An automated machine learning approach to language changes in Alzheimer’s disease and frontotemporal dementia across Latino and English-speaking populations

This project uses speech and language analysis (ASLA) as an objective, low-cost approach for dementia detection and monitoring in Latinos, a large and underserved minority. Leveraging a large cohort from Latin America and the United States (n = 2740), we will employ machine and deep learning to (a) test the diagnostic utility of ASLA markers; (b) correlate them with cognitive and neuroimaging features; and (c) identify those that are robust across, languages, dialects, and socio-biological variables. In the long term, this research will provide equitable tools for early diagnosis and monitoring of dementia in Latinos, reducing testing cost and time, avoiding biases of examiner-based tests, differentiating syndromes beyond common-cause confounds, and enabling timely adoption of neuroprotective life changes and pathology-targeted therapies.