference. It has been successfully employed with older adults at risk for cognitive impairment but the association between literacy, educational achievement, and LASSI-L performance in underserved Black/African American (B/AA) older adults remains unknown. 83 participants were evaluated as part of a longitudinal observational study to render a consensus diagnosis based on comprehensive clinical history and neuropsychological evaluation. Participants were classified into two groups: 43 were diagnosed with amnestic mild cognitive impairment (aMCI) and 40 were cognitively unimpaired (CU). The Wide Range Achievement Test-IV (WRAT-IV) was used to determine levels of literacy. Participants completed the LASSI-L, an AD sensitive cognitive challenge test which was not utilized for diagnosis to avoid circularity. The CU group outperformed the aMCI group across all LASSI-L measures after demographic covariate adjustment. Robust associations were found between education, literacy, and correct responses on different LASSI-L word lists (but not semantic intrusion errors). In contrast, aMCI participants largely exhibited no associations between LASSI-L indices and literacy effects. Stepwise logistic regression achieved an overall accuracy of 81% in classifying cognitive status (84% CU specificity, 71% aMCI sensitivity). Literacy was related to certain LASSI-L measures of learning in CU participants, but not among individuals with aMCI. Even when adjusting for demographic covariates, the LASSI-L effectively distinguished B/AA CU and aMCI groups. Results underscore the practicality of the LASSI-L to assess cognitive impairment in individuals with lower literacy proficiency, as our model correctly distinguish 84% of cognitively unimpaired individuals and 71% of individuals with amnestic mild cognitive impairment.