Diagnosing Parkinson’s disease can be quite challenging as it primarily relies on detecting motor symptoms, such as tremors, stiffness, and slowness, which often appear years after the disease onset. However, MIT’s Professor Dina Katabi and her team have developed an artificial intelligence (AI) model that can detect Parkinson’s by analyzing a person’s breathing patterns.
The AI tool in question is a neural network, which is a series of connected algorithms designed to work like a human brain. It can assess whether someone has Parkinson’s by analyzing their nocturnal breathing patterns while sleeping. Furthermore, the neural network can determine the severity of the disease and track its progression over time.
The team behind the research includes MIT PhD student Yuzhe Yang, postdoc Yuan Yuan, and 12 colleagues from other universities and medical centers. The research paper, published in Nature Medicine, details the team’s success in training the neural network to detect Parkinson’s disease.
Previous research focused on detecting Parkinson’s using cerebrospinal fluid and neuroimaging, but these methods are invasive, expensive, and require access to specialized medical centers. In contrast, the MIT team’s AI assessment of Parkinson’s can be performed every night at home while the person is asleep and without touching their body.
To achieve this, the team developed a device that looks like a home Wi-Fi router. However, instead of providing internet access, the device emits radio signals that analyze reflections off the surrounding environment to extract the subject’s breathing patterns. The breathing signal is then fed to the neural network for Parkinson’s assessment, which requires zero effort from the patient or caregiver.
This groundbreaking research has important implications for Parkinson’s drug development and clinical care. The results can help in the assessment of Parkinson’s patients in underserved communities, including those who live in rural areas or have limited mobility or cognitive impairment. Additionally, the approach can enable clinical trials with fewer participants and a significantly shorter duration, ultimately accelerating the development of new therapies.
From MIT Article in MitNews