Predicting CDR status over 36 months with a recall‐based digital cognitive biomarker

• DOI: 10.1002/alz.14213
• Publication Date: 2024-09-11T13:02:18Z
• Authors: Davide Bruno; Ainara Jauregi‐Zi...; Jason R. Bock

Word-list recall tests are routinely used for cognitive assessment, and process scoring may improve their accuracy. We examined whether Alzheimer's Disease Assessment Scale-Cognitive subscale (ADAS-Cog) derived, process-based digital biomarkers (DCBs) at baseline predicted Clinical Dementia Rating (CDR) longitudinally compared them to standard metrics. Analyses were performed with Neuroimaging Initiative (ADNI) data from 330 participants (mean age = 71.4 ± 7.2). conducted regression analyses predicting CDR 36 months, controlling demographics genetic risk, ADAS-Cog traditional scores DCBs as predictors. The best predictor of months was M, a DCB reflecting ability (area under the curve 0.84), outperforming scores. Diagnostic results suggest that M be particularly useful identify individuals who unlikely decline. These outperforms metrics supports word-list tests. More research is needed determine further applicability other populations. Process latent modeling more effective than scoring. Latent (M) decline months. top biomarker model had odds ≈ 90 times greater model. Particularly high negative predictive value literature on testing screen. Consideration both pathological outcomes needed.