The hands of a man with Parkinson's disease tremble. Strongly trembling hands of an older man
The hands of a man with Parkinson's disease tremble. [Astrid860/Getty Images]

Scientists in the Department of Electrical Engineering and Computer Science (EECS) at MIT and the MIT Jameel Clinic have developed an artificial intelligence (AI) tool that can detect early Parkinson’s disease by analyzing a person’s breathing patterns. The tool has the potential to significantly improve the diagnosis of this fast-growing neurological disease, which currently relies on the appearance of motor symptoms such as tremors, stiffness, and slowness that only become apparent several years after disease onset.

The research was performed in collaboration with the University of Rochester, Mayo Clinic, and Massachusetts General Hospital, and is sponsored by the National Institutes of Health.

The team, lead by Dina Katabi, a professor at EECS and a principal investigator at the MIT Jameel Clinic, developed the tool via a neural network that can discern the presence of Parkinson’s via a person’s nocturnal breathing patterns—the way they breath during sleep. The neural network, which was trained by MIT PhD student Yuzhe Yang and postdoc Yuan Yuan, can also diagnose the severity of the disease and track disease progression.

“A relationship between Parkinson’s and breathing was noted as early as 1817, in the work of Dr. James Parkinson. This motivated us to consider the potential of detecting the disease from one’s breathing without looking at movements,” said Katabi, senior author of the paper, which appears in Nature Medicine. “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.”

To collect the data needed to train the AI tool, the MIT team developed a non-invasive and non-touch device that resembles a WiFi router and can be placed in the room where a patient sleeps. To collect the breathing patterns without physical contact, the device emits radio signals, analyzes their reflections of the surrounding environment, and then extracts the subject’s breathing patterns. These data are then fed to the neural network for analysis, without the need for any tasks by the patient or caregiver.

The device has the potential to aid diagnosis of the disease and is an improvement over some previous efforts that have attempted to use cerebrospinal fluid and neuroimaging to detect markers of the disease. But these methods are both invasive and require patients to have access to specialized medical centers, making them impractical for the iterative early testing needed for early diagnosis.

Katabi said the new device has the potential to affect both clinical care and the development of new drugs targeting the disease.

“In terms of drug development, the results can enable clinical trials with a significantly shorter duration and fewer participants, ultimately accelerating the development of new therapies,” she noted. “In terms of clinical care, the approach can help in the assessment of Parkinson’s patients in traditionally underserved communities, including those who live in rural areas and those with difficulty leaving home due to limited mobility or cognitive impairment.”

According to Ray Dorsey, a professor of neurology at the University of Rochester, a Parkinson’s specialist and co-author of the paper: “We’ve had no therapeutic breakthroughs this century, suggesting that our current approaches to evaluating new treatments is suboptimal.” He added that the device developed by the Katabi team addresses a shortcoming of prior Parkinson’s research—the lack of data derived from manifestations of the disease in the natural environment.

“The analogy I like to draw [of current Parkinson’s assessments] is a street lamp at night,” Dorsey said, “and what we see from the street lamp is a very small segment…[Katabi’s] entirely contactless sensor helps us illuminate the darkness.”

This site uses Akismet to reduce spam. Learn how your comment data is processed.