© 2024 Texas Public Radio
Real. Reliable. Texas Public Radio.
Play Live Radio
Next Up:
0:00
0:00
0:00 0:00
Available On Air Stations

Science & Medicine: An app for detecting dementia decades early

Ways To Subscribe
Roberto Martinez
/
TPR

Nearly imperceptible changes in the way a person speaks could alert doctors to approaching dementia decades before it is typically diagnosed, and while people might not notice the changes, a digital tool that uses artificial intelligence to analyze speech patterns might.

Robin Hilsabeck, PhD, is a clinical neuropsychologist at UT Health San Antonio. She’s a Professor of Neurology and Chief of the Division of Clinical Neuropsychology. She has extensive experience in diagnosis and treatment of neurocognitive disorders, like Alzheimer’s, and she leads a team that’s developed an app to help people get diagnoses much earlier than they ever could before.

Robin Hilsabeck, PhD, ABPP Professor of Neurology Chief, Division of Clinical Neuropsychology The University of Texas Health Science Center at San Antonio
UT Health San Antonio
Robin Hilsabeck, PhD, ABPP
Professor of Neurology
Chief, Division of Clinical Neuropsychology
The University of Texas Health Science Center at San Antonio

“The specific task that we're using, we know, has the ability to identify decline 20 years before it actually shows up,” Hilsabeck said. “So we (run) this tool. If it shows that people aren't performing where we would expect them to for their age, then we might be a little bit concerned.”

Hilsabeck’s tool is designed to be simple so that it can be easily used in a primary care office. It combines a few questions assessing risk, then evaluates working memory and speech to determine if there are signs of cognitive impairment. If there are, the patient’s primary care doctor will refer them to appropriate specialists.

How does the tool use artificial intelligence to assess speech patterns? Patients are asked to record themselves sharing a personal memory.

“There's a total of 55 acoustic and linguistic variables that go into this machine learning algorithm,” Hilsabeck said. “The most important ones are things like, how many pauses there are. The complexity of the story that you tell. There are some aspects, like pronunciation that are part of the algorithm.”

The app then issues a score. If it falls within a certain range, the program might be detecting cognitive changes that could signal oncoming dementia.

When potential dementia is detected that early, a person might be able to get ahead of it with lifestyle changes and — in Alzheimer's — new medications that must be taken early in the course of the disease to make a difference.

“If you know about it sooner, you can start implementing changes sooner in your life,” said Hilsabeck.

This tool is currently being validated in two primary care clinics, to see how well it works and if primary care doctors find it useful and easy to use. A Spanish language version will soon be available in a third primary care clinic.

Science & Medicine is a collaboration between TPR and The University of Texas Health Science Center at San Antonio that explores how scientific discovery in San Antonio advances the way medicine is practiced everywhere.