AI Predicts Alzheimer’s via Speech Analysis

AI Model Predicts Alzheimer's Via Speech Analysis

Researchers at Boston University claim to have created an artificial intelligence (AI) tool that can identify with about 80% accuracy whether a person is at risk of getting Alzheimer’s disease by analyzing their speech patterns.

Determining whether someone has Alzheimer’s typically entails a battery of testing—interviews, brain imaging, blood and cerebrospinal fluid tests. Unfortunately, it is probably too late by then—long-standing personality features have started to change gradually, and memories have already begun to fade.

New groundbreaking medicines may slow down the disease’s progression if detected early, but there is no way to predict who will acquire Alzheimer’s dementia

However, researchers at Boston University claim they have created a novel and promising artificial intelligence model that may one day assist in predicting who will develop Alzheimer’s via voice analysis of patients.

About the Research

Boston University researchers have developed an AI tool that can identify with about 80% accuracy whether a person is at risk of getting Alzheimer’s by analyzing their speech patterns [1].

The researchers used data from one of the oldest and longest-running studies in the country, the Boston University-led Framingham Heart Study, to train and construct their new model.

Despite the Framingham study’s primary focus on cardiovascular health, individuals exhibiting cognitive deterioration are routinely assessed neuropsychologically and through interviews, yielding an abundance of longitudinal data on their cognitive health.

Researchers investigated whether mild cognitive decline patients would get Alzheimer’s within six years using a natural language processing model. Their study involved 166 individuals with varying levels of cognitive problems, aged between 63 and 97. Of these, 107 were women and 59 were men.

Every participant had been recorded for an hour-long interview as part of the study. These interviews provided the data used to examine the AI tool developed by the researchers.

Out of the cohort, 76 had stable cognitive performance, and 90 showed progressive declines. However, the researchers at Boston University found that by fusing machine learning and voice recognition technologies, they could monitor the relationships between speech patterns and cognitive loss by looking for biomarkers linked to cognitive decline.

According to the researchers, though the sample size was limited, their model predicted considerable cognitive impairment with 78.5% accuracy.

Researchers said that future applications of these models might include routine monitoring through interaction with an at-home app or care delivery to patients living far from medical facilities, thereby significantly raising the screening rate.

Future Prospects: A Dementia Diagnosis from Home

The researchers hope to use data from more casual, everyday talks in addition to conventional clinician-patient interviews, which follow a predetermined format and include prepared questions and back-and-forth.

To increase the model’s predictive accuracy, they are already planning to expand the current study beyond voice analysis—the Framingham tests also contain patient drawings and data on everyday living patterns—and investigate if AI might assist in diagnosing dementia using a smartphone app.

How might AI benefit both Alzheimer’s patients and medical professionals?

According to Emer MacSweeney, MD, CEO and consultant neuroradiologist of Re:Cognition Health in London, England. Clinicians and patients may have more choices if artificial intelligence helps identify Alzheimer’s early on [2]. 

The AI tool for predicting the course of Alzheimer’s disease offers several noteworthy advancements. It makes cognitive assessments more accessible through automated and remote screening. Allows for early intervention with treatments to slow the disease. And makes customized care plans based on predicted disease trajectories possible.

Early prescription is the key to the optimal efficacy of the current generation of Alzheimer’s medications. Amyloid is a protein found in the brain that is frequently used to identify possible cognitive deterioration. Since many individuals who test positive for amyloid in the brain do not go on to. Experience cognitive symptoms, the AI model can make more accurate predictions.

Does the use of AI for monitoring Alzheimer’s risk raise any concerns?

Although a nearly 80% accuracy rate is high, MacSweeney noted that there is still a chance of false positives or negatives and that automating healthcare could lead to issues. There will be instances where the tool wrongly predicts the course or stability of the condition. Which could cause unnecessary anxiety or false satisfaction.

While AI has the potential to transform some aspects of healthcare. There is a chance that doctors would over-rely on AI predictions without taking the larger clinical context into account. Which could result in incorrect diagnoses.

Alzheimer’s Research Association is a non-profit organization dedicated to helping caregivers of Alzheimer’s disease and dementia. We provide the latest information and news about the illness and helpful tips to help caregivers cope with their daily caregiving challenges. We realize the most important thing that a caregiver needs is financial assistance. Therefore, we provide grants to caregivers to ease their financial burden. Caregivers can apply for grants here: Alzheimer’s Grant Application

You can also help caregivers in their endeavor by donating as much as possible: Donation To Alzheimer’s Research Associations.

Comments

No comments yet. Why don’t you start the discussion?

Leave a Reply

Your email address will not be published. Required fields are marked *