AI Model Can Detect Parkinson's From Breathing Patterns 14
An anonymous reader quotes a report from MIT News: Parkinson's disease is notoriously difficult to diagnose as it relies primarily on the appearance of motor symptoms such as tremors, stiffness, and slowness, but these symptoms often appear several years after the disease onset. Now, Dina Katabi, the Thuan (1990) and Nicole Pham Professor in the Department of Electrical Engineering and Computer Science (EECS) at MIT and principal investigator at MIT Jameel Clinic, and her team have developed an artificial intelligence model that can detect Parkinson's just from reading a person's breathing patterns. The tool in question is a neural network, a series of connected algorithms that mimic the way a human brain works, capable of assessing whether someone has Parkinson's from their nocturnal breathing -- i.e., breathing patterns that occur while sleeping. The neural network, which was trained by MIT PhD student Yuzhe Yang and postdoc Yuan Yuan, is also able to discern the severity of someone's Parkinson's disease and track the progression of their disease over time.
The MIT researchers demonstrated that the artificial intelligence assessment of Parkinson's can be done every night at home while the person is asleep and without touching their body. To do so, the team developed a device with the appearance of a home Wi-Fi router, but instead of providing internet access, the device emits radio signals, analyzes their reflections off the surrounding environment, and extracts the subject's breathing patterns without any bodily contact. The breathing signal is then fed to the neural network to assess Parkinson's in a passive manner, and there is zero effort needed from the patient and caregiver. "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," Katabi says. "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." The research has been published in the journal Nature Medicine.
The MIT researchers demonstrated that the artificial intelligence assessment of Parkinson's can be done every night at home while the person is asleep and without touching their body. To do so, the team developed a device with the appearance of a home Wi-Fi router, but instead of providing internet access, the device emits radio signals, analyzes their reflections off the surrounding environment, and extracts the subject's breathing patterns without any bodily contact. The breathing signal is then fed to the neural network to assess Parkinson's in a passive manner, and there is zero effort needed from the patient and caregiver. "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," Katabi says. "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." The research has been published in the journal Nature Medicine.
Finally something usefull for AI (Score:2)
Can we now get rid of the AI buzz word?
Re: (Score:2)
With what accuracy (Score:3, Interesting)
Re:With what accuracy (Score:5, Informative)
Sleep apnea (Score:3)
Re:With what accuracy (Score:4, Informative)
The AI model can detect PD with an area-under-the-curve of 0.90 and 0.85 on held-out and external test sets, respectively
I feel like that explains it, but in case anyone needs an overview accuracy vs area under the curve. [baeldung.com] That said sample size was:
757 PD subjects (mean (s.d.) age 69.1 (10.4), 27% women) and 6,914 control subjects (mean (s.d.) age 66.2 (18.3), 30% women)
So definitely warrants a further study as it's likely, given these numbers, that a link indeed exists between the brain's regulation of breathing and PD, but if the qualifier is:
if I had Parkinson's I would want to be sure of it
Then yeah, this ain't your ship. A lot of test aren't 100%, hell most aren't even in the domain of two sigma. That said, you've got to remember that at the moment Parkinson's is a clinical disease, we don't have any kind of blood test for it. So if an AUC like this is pointing towards it most likely will get better with time, it's kind of one of those something might be better than nothing kinds of things. So just to really temper expectations here, they aren't pitching this kind of test as accurate every time. But they sure are pitching it as "a tool" in the empty tool chest doctors have for diagnosis of early on-set PD.
So I guess ask yourself if this test was only 50% accurate, would that be better than your doctor randomly guessing? Maths wise yeah, they sound about the same. But your doctor is likely to guess the same thing over and over, but you can have multiple test with varying results. I don't know, I just think it's a bit early in the game to be thinking absolute accuracy of diagnosis especially for a clinical disease. I feel this study was more about trying to point out that neurological tell-tells are a thing in early PD. But that's just my read.
When... (Score:3)
Re:When... (Score:5, Interesting)
Alexa (2025 edition, now with radar echo) in your bedroom could monitor this to add to Amazon's profile of you, so Amazon can sell that to interested parties (your medical insurer, your children's insurers, local specialist clinics who might want to target you, etc) or offer you products and medication.
Who wouldn't want this!
Re: (Score:2)
Good luck, Alexa! WIth my snoring, it's going to need anti-vibration mounts and some sort of AI to filter out the noise in the audio signal before it gets anywhere near the Parkinsons detector.
Only the lonely (Score:1)
where're data? (Score:2)
Would be nice if the linked article contained some statistical data like false positives/negatives, sample size, etc - or did I miss it somehow?
Without it it's pretty useless.