LSU Research Bites: AI Analysis of Everyday Speech Helps Detect Dementia Earlier
November 24, 2025
Early detection is one of the biggest challenges in dementia treatment and care.
By the time memory impairments are detectable by a patient or their loved ones, Alzheimer’s disease is typically advanced. Irreversible damage to brain cells makes treating or preventing the progression of this disease difficult.
For decades, clinicians have used memory and drawing tests to detect and diagnose dementia. They list out random words and ask patients to recall them later, or they ask patients to draw a clock to test their problem-solving, visuospatial, and attention skills.
However, the results of these tests aren’t always sensitive to early cognitive changes. They also aren’t reflective of daily life function, where being able to remember where things are in the kitchen while cooking dinner may be more pertinent. Clinicians need tools that can detect cognitive impairment earlier and that accurately reflect daily life function. LSU researchers, including Alex Cohen and PhD alum Ross Divers, leveraged artificial intelligence and subtle changes in speech to address this gap.
“AI really can be useful in detecting subtle behaviors that relate to cognition, memory, and thinking.” Divers said.
Divers and colleagues analyzed audio recordings from cognitive and memory tests administered to over 100 older adults. They used AI to look for patterns of speech that varied between healthy adults and those diagnosed with mild cognitive impairment or mild Alzheimer’s disease.
They found that longer pauses in speech, especially during memory tests, can reveal cognitive decline in its earlier phases, before other symptoms appear. This suggests that taking longer to remember something, even if ultimately being able to remember it, could be indicative of early mental changes associated with dementia.
“If we can detect disease in earlier stages, we can help people get targeted treatments earlier,” Divers said.
The project was funded by a CART grant and conducted in collaboration with Marymount University, the University of Colorado, and the University of Tromso in Norway.
Divers earned his Ph.D. from LSU. In his research, he is interested in developing new tech-based assessments of dementia and other disorders, such as depression. He is currently a postdoctoral fellow in neuropsychology at the Ohio State University Wexner Medical Center.
Read the paper: Divers, R., Cohen, A. S., Elvevåg, B., Chandler, C., Turner, R. S., Reynolds, B., & Diaz-Asper, C. (2025). Speech production as an artificial intelligence-based ‘process’ measure of cognition sensitive to mild cognitive impairment and Alzheimer’s disease. The Clinical Neuropsychologist, 1-16.
A meta-analysis replicating the findings from this study in the larger literature has been accepted for publication: "Speech pause and speech rate for evaluating Alzheimer’s and Mild Cognitive Impairment: A Meta-analysis.”
Let LSU put you on a path to success! With 330+ undergraduate programs, 70 master's
programs, and over 50 doctoral programs, we have a degree for you.Next Steps