AI Predicts Alzheimer’s Onset with 78.5% Accuracy Using Speech Analysis

Boston University researchers have developed an artificial intelligence (AI) program that can predict with 78.5% accuracy whether someone with mild cognitive impairment will develop Alzheimer’s-associated dementia within six years. The AI model, designed by a team led by Ioannis Paschalidis, director of the BU Rafik B. Hariri Institute for Computing and Computational Science & Engineering, uses machine learning to analyze a patient’s speech patterns. The team believes their work could make cognitive impairment screening more accessible and efficient, potentially enabling earlier diagnoses and interventions. The findings were published in Alzheimer’s & Dementia, the journal of the Alzheimer’s Association.

AI Model Predicts Alzheimer’s Disease Progression Through Speech Analysis

Boston University researchers have developed an artificial intelligence (AI) model that can predict the likelihood of a person with mild cognitive impairment developing Alzheimer’s-associated dementia within six years. The model, which analyzes speech patterns, has an accuracy rate of 78.5 percent. This innovative approach could potentially revolutionize the early diagnosis of Alzheimer’s disease, making it more accessible and less reliant on expensive lab tests and imaging exams.

The AI model is powered by machine learning, a subset of AI where a program is taught to independently analyze data. “We wanted to predict what would happen in the next six years—and we found we can reasonably make that prediction with relatively good confidence and accuracy,” says Ioannis (Yannis) Paschalidis, director of the BU Rafik B. Hariri Institute for Computing and Computational Science & Engineering. The team’s findings were published in Alzheimer’s & Dementia, the journal of the Alzheimer’s Association.

Utilizing Longitudinal Data for AI Training

To train their model, the researchers used data from the BU-led Framingham Heart Study, one of the nation’s oldest and longest-running studies. Although primarily focused on cardiovascular health, the study also provides a wealth of longitudinal information on cognitive well-being. The researchers were given audio recordings of initial interviews with 166 individuals, aged between 63 and 97, diagnosed with mild cognitive impairment. They then used a combination of speech recognition tools and machine learning to train a model to spot connections between speech, demographics, diagnosis, and disease progression.

The Power of AI in Predicting Alzheimer’s Disease

The AI model does not rely on acoustic features of speech, such as enunciation or speed. Instead, it analyzes the content of the interview—the words spoken and how they’re structured. Despite the low-quality and background noise in the recordings, the model was able to make accurate predictions. This demonstrates the potential of AI in making the process of dementia diagnosis more efficient and automated, with minimal human involvement.

Expanding Accessibility of Alzheimer’s Disease Screening

The researchers believe that their model could be used to bring care to patients who aren’t near medical centers or to provide routine monitoring through interaction with an at-home app. This could drastically increase the number of people who get screened. According to Alzheimer’s Disease International, the majority of people with dementia worldwide never receive a formal diagnosis, leaving them shut off from treatment and care. The use of AI in this context could create “equal opportunity science and healthcare,” says Rhoda Au, a coauthor on the paper.

Future Research and Applications

In future research, Paschalidis is interested in exploring the use of data from more natural, everyday conversations, not just formal clinician-patient interviews. He is also considering the possibility of diagnosing dementia via a smartphone app and expanding the current study beyond speech analysis to include patient drawings and data on daily life patterns. This could potentially enhance the model’s predictive accuracy. “Digital is the new blood,” says Au. “You can collect it, analyze it for what is known today, store it, and reanalyze it for whatever new emerges tomorrow.”