AI Breakthrough: Predicting Alzheimer's Through Speech Patterns

Dementia affects over 55 million people globally, with Alzheimer’s disease affecting up to 70% of them.
Representational image
Representational image

Researchers at Boston University have created a tool using artificial intelligence that can predict with nearly 80% accuracy whether someone might develop Alzheimer’s disease based on how they speak.

They used a natural language processing model to study if people showing early signs of cognitive decline would progress to Alzheimer’s over six years. The study involved 166 participants (107 women and 59 men) aged 63 to 97, all with varying levels of cognitive issues.

These participants had previously taken part in the Framingham Heart Study led by Boston University and had been recorded during hour-long interviews. The AI tool analyzed these recordings to make its predictions.

Out of the group studied, 90 individuals experienced worsening cognitive function, while 76 showed no significant change.

The researchers at Boston University found that by using speech-recognition tools and machine learning together, they could identify links between speech patterns and cognitive decline, using biomarkers linked to such decline. Despite the study's limited size, their model achieved a 78.5% accuracy in predicting substantial cognitive decline, according to the researchers.

Dementia affects over 55 million people globally, with Alzheimer’s disease affecting up to 70% of them. Alzheimer’s is characterized by the loss of brain cells due to the buildup of amyloid and tau proteins, leading to symptoms like memory loss, cognitive issues, difficulty with speech, spatial awareness, and changes in behavior and personality.

The disease progresses gradually, starting with mild symptoms that worsen over time. While there is no cure, treatment typically involves medication, lifestyle adjustments, and support groups for both patients and caregivers.

According to a report from the Alzheimer’s Association, Alzheimer’s disease, the most prevalent form of dementia, impacts 3% of Americans aged 65–74. The risk rises with age: 17% of those aged 75–84 and 32% of those aged 85 and older are diagnosed with dementia.

There's a common misconception that dementia is solely genetic, meaning if a family member has it, others will definitely get it too. This isn’t accurate.

While genetics play a role in certain types of dementia, most cases aren't strongly linked to genes. As mentioned earlier, age is the primary risk factor for dementia. However, if a parent or grandparent developed Alzheimer’s before age 65, the genetic risk can be higher.


Representational image
Manipur: Rahul Gandhi Visits Relief Camps in Jiribam, Meets Victims of Violence


Top Headlines

No stories found.
Sentinel Assam