Project Summary/Abstract The proposed ancillary study looks to better understand racial/ethnic disparities in Alzheimer's disease (AD) by investigating spontaneous speech as part of the Washington Heights/Inwood Columbia Aging Project (WHICAP) Offspring Study. Limitations of existing cognitive screening instruments is a known contributor to racial/ethnic disparities in AD and recent work suggests that speech analysis may be a sensitive marker of cognitive decline and future AD. New technological advances raise the possibility that automated speech analysis could be conducted using smartphones in community settings, creating an inexpensive and scalable approach to early detection. However, many challenges remain in developing this method, including a very limited understanding of the biological underpinnings of spontaneous speech in at-risk individuals (e.g., hippocampal volume, amyloid deposition, contribution of executive function vs. language). We propose to examine spontaneous speech as part of the ongoing WHICAP Offspring Study (RF1AG054070, RF1AG058067). This large project is enrolling a diverse cohort (>70% minorities) of pre-clinical participants with varying degrees of risk for AD using medical, genetic, and neuropsychological testing (N = 3000), structural MRI (N = 1000), and amyloid and tau PET imaging (N = 150 for each). Both English and Spanish speakers are included. One measure of spontaneous speech has been included since study onset and a second brief measure would be added. We hypothesize that spontaneous speech will reflect a combination of both biological and sociocultural risk factors to AD and that bilingualism will show protective effects. By leveraging the rich data collection of the WHICAP Offspring Study, this application would generate substantial new insight into the basic underpinnings of spontaneous speech in diverse pre-clinical samples, as well as determine the feasibility of using speech analysis as a novel biomarker of AD risk. Additionally, should the cohort be followed longitudinally, the ability of spontaneous speech to predict future incidence of AD in racially/ethnically diverse samples over time could also be evaluated.