The Combination of Neuropsychiatric Symptoms and Blood-Based Biomarkers Enhances Early Detection of Mild Cognitive Impairment
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Combing behavioral assessments with blood-based biomarkers (BBM) may improve detection of Mild Cognitive Impairment (MCI) linked to early-stage neurodegenerative disease. Neuropsychiatric symptoms (NPS) often precede or accompany cognitive decline and provide observable behavioral signals, while BBM reflect underlying neuropathological changes. We investigated if integrating biological (plasma biomarkers) and behavioral (NPS) measures improves differentiation of MCI from cognitively normal (CN) individuals in a multi-ethnic Southeast Asian cohort— an underrepresented population in dementia research.
Methods
This cross-sectional analysis included 678 community-dwelling adults (mean age 59.2±11.0, 60.5% female) from the Biomarkers and Cognition Study, Singapore (BIOCIS), comprising participants recruited from the community, at Dementia Research Centre (Singapore) from February 2022 to March 2024. Participants underwent behavioral assessments using the Mild Behavioral Impairment Checklist (MBI-C) and the Depression, Anxiety, and Stress Scales (DASS). Plasma biomarkers measured were amyloid-beta (Aβ40, Aβ42), phosphorylated tau (p-tau181), neurofilament light (NfL), and glial fibrillary acidic protein (GFAP). Logistic regression and receiver operating characteristic (ROC) analyses evaluated the discriminative power of NPS, BBM, and their combination for identifying MCI risk.
Results
MBI-C total scores and subdomains (Mood, Interest, Control) and plasma biomarkers (Aβ40, NfL, GFAP) were significantly elevated in MCI compared to CN participants. Multivariate analysis showed elevated plasma GFAP (OR=3.64, 95% CI:1.96–6.75, p<0.001) and higher MBI-C Mood scores (OR=2.61, 95% CI:1.54–4.44, p<0.001) as the variables most associated with MCI. The combined model integrating NPS and BBM achieved a higher discriminative ability (AUC = 0.786) for MCI than models using NPS (AUC = 0.593) or BBM (AUC = 0.697) alone. The integrated model yielded 64.7% sensitivity and 84.9% specificity for distinguishing MCI from CN, outperforming single-domain approaches.
Conclusions
Integrating biological and behavioral markers improves identification of individuals with early cognitive impairment. Notably, GFAP-driven neuroinflammation and mood disturbances emerged as key features of prodromal dementia, highlighting astrocytic activation and affective changes as promising biomarkers and early intervention targets. This dual-domain, multimodal framework offers translational potential for earlier detection, risk stratification, and timely intervention for Alzheimer’s disease and other dementias.
