Examining the diagnostic utility of the mnemonic discrimination task for classification of cognition and amyloid‐β burden. (20th December 2022)

• Record Type: Journal Article
• Authors: Kim, Soyun; Chappel‐Farley, Miranda G.; Keator, David B.; Janecek, John; Mikhail, Abanoub; Hollearn, Martina K; McMillan, Liv; Yassa, Michael A.
• Is Part Of: Alzheimer's & dementia. Volume 18(2022)Supplement 5
• 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.060714
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Developing non‐invasive diagnostics that can effectively predict amyloid‐β (Aβ) burden is crucial, as this technology could help to identify individuals with increased risk for AD and provide a window for potential treatment or intervention. Mnemonic discrimination tasks (MDTs) are designed to assess the ability to discriminate highly similar episodic memory representations (e.g., objects or spatial locations). The relationship between the MDT performance and Aβ burden, however, has not yet been directly tested. In this study, we investigated whether MDT performance can successfully classify cognitive status or Aβ load in older adults. Method: A sample of 85 older adults (60‐91 yrs old) with normal cognition and 9 older adults with a clinical diagnosis of Mild Cognitive Impairment (MCI) (62 ‐89 yrs old) underwent either cerebrospinal fluid (CSF) lumbar punctures or 18 F‐florbetapir‐PET imaging. Cognitively normal individuals were stratified into Aβ+ and Aβ‐ subgroups based on quantitative thresholds for Aβ positivity (SUVR >1.11, CSF Ab1‐42/1‐40 < 0.062). Participants performed the object and spatial versions of the MDT. We used a random forest classification algorithm, including MDT performance metrics and demographic factors (age, gender, education) as features to classify cognitive status (normal vs. MCI) and Aβ status (positive vs. negative). We evaluated the area under the receiver operating characteristic curve (AUC) using the classifiers in a leave‐one‐out cross validation design. Result: For classification of cognitive status, the best model performance was achieved by including all the features of object and spatial MDT metrics and demographics (AUC = 0.80). For classification of Aβ status in cognitively normal participants, the best model performance was also achieved by including all the features of object and spatial MDT metrics and demographics (AUC = 0.65). Conclusion: Our findings suggest that MDT performance measures may have clinical utility as a diagnostic tool to predict cognitive status and in vivo Aβ burden in preclinical older adults.