August 26, 2022
2 min read
August 26, 2022
2 min read
Katabi reports receiving research funding from NIH and the Michael J. Fox Foundation, being a co-founder of Emerald Innovations Inc., and serving on the scientific advisory board for Janssen and the data and analytical advisory board for Amgen. Please see the study for all other authors’ relevant financial disclosures.
An at-home, artificial intelligence-based system identified individuals with Parkinson’s disease and predicted disease severity and progression using nocturnal breathing signals, according to a study in Nature Medicine.
“A relationship between Parkinson’s and breathing was noted as early as 1817, in the work of Dr. James Parkinson,” Dina Katabi, PhD, Thuan and Nicole Pham Professor and director of the Center for Wireless Networks and Mobile Computing at the Massachusetts Institute of Technology, said in a related MIT press release. “This motivated us to consider the potential of detecting the disease from one’s breathing without looking at movements.
“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 and colleagues evaluated the AI model using a dataset of 7,671 individuals from several sources, including the Mayo Clinic, Massachusetts General Hospital sleep lab and observational clinical trials. The dataset contained 11,964 nights of more than 120,000 hours of nocturnal breathing signals from 757 PD patients (mean age, 69.1 years; 27% women) and 6,914 controls (mean age, 66.2 years; 30% women).
Researchers divided data into breathing belt datasets, from polysomnography sleep studies that use a breathing belt for recordings throughout the night, and wireless datasets, which detect nocturnal breathing using a contactless radio device that extracts “breathing from radio waves that bounce off a person’s body during sleep.”
According to study results, nights measured with a breathing belt achieved an area under the curve (AUC) of 0.889 with a sensitivity of 80.22% (95% CI, 70.28-87.55) and specificity of 78.62% (95% CI, 77.59-79.61). With the wireless signal, researchers reported 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).
Researchers also reported that the AI model could predict PD severity and progression based on the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (R = 0.94).
“Our study demonstrates the feasibility of objective, noninvasive, at-home assessment of PD and also provides initial evidence that this AI model may be useful for risk assessment before clinical diagnosis,” the authors wrote.
According to Katabi, these study findings have important implications for PD treatment and care. “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,” Katabi said in the release. “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.”
Artificial intelligence model can detect Parkinson’s from breathing patterns. https://news.mit.edu/2022/artificial-intelligence-can-detect-parkinsons-from-breathing-patterns-0822. Published Aug. 22, 2022. Accessed Aug. 25, 2022.