Researchers at MIT have developed a sensor, about the size and shape of a WiFi router, which they say can help track Parkinson’s patients’ breathing while they sleep. Tracking is completely contact-free, and the device alerts caregivers to any progression of the condition—and could also be used to diagnose Parkinson’s.
The device emits radio waves and captures their reflection to read small changes in its immediate environment. It works like a radar, but in this case, the device senses the rise and fall of a person’s chest.
An artificial intelligence-powered system then takes that information and analyzes it for patterns that can be linked to some of the earliest signs of Parkinson’s or to record changes in the severity of the disease over time.
“A relationship between Parkinson’s and breathing was noted as early as 1817, in the work of Dr. James Parkinson,” professor Dina Katabi, principal investigator at the university’s AI-focused Jameel Clinic, said to MIT News. “This motivated us to consider the potential of detecting the disease from one’s breathing without looking at movements.”
Parkinson’s is traditionally diagnosed by a more subjective examination of muscle stiffness, slowness or tremors. However, the researchers said these symptoms can often become apparent long after the disease has taken hold.
“Some medical studies have shown that respiratory symptoms manifest years before motor symptoms, meaning that breathing attributes could be promising for risk assessment prior to Parkinson’s diagnosis,” Katabi said.
Other diagnostic options include brain scans or the collection of cerebrospinal fluid samples. The AI model, by comparison, can easily collect data on a daily basis. Trained by MIT Ph.D. student Yuzhe Yang and postdoc Yuan Yuan, the neural network program was the subject of a study published this week in the journal Nature Medicine.
When used to analyze readings from over 7,600 individuals collected from sensors in several U.S. hospitals and sleep labs, as well as from other public datasets—including 757 people with Parkinson’s—the model demonstrated a high degree of accuracy in detecting Parkinson’s disease. It correctly identified positive cases 80% of the time and negative cases 82% of the time.
The approach could also be used to help advance the development of new therapies for Parkinson’s, the researchers said, by making it easier to catch a clear signal when a treatment is working.
Yang and Yuan were joined by colleagues from Rutgers University, the University of Rochester Medical Center, the Mayo Clinic, Massachusetts General Hospital and the Boston University College of Health and Rehabilitation. The study was sponsored by the National Institutes of Health, with support from the National Science Foundation and the Michael J. Fox Foundation.
“We’ve had no therapeutic breakthroughs this century, suggesting that our current approaches to evaluating new treatments is suboptimal,” the paper’s co-author, Ray Dorsey, a professor of neurology at the University of Rochester, said to MIT News. “We have very limited information about manifestations of the disease in their natural environment and [Katabi’s] device allows you to get objective, real-world assessments of how people are doing at home.”