PROJECT SUMMARY/ABSTRACT Alzheimer’s disease (AD) is responsible for 1 in 3 deaths among older adults, and methods of accurate, equitable, and widely available detection are needed. Digital biomarkers, such as the Visuospatial Memory Eye- Tracking Test (VisMET) show promise in their ability to discern subtle neuropathological changes in early AD and address limitations of existing detection techniques and provide unique opportunities to efficiently evaluate cognitive change via novel, passive mechanisms. This proposal aims to formally evaluate VisMET’s ability to detect subtle memory impairment in preclinical AD and its capacity to function as a multi-faceted digital indicator free from demographic bias. Successful completion of this proposal will be an important step toward our long- term goal of deploying VisMET clinically to diverse older adult populations to provide quick, accessible, and accurate detection of early pathological change in AD. Our general hypothesis is that VisMET will be shown to detect preclinical decline in visuospatial memory and will predict longitudinal cognitive decline in preclinical AD, and that this detection strength will not be impacted by race, sex, or education level. This study will recruit older adults with and without cognitive impairment from ongoing studies at Emory University in Atlanta, GA and from a diverse national cohort, the Bio-Hermes study. Aim 1 proposes to replicate past findings that VisMET performance is a strong predictor of global cognition on a widely recognized screening measure, the Montreal Cognitive Assessment (MoCA). These findings will then be extended to discern VisMET’s ability to predict preclinical AD status in healthy individuals (based on CSF biomarkers) and its ability to predict longitudinal decline in performance over a two-year interval. Aim 2 will rigorously evaluate the impact of demographic factors, including age, sex, race, and education, on VisMET performance. To achieve this, the relationship between task performance and personal factors will be evaluate in both a large, diverse sample from Emory and in a large, diverse sample from external national sites via the Bio-Hermes study. In Aim 3, additional metrics extracted via machine learning algorithms from VisMET and will be tested for their convergent validity with comparable neuropsychological measures. Successful completion of this proposal is expected to significantly impact detection of preclinical AD on a local and national scale by supplying key information about VisMET and its generalizability. Given its ability to address limitations of traditional neuropsychological measures, this proposal represents a critical step that will support broad deployment of VisMET in clinical and non-clinical settings. This proposal also stands to provide a model for extracting additional cognitive metrics from emerging digital indicators using advanced statistical techniques like machine learning algorithms. Critically, this researc...