Let's say that you, like 50 million other people in America alone, suffer from some kind of sleep disorder. A logical step would be to visit a sleep clinic. But is that really the best way to have your sleep patterns analyzed? For one thing, no one really sleeps that well in a new space. For another, the electrodes attached to your skin could wake you up when they pull and pinch as you toss and turn. MIT researchers believe they have a better solution: beaming you with radio waves in your own home.

The new method builds upon previous work conducted at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) that made use of radio frequency (RF) signals to determine emotions and measure walking speed in the home. In both of those cases, radio waves are broadcast toward a person and then reflected back to a receiver. By analyzing the way in which the waves change, the system is able to make determinations about the individual being studied – even down to the level of detecting changes in breathing and heart rates in the case of analyzing emotions.

UPGRADE TO NEW ATLAS PLUS

More than 1,500 New Atlas Plus subscribers directly support our journalism, and get access to our premium ad-free site and email newsletter. Join them for just US$19 a year.

UPGRADE

"It's a smart Wi-Fi-like box that sits in the home and analyzes these reflections and discovers all of these changes in the body, through a signature that the body leaves on the RF signal," said Dina Katabi, MIT professor of electrical engineering and computer science, who led the study.

In applying the RF technique to sleep, the team of researchers from both CSAIL and Massachusetts General Hospital turned to a deep neural network to analyze the radio frequencies captured by the receiver while test subjects slept. The trick was to weed out all of the other radio waves bouncing back from other items in the room and isolate just those coming from a sleeping person. So they developed a new artificial intelligence algorithm that did exactly that.

"The surrounding conditions introduce a lot of unwanted variation in what you measure," said study co-author Tommi Jaakkola, a professor of electrical engineering and computer science at MIT. "The novelty lies in preserving the sleep signal while removing the rest."

The team found that in tests involving 25 volunteers, the system was sensitive enough to determine whether people were in light, deep or rapid-eye-movement levels of sleep with an 80 percent accuracy, an achievement they say is comparable to the EEG systems often used in sleep clinics. But, because the system works without any sensors, it holds a distinct advantage over EEGs.

"Our device allows you not only to remove all of these sensors that you put on the person, and make it a much better experience that can be done at home, it also makes the job of the doctor and the sleep technologist much easier," said Katabi. Katabi also points out that a home-based system means that constant monitoring is possible as opposed to sporadic monitoring in a sleep clinic.

"Imagine if your Wi-Fi router knows when you are dreaming, and can monitor whether you are having enough deep sleep, which is necessary for memory consolidation," she said. "Our vision is developing health sensors that will disappear into the background and capture physiological signals and important health metrics, without asking the user to change her behavior in any way."

Next, the researchers will use their system to study the impact Parkinson's disease has on sleep.

"When you think about Parkinson's, you think about it as a movement disorder, but the disease is also associated with very complex sleep deficiencies, which are not very well understood," Katabi said.

The RF system could also be used to study the way other diseases such as Alzheimer's and epilepsy impact sleep, as well as helping reveal more information about sleep disorders like insomnia and sleep apnea.

The research will be presented at the International Conference on Machine Learning that will be held in Sydney on August 9, while a paper describing the work is available online.

Source: MIT/CSAIL