Truth, Inspiration, Hope.

Diagnosing Parkinson’s Disease With Artificial Intelligence and Breathing

Jonathan Ferng, MD, MBA, MS
Jonathan Ferng is an internal medicine physician who has a wide range of interests spanning healthcare, business, consulting, research, and music. He enjoys meditating, learning new skills, and sharing positivity with the world.
Published: September 1, 2022
A new MIT study diagnosed Parkinson's Disease using artificial intelligence and breathing measurements
Measuring breathing patterns while sleeping may be the key to diagnosing Parkinson’s Disease early. (Image: SHVETS Production via Pexels)

Currently, there are few effective biomarkers for diagnosing Parkinson’s disease (PD), and a dearth of therapeutic options exist for treating or slowing down its disease course. PD is often diagnosed late because clinical symptoms such as tremor or rigidity typically occur years after disease onset.

While biomarkers from cerebrospinal fluid or blood and neuroimaging are accurate, they are costly, invasive, and often only available at specialized facilities. In contrast, the link between PD and breathing, first reported by James Parkinson himself in 1817, may be the key to early diagnosis.


On Aug. 22, Massachusetts Institute of Technology (MIT) researchers published a paper titled Artificial Intelligence-enabled Detection and Assessment of Parkinson’s Disease Using Nocturnal Breathing Signals in the prestigious Nature Medicine journal.

Researchers from Rutgers University, the University of Rochester Medical Center, the Mayo Clinic, Massachusetts General Hospital, and the Boston University College of Health and Rehabilitation contributed to the study.

Breathing belts and wireless measurements

Nearly one million American citizens have PD, a number expected to climb to 1.2 million by 2030. More than 10 million people worldwide have been diagnosed with the disease, and the financial burden in the U.S. alone is estimated to be $52 billion USD.

According to The Parkinson’s Foundation, 10 early signs of PD are shaking tremors, small handwriting, loss of smell, trouble sleeping, trouble moving or walking, constipation, a soft or low voice, a masked face with a more serious, depressed, or mad look, dizziness or fainting, and stooping or hunching over.

The MIT researchers presented a new system based on artificial intelligence (AI) monitoring of nocturnal breathing to help facilitate early PD diagnosis, track disease severity, and monitor disease progression over time.

Breathing data was collected either through a specialized breathing belt worn by the subject or transmission of a “low power radio signal and analyzing its reflections off the person’s body.”

A total of 11,964 nights with more than 120,000 hours of data from 757 PD subjects contrasted against a 6,914 subject control group were included in the study. One to two nights of data were recorded for the breathing belt datasets, compared to up to one year of longitudinal data for the wireless radio devices.

The accuracy of diagnosing PD through one night of recording was assessed through receiver operating characteristic (ROC) curves.

The accuracy was found to be high, with the breathing belt method achieving “an area under the ROC curve (AUC) of 0.889 with a sensitivity of 80.22% (95% confidence interval (CI) (70.28%, 87.55%)) and specificity of 78.62% (95% CI (77.59%, 79.61%)),” the study stated. 

Researchers continued, “For nights measured using wireless signals, the model achieves an AUC of 0.906 with a sensitivity of 86.23% (95% CI (84.08%, 88.13%)) and specificity of 82.83% (95% CI (79.94%, 85.40%)).”

Higher AUC values are associated with higher accuracy. Sensitivity measures a test’s ability to correctly identify patients with a disease, while specificity measures a test’s ability to correctly identify patients without the disease.

The wireless radio device method had both a higher sensitivity and specificity. The authors also found that when several nights of data were combined for an individual, sensitivity and specificity of the test reached near 100 percent, with test-rest reliability improving significantly.

A promising future

Professor and principal investigator of the study at MIT Jameel Clinic, Dina Katabi, told MIT News, “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.”

“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,” Dr. Katabi added.

Dr. Ray Dorsey, Professor of Neurology and Director of the Center for Health and Technology at the University of Rochester Medical Center, stated, “We’ve had no therapeutic breakthroughs this century, suggesting that our current approaches to evaluating new treatments is suboptimal. 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.” 

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