AI detects or predicts Parkinson's and potential disease-modifying drugs
Dina Katabi, principal investigator for AI and health at the MIT Abdul Latif Jameel Clinic for Machine Learning in Health (MIT Jameel Clinic) is a pioneer in the field of passive monitoring of patients with Parkinson's disease. Dina's devices utilise very low radio signals to monitor a patient's movement and breathing remotely, gathering data which is interpreted by AI to determine disease progression. Despite concern around population-level data generation and the implications of data use, Dina's company, Emerald Innovations, ensures that the technology is properly utilised by researchers, and not accessed by consumers, stating, “We are focused on getting the science properly done for patients and doctors. We want it used within the health care system with proper protocols.”
A smart watch, a 10-second electrocardiogram, an innocuous wall-mounted device in the home, and other new, objective tools may help diagnose, stage, or predict the onset of Parkinson's disease (PD) if recent studies using artificial intelligence (AI) are confirmed, some neurologists say.
AI programs already have identified drugs developed for other indications that may serve as treatments for PD. And multiple studies have shown the potential of AI to detect PD up to 10 years before a clinical diagnosis. The studies rely on data from devices including a smart phone's accelerometer, a standard 10-second electrocardiogram, and an investigational device that works like a bat to analyze radio waves bouncing off individuals as they sleep.
“Because Parkinson's disease involves a disturbance of movement, I do think that AI will be quite useful in identifying patterns of movement we have not been able to parse out through the clinical exam,” said Samuel Goldman, MD, MPH, professor emeritus at the University of California, San Francisco, where he studies environmental risk factors for PD and other neurodegenerative disorders.