Innovative Research will Use AI to Understand Dementia's Linguistic Footprint in Latin America

Atlantic Fellows for Equity in Brain Health Adolfo Garcia and Agustín Ibáñez discuss the potential of advanced technologies to revolutionize dementia care in Latin America.

Medical doctor attending a patient - stock photo

Latin America faces an imminent and daunting challenge: The prevalence of dementia is projected to surge by 180-250% by 2050. Access to a timely and appropriate health system remains a significant hurdle, with an estimated 70-90% of patients in the region left underserved. However, amidst this bleak landscape, advanced technologies, particularly artificial intelligence and natural speech analysis, stand poised to revolutionize dementia care in the region.

Throughout history, language has been a vital ally for Latinos. It has enabled us to unite across national boundaries, maintain our cultural identity, fight for social justice, and disseminate unparalleled fiction and lore. In the era of artificial intelligence, language may offer Latinos yet another powerful tool to favor equity in the assessment of dementia.

The Looming Dementia Crisis in Latin America

There are various types of dementia, but Alzheimer’s and frontotemporal dementia (FTD) are among the most common and challenging. Think of them as the primary culprits causing the most disruptions. Beyond the individual, these conditions weigh heavily on families emotionally and financially. Here's a crucial point: Latinos, more susceptible to these conditions, often do not receive the top-notch care they need. Highlighting and addressing this discrepancy is essential to ensure everyone gets proper support.

This scenario reveals a need for scalable, low-cost approaches to complement standard testing. Automated speech and language assessments (ASLA) simply require participants to speak, producing acoustic and linguistic features that can be processed through digital technologies. Such features can be used to characterize disorders and predict symptom severity and brain atrophy. Yet, these findings come mainly from high-income countries, and they are minimal in Latinos, whose social and biological particularities can significantly affect the course of the disease.

Leveraging Technology for Dementia Assessment

In response to this challenge, an interdisciplinary group of researchers has developed a large ASLA study involving over 2500 participants from Latin America and the United States. The project, part of the Multi-partner consortium to expand dementia research in Latina America (ReDLat), is titled “An automated machine learning approach to language changes in Alzheimer’s disease and frontotemporal dementia across Latino and English-speaking populations” and it is funded by the National Institute On Aging of the National Institutes of Health (R01AG075775).

The initiative will combine a state-of-the-art ASLA app (called TELL) with machine and deep learning tools to (a) test the diagnostic use 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.

We hope that the project will help revolutionize the diagnosis and treatment of dementia in Latin America, focusing on several key areas. First, we aim to develop and validate a diagnostic framework using ASLA tools tailored to Latin American populations. This initiative leans heavily towards equity, as it seeks to address the existing disparities in medical care in the region, allowing for early diagnoses and effective monitoring of dementia in the Latino community. Additionally, the technological innovations of the project, which combine ASLA with machine learning, are expected to lead to significant reductions in both costs and testing times. These tools are designed to be adaptable, considering the rich linguistic and cultural diversity of Latin America. Likewise, with this project, we aim to promote early interventions and specific therapies in the future, based on accurate diagnoses, strengthening collaboration between institutions and experts from different countries. Lastly, with the inclusion of thousands of participants, we hope to significantly expand knowledge about dementia in Latino populations.

Meet the Leaders Behind the Research

The effort is led by Adolfo M. García (Cognitive Neuroscience Center at Universidad de San Andrés, Argentina; Global Brain Health Institute, University of California, San Francisco, USA; Universidad de Santiago de Chile, Chile; Include Network); Maria Luisa Gorno-Tempini (Memory and Aging Center, University of California, San Francisco, USA), and Agustín Ibáñez (Latin American Brain Health Institute- BrainLat at Universidad Adolfo Ibanez, Chile; and the Precitive Brain Health Modellig Group at Global Brain Health Institute, Trinity College Dublin, Ireland).

The project also involves a stellar team of site Investigators including Drs. Diana Matallana (Pontificia Universidad Javeriana, Colombia), Francisco Lopera (Grupo de Neurociencias de Antioquia, Universidad de Antioquia, Colombia), Claudia Durán-Aniotz (BrainLat Institute, Chile), Nilton Custodio (Instituto Peruano de Neurociencias, Peru), and José Alberto Ávila Funes (Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico), with additional support from leading regional researchers, such as Maria Isabel Behrens (Centro de Investigación Clínica Avanzada - CICA, Chile), Andrea Slachevsky (Facultad de Medicina – Universidad de Chile, Chile), and Martin Bruno (Universidad Católica de Cuyo, Argentina).