Repeated testing of plasma biomarkers and digital cognition revealed that phosphorylated tau 181 exhibited the poorest test-retest reliability among evaluated blood markers, serving as the only metric where collecting a repeat sample improved accuracy in discriminating between diagnostic groups.
In contrast to the blood marker variability, none of the digital cognitive measures demonstrated improvements in diagnostic accuracy when repeated across longitudinal sessions, indicating high baseline test-retest reliability for computerized assessment tools.
The investigation evaluated paired plasma biomarker collection and digital cognitive assessments at zero, three, and six months across a cohort of 92 individuals, comprising 55 cognitively unimpaired healthy controls and 37 patients diagnosed with early Alzheimer’s disease dementia.
Cognitive performance was quantified using the Oxford Cognitive Testing Portal digital platform, while the plasma panel incorporated phosphorylated tau 181, phosphorylated tau 217, neurofilament light chain, glial fibrillary acidic protein, and the amyloid-beta 42 to 40 ratio.
Individual metrics displayed distinct reliability profiles depending on whether variability was quantified through the coefficient of variation or the intraclass correlation coefficient, and varied within groups (healthy controls and early Alzheimer's disease dementia).
Researchers emphasized that establishing test-retest reliability across intermediate three- and six-month intervals is essential for accurately interpreting meaningful therapeutic effects over time in screening programs and clinical trials evaluating interventions for Alzheimer’s disease.
The emergence of both plasma biomarkers and computerized cognitive testing has established new avenues for widespread population screening and for tracking therapeutic efficacy during Alzheimer’s disease clinical trials. However, interpreting longitudinal outcomes requires a precise understanding of how consistently these assessment tools perform upon repeated administration. Prior scientific investigations have primarily documented biomarker fluctuations across either very brief intervals or across full annual periods, leaving the intermediate three-to-six-month window largely unexamined.
To resolve these uncertainties, investigators instituted a paired assessment protocol tracking both biological and computerized cognitive parameters longitudinally across baseline, three-month, and six-month intervals. The study cohort comprised ninety-two participants, separating fifty-five cognitively unimpaired healthy control individuals from thirty-seven patients diagnosed with early Alzheimer’s disease dementia. Blood evaluations captured five distinct plasma biomarkers: phosphorylated tau 217, phosphorylated tau 181, neurofilament light chain, glial fibrillary acidic protein, and the amyloid-beta 42 to 40 ratio. Alongside blood collections, participants completed computerized evaluations administered through the Oxford Cognitive Testing Portal, known as OCTAL, to evaluate cognitive performance across identical time points.
The research demonstrated that individual measurement stability profiles diverged considerably depending on the specific mathematical metric selected to quantify longitudinal variability. Specifically, reliability determinations differed when researchers calculated the coefficient of variation versus the intraclass correlation coefficient across consecutive assessments. In addition to metric-dependent variations, individual metrics varied within groups (healthy controls and early Alzheimer's disease dementia).
The comparative analysis revealed clear differences between biological and digital evaluation tools over the six-month study duration. Phosphorylated tau 181 demonstrated the poorest test-retest reliability among all tested plasma analytes, and it was the sole biomarker where obtaining a repeated measurement yielded higher diagnostic accuracy when separating healthy controls from early dementia patients. In contrast, computerized assessments recorded through the Oxford platform showed no improvement in diagnostic discrimination when repeated across subsequent visits. This stability confirms that digital cognitive measures maintain high baseline test-retest reliability without requiring duplicative sampling to achieve optimal diagnostic accuracy across intermediate clinical intervals.
Establishing robust test-retest reliability benchmarks across intermediate intervals is essential for the future implementation of disease-modifying therapies and screening protocols. The documented performance profiles across both plasma analytes and digital testing platforms provide necessary parameters for distinguishing true intervention responses over time.
Subject of Research: Test-retest reliability of plasma biomarkers and digital cognitive tests in Alzheimer's disease
Article Title: Test-retest reliability of plasma biomarkers and digital cognitive metrics at clinically meaningful timescales
Article References: Toniolo, S., Zhao, S., Broulidakis, M., Fower, A., Gendarini, C., Scholcz, A., Griffiths, T. J. C., Zetterberg, H., Thompson, S., Manohar, S., & Husain, M. (2026). Test-retest reliability of plasma biomarkers and digital cognitive metrics at clinically meaningful timescales. Alzheimer's Research & Therapy. https://doi.org/10.1186/s13195-026-02205-2
Image Credits: AI Generated
DOI: 10.1186/s13195-026-02205-2
Keywords: Alzheimer's disease, plasma biomarkers, digital cognitive testing, pTau181, test-retest reliability, OCTAL, clinical trials, Test-retest, reliability, plasma, biomarkers, digital
News Source: Ophelia Keating. (November 10, 2026). Plasma Biomarker and Digital Cognitive Test Reliability Evaluated Over Six Months. Scienmag.
