Neuropsychologically derived errors as predictors of neurocognitive phenotypes: The Framingham Heart Study. (20th December 2022)

• Record Type: Journal Article
• Authors: De Anda‐Duran, Ileana; Drabick, Deborah AG; Hwang, Phillip H; Emrani, Sheina; Matusz, Emily; Wasserman, Victor; Swenson, Rodney; Andersen, Stacy L.; Liu, Chunyu; Au, Rhoda; Libon, David
• Is Part Of: Alzheimer's & dementia. Volume 18(2022)Supplement 11
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-12-20
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.066449
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Data‐driven approaches are powerful for identifying impaired neurocognitive constructs and aid in characterizing neurocognitive impairment patterns. There is, however, a significant clinical need for time‐efficient diagnostic tools compared to traditional more time‐consuming neuropsychological (NP) assessments. We evaluated how NP errors can compress the number of tests needed to differentiate among neurocognitive phenotypes. Method: We studied 1, 915 Framingham Heart Study (FHS) Offspring participants (age: 70 ± 8 years, 56% women). Three neurocognitive phenotypes were identified by Latent Profile Analysis (LPA) using nine NP summary scores: Cognitively normal (CN; n = 1060, 88.7%); Moderate‐Mixed (n = 72, 6.0%); and Dysexecutive (n = 63, 5.3%). Receiver operating characteristic (ROC) analysis compared the predictive accuracy of NP errors in identifying the distinct neurocognitive phenotypes. Regression models evaluated associations between significant NP errors (i.e., with an area under the curve (AUC) ≥ 0.64) and neurocognitive phenotypes, adjusting for age, sex, and education. The CN group was used as reference for comparisons. Result: Dysexecutive subgroup participants were older (77 ± 8 years) with less education (8 ± 3 years). Individuals producing any of the errors displayed in Figure 1 were 1.3 times more likely to be in the moderate/mixed and dysexecutive group [OR‐range = 1.3 ‐ 3.4, all p's < 0.001]. Participants generating Trail Making Test‐Part B (TMT‐B) examiner‐corrected non‐perceptual errors were 3 times more likely to be classified into the dysexecutive [OR = 3.44, 95% CI (2.7, 4.4), p < 0.001] and moderate/mixed phenotypes [OR = 3.21, 95% CI (2.13, 4.85), p < 0.001]. Participants were more likely to have a moderate/mixed phenotype with Boston Naming Test (BNT) perceptual [OR = 2.1, 95% CI (1.7, 2.5), p < 0.001] and semantic errors [OR = 1.6, 95% CI (1.4, 21.9), p < 0.001]. Conclusion: An analysis of errors derived from the TMT‐B and BNT provided evidence for the presence of two distinct but complementary and impaired neurocognitive constructs: dysexecutive/ inhibition impairment (TMT‐B) and struggling to access lexical information (BNT). Since cognitive screeners (e.g., MoCA) incorporate briefer versions of these tests, these errors may aid in characterizing brain‐behavior relationships and identifying persons at risk for the emergence of dementia‐related syndromes in a less costly and burdensome manner than standard NP protocols.