Summary: A newly developed artificial intelligence model can detect Parkinson’s disease by reading a person’s breathing patterns. The algorithm can also recognize the severity of Parkinson’s disease and track progression over time.
Parkinson’s disease is notoriously difficult to diagnose because it relies primarily on the appearance of motor symptoms such as tremors, stiffness, and slowness, but these symptoms often appear years after the onset of the disease.
Now Dina Katabi, Thuan (1990) and Nicole Pham Professor in the Department of Electrical Engineering and Computer Science (EECS) at MIT and Principal Investigator at the MIT Jameel Clinic and her team have developed an artificial intelligence model that can detect Parkinson’s disease. from reading a person’s breathing patterns.
The tool in question is a neural network, a series of interconnected algorithms that mimic the way the human brain works, and that are able to judge whether someone has Parkinson’s disease from nocturnal breathing—that is, from the breathing patterns that occur during sleep.
A neural network trained by MIT Ph.D. student Yuzhe Yang and postdoc Yuan Yuan is also able to recognize the severity of someone’s Parkinson’s disease and track the progression of their disease over time.
Yang and Yuan are joint first authors of a new paper describing the work, published today in the Natural medicine. Katabi, who is also an affiliate of the MIT Computer Science and Artificial Intelligence Laboratory and director of the Center for Wireless Networks and Mobile Computing, is the lead author.
They are joined by 12 colleagues from Rutgers University, University of Rochester Medical Center, Mayo Clinic, Massachusetts General Hospital and Boston University College of Health and Rehabilitation.
Over the years, researchers have explored the potential of detecting Parkinson’s disease using cerebrospinal fluid and neuroimaging, but these methods are invasive, expensive and require access to specialized medical centers, making them unsuitable for frequent testing that might otherwise provide early diagnosis or continuous monitoring. disease progression.
MIT researchers have demonstrated that an artificial intelligence assessment of Parkinson’s disease can be performed every night at home while the person is asleep and without touching their body.
To this end, the team developed a device with the appearance of a home Wi-Fi router, but instead of providing access to the Internet, the device transmits radio signals, analyzes their reflections from the surrounding environment and extracts the breathing patterns of the subject without any physical contact.
The breath signal is then fed into a neural network to assess Parkinson’s disease in a passive way, requiring no effort from the patient and caregiver.
“The relationship between Parkinson’s disease and breathing was noted as early as 1817 in the work of Dr. James Parkinson. This motivated us to consider the potential of detecting disease from one’s own breathing without looking at movements,” says Katabi.
“Some medical studies have shown that respiratory symptoms appear years before motor symptoms, meaning that attributes of breathing could hold promise for risk assessment before a Parkinson’s diagnosis.”
The fastest growing neurological disease in the world, Parkinson’s disease is the second most common neurological disorder after Alzheimer’s disease. It affects more than 1 million people in the United States alone and has an annual economic burden of $51.9 billion. The research team’s device was tested on 7,687 individuals, including 757 patients with Parkinson’s disease.
Katabi notes that the study has important implications for Parkinson’s drug development and clinical care. “When it comes to drug development, the results can enable clinical trials with significantly shorter durations and fewer participants, ultimately speeding up the development of new therapies.
“In terms of clinical care, this approach can aid in the assessment of Parkinson’s patients in traditionally underserved communities, including those living in rural areas and those who have difficulty leaving home due to limited mobility or cognitive impairment,” he says.
“We haven’t seen any therapeutic breakthroughs this century, which suggests that our current approaches to evaluating new treatments are suboptimal,” says Ray Dorsey, a professor of neurology at the University of Rochester and an expert in Parkinson’s who co-authored the paper. Dorsey adds that the study is probably one of the largest sleep studies ever conducted in Parkinson’s disease.
“We have very limited information about disease manifestations in their natural environment and [Katabi’s] the device allows you to get objective, real-world assessments of how people are doing at home.
“A simile 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] a completely contactless sensor helps us illuminate the darkness.”
About this AI and Parkinson’s research news
Author: Anne Trafton
Contact: Anne Trafton-MIT
Picture: The image is in the public domain
Original Research: Open access.
“Detection and Assessment of Parkinson’s Disease Using Nocturnal Breathing Signals Using Artificial Intelligence” by Yuzhe Yang et al. Natural medicine
Detection and assessment of Parkinson’s disease using nocturnal breathing signals using artificial intelligence
There are currently no effective biomarkers for diagnosing Parkinson’s disease (PD) or monitoring its progression.
Here, we developed an artificial intelligence (AI) model to detect PD and track its evolution from nocturnal respiratory signals. The model was evaluated on a large dataset of 7,671 individuals, using data from several hospitals in the United States as well as several public datasets.
The AI model can detect PDs with an area under the curve of 0.90 and 0.85 on the retained and external test sets. The AI model can also estimate the severity and progression of PD according to the Movement Disorder Society’s Unified Parkinson’s Disease Rating Scale (R= 0.94, P= 3.6 × 10-25).
The AI model uses an attention layer that allows it to interpret its predictions with respect to sleep and electroencephalogram. In addition, the model can non-contactly assess PD in the home environment by extracting breathing from the radio waves that are reflected from the human body during sleep.
Our study demonstrates the feasibility of an objective, noninvasive, home-based assessment of PD and also provides initial evidence that this AI model may be useful for risk assessment prior to clinical diagnosis.
#Artificial #intelligence #model #detect #Parkinsons #disease #breathing #patterns #Neuroscience #News