Motion Detection and Machine Learning Device Used in Study of Chronic Itch Patients

    By Kathleen Berger, Executive Producer for Science & Technology

    A completely new approach to medicine is all in a box that looks like a WiFi router on your wall. It’s a touchless sensor that picks up on every motion, a machine learning platform for health analytics. The device is called Emerald.

    For Dr. Brian Kim, MD, Emerald is a gem his lab couldn’t pass up, as he is most familiar with diseases and conditions that can be difficult to diagnose and treat. As co-director of the Center for the Study of Itch and Sensory Disorders at Washington University School of Medicine in St. Louis, Dr. Kim embraced the new technology.

    “My laboratory focuses on very specific biologic mechanisms that drive itch. But then my other job is to see patients who have chronic itch, and one of the things that’s held us up in the field of medicine, the field does not do a good job of recognizing conditions that subjectively, seemingly subjectively, affect people,” said Kim.

    Understanding chronic itch and how or why someone may experience itch can be complicated and when it comes to drug therapies and clinical trials, measuring success can be complicated as well.

    “Companies often ask, ‘ how do we know that the itch got better? What if they’re lying? What if they’re just stressed out?’” explained Kim. “The million dollar question right now in the field of clinical, chronic itch is can we measure itch instead of the patient having to tell us subjectively how they feel? Can we just measure itch and look at the toll that this symptom has in their life?”

    The associate professor of medicine and dermatologist believes in this new tool for measurement, so Kim created a clinical trial using Emerald for itch patients, mounting the technology to patient walls.

    Emerald is based on award-winning wireless and machine learning research from Massachusetts Institute of Technology (MIT) using wireless signals. The software allows for remote monitoring of patients’ movements while awake and asleep. Emerald’s sensor and AI algorithms analyze radio waves to reverse engineer motion, even the smallest of movements.

    “One of the issues with medicine is that I’ll see a patient maybe every two to three months, but I have no idea how they were doing every other day between our visits,” he said. “Emerald became a perfect fit for itch studies because of the ability to detect things like movement. We wanted to detect scratching. Can we also have the machine learning detect other things about itching and the impact of itching and scratching on patients that we did not know? For instance, we’re detecting that patients who end up scratching at night, it seems that they don’t sleep as well, unbeknownst to even them. So in other words, the pattern of their sleeping changes. As we know, that can have a huge impact on their overall health.”

    “Machine learning can figure out, perhaps hypothetically, how they’re sleeping, whether they’re depressed, whether patients are anxious, whether there are other kinds of comorbidities. So we decided to do a trial to answer these multiple questions, the primary objective being – can we measure it?”

    Kim said preliminary clinical trial results are promising.

    “The more Emerald learns from each and every movement, the better it becomes at diagnostics,” Kim explained. “This device actually uses a lot of artificial intelligence. So the reason why this is very powerful is not just that you can detect breathing and movement without even being in the room, but that the computer can log all that data and over time say, ‘hey, we are seeing a pattern.’”

    Emerald is designed to monitor vital signs and other health metrics, helping doctors know when a patient needs to be seen or go to the hospital.

    “What we’re able to do is train the device,” he explained “It’s not that the device is going to learn so much from just one patient, but it’s going to learn so much from many patients, perhaps thousands of patients. And as it gets further along, it gets more and more accurate. “

    “It can innovate and detect something that even if I, as a clinician, had seen a million patients would perhaps never have even found that pattern.”

    Kim said helping itch patients is just one application.

    “Any disease in which movement can be picked up in a very subtle way over a long period of time, this device could be potentially useful for.”