Capturing social and cognitive health dynamics as digital risk biomarkers of Alzheimer's disease

About This Grant

Poor social health (high loneliness, high social isolation, low social support) is associated with heightened risk for Alzheimer's disease and related dementias (AD/ADRDs). Theoretical models posit that social health's long-term effects on cognition result from cumulative risks associated with short-term, dynamic social processes. For example, loneliness may increase AD/ADRD risk by reducing healthy behaviors (physical activity) and increasing unhealthy behaviors (alcohol consumption), impacting momentary cognitive performance and variability – which, themselves, predict future AD/ADRD diagnosis. Digital biomarkers that capture short-term, dynamic associations between social and cognitive health hold promise to detect early clinical signatures of cognitive decline and identify personalized targets for AD/ADRD risk-modifying behavioral interventions. Three major gaps obfuscate mechanisms linking social and cognitive health, stalling digital biomarker development. First, there is a conceptual gap: theoretical models highlight the cumulative impact of daily social experiences in AD/ADRD-related cognitive decline, yet empirical research relies largely on cross-sectional or limited longitudinal designs ill-suited to capture this granularity. Second, there is a measurement gap: ecologically valid tools differentiating the ABCs of social health have yet to be validated for ecological momentary assessment (EMA), hampering interpretation of dynamic, within-person associations. Finally, there is a computational gap: AD/ADRDs are highly heterogeneous, yet current analytic approaches do not adequately accommodate individual differences in AD/ADRD pathogenesis. This project is strategically poised to address these gaps through three aims. We will (1) develop mobile assessments of social health; (2) delineate short-term cognitive impacts of daily social dynamics; and (3) generate individualized models of social determinants of AD/ADRD risk. We hypothesize that, within individuals, cognition will be reduced when social health is low. We further expect to identify social and behavioral mediators of within-person associations between social health and cognition, as well as multiple combinations of social risk factors that predict AD/ADRD risk, consistent with multiple pathways to AD/ADRD. We will test hypotheses in a sample of N=250 adults (45-74 years) who provide intensive longitudinal social and cognitive EMA data. Successful completion of study aims significantly advances measurement, mechanistic understanding, and early risk assessment in AD/ADRDs, with implications for reducing the burden of AD/ADRDs through behavioral intervention. Successful completion of study aims is innovative because it provides the field with novel, open-source EMA scales optimized to measure fluctuations in multiple dimensions of social health. Race and ethnicity will be representative of 2060 United States population projections, ensuring that products of our work–including open-source scales and dynamic modeling insights–will support accurate stratification of midlife risk as population demographics shift.