It takes a lot of time – and money – to diagnose Alzheimer’s disease. After running lengthy in-person neuropsychological exams, clinicians should transcribe, review, and analyze each response in detail. But researchers at Boston University have created a new tool that can automate the process and eventually allow it to work online. Their machine learning-powered computational model can detect cognitive impairment from audio recordings of neuropsychological tests-no in-person appointment required. Their findings are published in Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association.
“This approach brings us one step closer to early intervention,” said Ioannis Paschalidis, a co -author of the paper and a BU College of Engineering Distinguished Professor of Engineering. He says faster and earlier detection of Alzheimer’s could push for larger clinical trials targeting individuals in the early stages of the disease and possibly enable clinical interventions to slow the progression. mindset: “This could be the basis of an online tool that can reach everyone and can increase. the number of people screening early.”
The research team trained their model using audio recordings of neuropsychological interviews from more than 1,000 individuals in the Framingham Heart Study, a long-standing BU-led project looking at cardiovascular disease and other physiological factors. condition. With the automated online speech recognition tool – think, “Hey, Google!” -and a machine learning technique called natural language processing that helps computers understand text, they have their program transcribe the interviews, then encode them into numbers . A final model was trained to assess the likelihood and severity of a person’s cognitive impairment using demographic data, text encodings, and actual diagnosis from neurologists and neuropsychologists.
Paschalidis said the model not only makes an accurate identification between healthy individuals and those with dementia, but also finds differences between those with mild cognitive impairment and dementia. And, it turns out, the quality of the recordings and how people speak – whether their language is fast or constantly weak – is less important than the content of what they say.
“We’re surprised that the flow of speech or other parts of the audio isn’t so critical; you can automatically transcribe interviews reasonably well, and rely on text analysis via AI to analyze mental retardation, ”said Paschalidis, who is also the new director of BU’s Rafik B. Hariri Institute for Computing and Computational Science & Engineering. Although the team still needs to validate its results against other data sources, the findings suggest that their tool can support clinicians in diagnosing cognitive impairment using audio recordings, including from a virtual or telehealth appointment.
Check before Symptoms
The model also provides insight into which aspects of neuropsychological examination may be more important than others in determining whether an individual has a cognitive impairment. The researchers model divides the examination transcripts into different sections based on the clinical trials performed. They discovered, for example, that the Boston Naming Test – in which clinicians ask individuals to label a photo with a word – is the most informative for an accurate diagnosis of dementia. “This may enable clinicians to allocate resources in a way that allows them to perform additional screening, even before symptoms begin,” Paschalidis said.
Early diagnosis of dementia is not only important for patients and their caregivers to develop an effective plan for treatment and support, but it is also important for researchers working on therapies to relieve and prevent the progression of Alzheimer’s disease. “Our models help clinicians evaluate patients in terms of their chances of dementia,” says Paschalidis, “and then best tailor resources to them by to make more testing of those with a higher likelihood of dementia. “
Want to Participate in a Research Effort?
The research team is looking for volunteers to take an online survey and submit an anonymous cognitive test – the results will be used to provide personalized cognitive assessments and also help the team refine their AI model.
Materials provided by University of Boston. Originally written by Gina Mantica. Note: Content can be edited for style and length.