ChatGPT’s AI could help detect Alzheimer’s earlier

Feb. 3, 2023 — Artificial intelligence that can write essays and pass tests may also help identify dementia.

Researchers at Drexel University in Philadelphia used the AI ​​behind ChatGPT (who made headlines for writing credible term papers and passing bar exams) to analyze speech, and the system correctly identified patients with Alzheimer’s disease 80% of the time, according to the to study published in the journal PLOS digital health.

The researchers used GPT-3, the language model that drives ChatGPT, to analyze audio clips of people describing an image in a standard dementia test.

Patients with Alzheimer’s disease often repeated themselves, strayed from describing the content of the image, did not complete their thoughts, and vaguely referred to objects as “thing” or “something”.

“GPT-3 is able to capture such a subtle difference reflected in text,” says study author Hualou Liang, PhD, professor of biomedical engineering at Drexel.

The software analyzed text transcribed (also by software) from 10-second recordings of healthy adults and patients with Alzheimer’s disease. The text trained the GPT-3 model to identify subtle differences between normal language and the speech of a person with cognitive decline.

GPT-3 machine learning models understand passages of text by converting words into mathematical representations called embeddings. Incorporations are multi-dimensional signals, which allow AI to identify subtle differences and similarities that even experienced doctors cannot hear. GPT-3 compares passages of text by measuring the distance between these signals in embeddings.

Since GPT-3 only analyzes written text, the process bypasses pauses and other non-word sounds in spoken language. In this case, it turned out to be an advantage: GPT-3 analysis outperformed some machine learning models developed by other labs that included these sounds.

Other studies, however, have shown that “ah”s and “ums” in speech may be important in revealing Alzheimer’s disease. A study 2021 who coded these pauses enabled a machine learning model to detect Alzheimer’s disease with 90% accuracy, and a separate study conducted in Slovenia which combined textual and acoustic features achieved an accuracy of 94%.

“The best combination tends to combine both types of functionality,” says Frank Rudzicz, PhD, associate professor of computer science at the University of Toronto. “There’s a lot of information in the words and structure of the transcripts, but also in our tone of voice.”

Using voice to spot Alzheimer’s disease

More and more researchers are studying voice as a biomarkera means of detecting various diseases including Alzheimer’s.

Worldwide, Alzheimer’s cases are only successfully detected 48% of the time, according to estimates by the World Health Organization. High-income countries achieve a 54% diagnosis rate, while low- and middle-income countries identify only 24% of Alzheimer’s cases.

Researchers in this field hope to fill this gap by developing a tool that can detect Alzheimer’s disease early, when the effects may be too subtle for a doctor to notice. “There is no cure for Alzheimer’s disease yet, but certain life changes can delay some of its effects, so early diagnosis is always important,” says Rudzicz, who co-founded a mobile analytics app of the spoken word winter light. “These types of technologies could also be applied to other disorders, including Parkinson’s disease, depression, etc.”

Doctors could potentially use a device or computer program to test a patient’s cognitive abilities in their office. Brain scans or other clinical tests could then confirm the diagnosis of Alzheimer’s.

Another app can use smart devices like Alexa and Siri to monitor your regular conversations (with your consent) and alert you if it notices a disturbing word. It can even detect other psychological issues like depression and stress.

“The analysis could be done in a privacy-preserving way once the system is fully functional,” Liang says. “As such, it could have an immediate and significant impact on alleviating the problem of dementia in the older community.”