The Emerald device is a WiFi-like box that runs customized machine learning algorithms to learn digital biomarkers from the wireless signals in a patient's home. In this exclusive BWB TV interview filmed at Biotech Week Boston in September 2018, Dina Katabi, MacArthur Fellow and Professor at MIT, explains how it is possible to measure someone's vital signs and movements from a distance without any sensors on their body.
So, you guys have developed this new device. Can you tell me a little bit more about it and what it does?
DK: Of course you have a WiFi box in your home, so imagine if that WiFi box can just sit in the home and monitor breathing, heart beats, gait, sleep, even sleep stages, like when the person is in REM stage versus light or deep sleep. All of these physiological signals, even things related to mental health and the depression-related metrics. Imagine that WiFi box can do all of that without asking the person in the home to wear any sensors on their body - just from distance by analyzing the wireless signals in the environment.
This is exactly what we do in my group at MIT. We invented a WiFi-like box or smart version of your WiFi box, and it sits in the home and just uses the electromagnetic waves, the wireless signals that are in the environment, to analyze and get breathing, heartbeat, gait, falls of elderly people, sleep stages, depression-related metrics etc etc.
How is it possible to measure someone's vital signs and movements from a distance without any sensors on their body?
DK: This is the research that we do in my group at MIT. Anything, any movement that you do - whether you breathe or the pulsing of your blood - everything affects the electromagnetic waves around you. We are surrounded with so many electromagnetic waves and any of those movements affect and change the electromagnetic waves. We have a new smart WiFi box that analyzes those changes and discovers that you took a breath, or this is the pulse rate of your blood or this is the movement of the person, and we can analyze all of those signals.
Now, the real enabler that enables that analysis is advancement in machine learning because, as you can guess, while the physics mean that everything affects the electromagnetic waves, actually extracting those physiological signals from the electromagnetic waves around you is very complex. Advancements in machine learning make it possible to do that analysis.
And what are the most important applications for this technology?
DK: It has many applications, but let me focus on applications in biotech and pharma. So, if you think today of how clinical trials are done - the patient goes to the site perhaps once every two weeks and then during that whole time, the full two weeks, the patient is in his or her home, you have no information about them. So, imagine if you can have a smart WiFi box that sits in the home and collect breathing, heartbeat, movements, sleep, a variety of physiological signals without asking the patient to do anything. Just live their lives. You get that continuous information, which dramatically [improves] the clinical trial, both in terms of efficacy and also in terms of safety.
What about privacy issues?
DK: That's a very important question. Of course, once you start looking at these electromagnetic signals and being able to discover so much about people without even having to touch them, the issue of privacy and how we manage this is very important. So, for us of course, whatever we do in terms of extracting information is based on consent. The patient or the person who's monitored decides what kind of information is extracted from the electromagnetic waves and also what to do with this information. This is one aspect. The other thing is that we stipulate that for the data from anything that is identifiable, that has an identity, the person and both types of data are encrypted and stored in the most secure way.