YOUR HEALTH: Using speech to detect Dementia

According to the World Health Organization, over 10 million people are diagnosed with dementia each year, worldwide.
Published: Nov. 15, 2022 at 6:05 AM CST|Updated: Nov. 15, 2022 at 6:59 AM CST
Email This Link
Share on Pinterest
Share on LinkedIn

ORLANDO, Fla. (Ivanhoe Newswire) – According to the World Health Organization, over 10 million people are diagnosed with dementia each year, worldwide. As Ivanhoe reports, new research using a simple recording of your voice and a special computer algorithm could help early detection of dementia.

Neurological tests to determine a person’s cognitive ability can take a lot of time because clinicians have to transcribe, review, and analyze every response in vivid detail.

But now, researchers at Boston University have developed a new tool that could automate the process. The machine-learning computer model can detect cognitive impairment from audio recordings of neuropsychological tests without you even having to go to the doctor.

“Why have dementia if we can reduce those things we know are modifiable that are strongly associated with the risk of dementia?” says James E. Galvin, MD, MPH, University of Miami Miller School of Medicine.

By using automated online speech recognition tools like, “hey, google!”, and a machine learning technique called “Natural Language Processing” it helps computers understand your recorded text allowing the model to access the likelihood and severity of a person’s cognitive impairment.

Faster and earlier detection of Alzheimer’s could drive larger clinical trials that focus on people in early stages of the disease and potentially enable clinical interventions that slow cognitive decline.

Doctor Galvin says, “The idea is that instead of waiting for disease to happen, we try to prevent it from happening first.”

The model was not only able to accurately distinguish between healthy people and those with dementia, but it also detected differences between those with mild cognitive impairment and dementia. As it turns out, the quality of the recordings and how people spoke were less important than the content of what they were actually saying.

The research team trained their model using audio recordings from more than one thousand neuropsychological interviews and still needs to validate its results against other sources of data. But, the findings suggest their tool could support clinicians in diagnosing cognitive impairment using audio recordings, including those from virtual or telehealth visits.

Click here to report a typo.