Researchers Can Use Breathing Patterns to Detect Parkinson’s Disease

Parkinson’s disease is difficult to diagnose primarily due to the late onset of typical symptoms such as tremors and slowness. Researchers from the Department of Electrical Engineering and Computer Science (EECS) at MIT and MIT Jameel Clinic have developed an AI model to detect Parkinson’s at an early stage. The new AI model offers diagnosis based on the patient’s breathing pattern during his sleep. 

The model consists of a complex neural network that can detect Parkinson’s disease, the severity of the condition, and the progression of the disease over time. The data from several hospitals in the United States and multiple public datasets were used to test the algorithm on 7,687 individuals, including 757 Parkinson’s patients. 

The new AI device is small and has a non-contact sensor system that looks like a home Wi-Fi router. The device emits radio signals that are then analyzed by an AI algorithm as they reflect off from the surrounding environment. The reflection of radio signals provides the patient’s breathing pattern without any physical contact with the patient. 

Parkinson’s is one of the fastest-growing neurological diseases in the world. More than 1 million people in the U.S. are affected. Currently, there is no cure for Parkinson’s. Early diagnosis through breathing patterns could help participation in short and small-sized clinical trials, which would help fasten the efforts to create a new therapy for treatment. 

sleep
therapy