Using Invisible Radio Waves To Detect Parkinson’s With Help From AI

Data and machine learning hold transformative potential for the healthcare sector, propelling it towards a future of personalized and preventative medicine. As we amass vast volumes of health-related data from sources like electronic medical records, wearable devices, and genetic tests, machine learning algorithms can analyze this data to extract valuable insights. These insights can enhance diagnosis by identifying patterns undetectable by the human eye, predict disease progression by learning from past medical histories, and even personalize treatment plans by considering individual genetic makeup and lifestyle factors. The fusion of data and machine learning has the potential to revolutionize healthcare, making it more accurate, efficient, and patient-centric.

Dina Katabi, MIT CSAIL member, Andrew & Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT, and co-founder of Emerald Innovations, is transforming healthcare by enabling continuous collection of data about patience using insights from WiFi fields computed by machine learning. Prof. Katabi sees the future of healthcare firmly anchored in a data-centric approach. The key to future healthcare transformations, lies in the continuous collection of clinical data. The future has us gathering ongoing clinical data from patients within the comforts of their homes. Using this data, it will be possible to monitor symptoms in real time, track the patient progression over long periods of time, and use machine learning to gain preemptive insights. This will empower doctors to intervene at early stages, averting hospitalizations, and ultimately, enhancing patient outcomes.

Today’s practice for sleep studies is to tether subjects to an array of needles, electrodes, and intricate monitoring devices, prompting the audience to reflect on the emotional and physical implications of such a scenario. However, comfort is not the only drawback of today’s approaches for collecting patient data. “(In traditional methods) we are getting sporadic data, we get some data points here,” says Katabi. “And another data point later, six months later when the patient takes another test …(this way) we are unable to track the dynamics of symptom evolutions in diseases.”

The alternative is ongoing data collection, with the novel technical solution provided by wireless systems that use ubiquitous radio signals to get patient data on vitals and more, at home or anywhere else, all of the time. Using machine learning, wireless systems can discern individuals’ activities, such as sitting or moving around. Such vital data, when linked with real-life contexts, significantly enhances the accuracy and sensitivity of the well-being assessment and diagnosis of a range of conditions: sleep monitoring, Parkinson’s disease and Alzheimer’s disease.

Sleep can serve as a mirror for various health conditions. Indicators such as early rapid eye movement during sleep stages may hint at depression, and interrupted slow-wave sleep might signal the onset of Alzheimer’s disease. Moreover, these systems have the potential to diagnose Parkinson’s disease, which is currently the world’s most rapidly proliferating neurological condition, through patient breathing analysis.

Current methods of diagnosing Parkinson’s often fall short, identifying the disease only when significant damage has already occurred. As Prof. Katabi notes, “By the time we diagnose the disease, almost 50% to 80% (of the eventual) impairment already exists in the brain, and our dopamine capability. Dopamine is the underlying problem with Parkinson’s. So … by the time we diagnose it, it can be 80% (of) the brain is damaged. And you can understand why it’s very hard to develop drugs for Parkinson’s, or to treat Parkinson’s.”

Offering an alternative, Prof. Katabi’s Emerald system has demonstrated remarkable potential, achieving up to 90% accuracy in preliminary findings. This level of accuracy is based on follow-up data from a comprehensive study involving roughly 7600 patients.

Incorporating continuous time data collection in healthcare using ambient WiFi detectable by machine learning promises an era where early and accurate diagnosis becomes the norm rather than the exception. This technology has the potential to significantly improve treatment outcomes but paves the way for a proactive approach to health management.