Detecting Alzheimer's disease using digital virtual reality cognitive tests

Area of Science:

• Neuroscience
• Biomarkers
• Digital Health

Background:

• Digital biomarkers for Alzheimer's disease (AD) detection face challenges with accuracy and correlation to core AD pathology.
• Existing methods for AD diagnosis can be invasive or costly.

Purpose of the Study:

• To assess the association between artificial intelligence (AI)-assisted virtual reality (VR) cognitive scores and established AD biomarkers.
• To evaluate the efficacy of a novel AI-VR cognitive test for identifying cognitive impairment and AD pathology.

Main Methods:

• Recruited 428 Chinese older adults for the study.
• Utilized generalized linear models to analyze associations between AI-VR cognitive scores and plasma biomarkers (Aβ42/Aβ40, p-tau217, GFAP, NfL), Aβ PET, tau PET, and neuroimaging markers.
• Controlled for covariates including age, sex, education, and apolipoprotein E ε4 status.

Main Results:

• AI-VR cognitive scores significantly correlated with key AD biomarkers including plasma Aβ42/Aβ40, p-tau217, GFAP, NfL, and PET imaging results.
• The AI-VR test demonstrated high accuracy in distinguishing cognitively impaired from unimpaired individuals (AUC: 0.908-0.969).
• The test also effectively classified individuals positive for amyloid-beta positron emission tomography (AUC: 0.771-0.991).

Conclusions:

• A 5-minute AI-VR cognitive test shows significant potential as a reliable digital marker for early AD detection.
• This digital tool can identify abnormal cognition and correlate with core AD biomarkers in a community setting.
• AI-VR cognitive testing offers a promising non-invasive approach for AD screening and diagnosis.