Identifying Digital Phenotypes of Risk for Alzheimer's Disease and Related Dementias Among Hispanics/Latinos

点击翻译按钮获取中文摘要

PROJECT SUMMARY Hispanics/Latinos are at increased risk for developing Alzheimer’s disease and related dementias (ADRD) compared to non-Hispanic Whites. While factors underlying this disparity are not well understood, increased cardiovascular disease (CVD) risk among Hispanics/Latinos is likely to be a contributory factor. Pathological changes of ADRD begin years before clinical symptoms become evident and interventions are most likely to confer benefit in the earliest stages of ADRD. There is a pressing need to develop tools to detect the earliest manifestations of ADRD, particularly in Hispanics, who develop symptoms of ADRD at earlier ages, yet are diagnosed at more advanced disease stages than other groups. A variety of behaviors show changes in the preclinical stages of ADRD, including sleep, gait speed, and physical activity, among others. Recent innovations in mobile technology now offer novel ways to collect, track, and analyze these behaviors passively and unobtrusively, as a person engages in their daily life. Our preliminary work demonstrated that the application of machine learning models to passively-collected digital health data from smartphones and wearables differentiated persons with and without mild cognitive impairment with 85% accuracy in a primarily non-Hispanic White sample. Guided by the NIA Health Disparities Research Framework, we propose to leverage artificial intelligence (AI)-powered analytics and insights, coupled with readily available sensors in consumer electronics (smartphones, wrist-worn wearables), to identify digital biomarkers of ADRD risk, with a focus on vascular contributions to dementia, among Hispanics/Latinos. Augmenting an existing cohort study of Hispanics/Latinos residing in Southern California, the proposed study has three principal aims: Aim 1 involves determining digital signatures of ADRD risk among Hispanics (N=300; aged 50-70 years) using integrated passive mobile sensing features, derived from smartphones and wrist-worn wearables, and machine learning methods. ADRD risk will be defined by cognitive status and CVD risk burden (diabetes, hypercholesterolemia, hypertension, obesity, smoking) and will also incorporate apolipoprotein E (APOE-ε4) and plasma-based AD biomarkers for further classification of ADRD risk. Aim 2 investigates sex differences in digital signatures of ADRD risk and Aim 3 examines the impact of sociocultural factors (e.g., language use, acculturation) on these signatures. We will also investigate whether changes in digital data features predict longitudinal neurocognitive change over a span of three years in a subset of Hispanics with and without ADRD risk. Housed within a renowned research institution at the vanguard of ADRD research and engineering innovations, the proposed study includes a multidisciplinary team with expertise across all aspects of this cutting-edge proposal. Recognizing the value of a community- engaged research approach, we have partnered with community stakeholders to ensure the relevance of our study to the Hispanic community. Our work could revolutionize early detection of ADRD and reduce ADRD disparities by developing a low burden, low-cost approach to identify ADRD risk among Hispanics.