MIT researchers are back once again to deliver us new technologies, this time revolving around sleep and dreaming. They have developed a wireless, artificially intelligent sensor which is able to detect all sleep’s stages, including REM or rapid eye movement stage that is connected to dreaming. This new system is on its way to change how clinicians diagnose sleep disorders and other health complications.
The system involves the use of wi-fi radio signals. There have been used previously for various intents and purposes due to their sensitivity, such as seeing through walls, tracking a person’s movements and mood, as well as for monitoring the breathing and heart rates of sleeping children. This last way of use was what inspired the researchers from MIT to develop a sensor that detects the three big phases of sleep – light (which strengthens cognitive processes), deep (associated with memory consolidation), and REM (associated with the most intense dreaming).
Dina Katabi and her team from MIT partnered up with researchers from Massachusetts General Hospital and used an artificially intelligent algorithm called deep neural network. The system works by analyzing the small changes in the frequencies of reflected radio frequency (RF) signals as they bounce off a sleeping patient and then connecting them to the various phases of sleep, which also includes wakefulness. The system’s main advantage compared to others’ sleep tracking systems is that this one is able to weed out superfluous movements and background noise, focusing on pulse and breathing rate instead.
The low-power RF signal-emitting sensor is as big as a laptop and is supposed to be hanged next to an asleep patient. 25 volunteers were put to various tests that showed the technology having the 80% accuracy rate, which is the same as EEG measurements. This device’s advantage, though, is that it isn’t clunky and doesn’t use any wires or hookups, making it far less intrusive for patients that are trying to sleep.
Katabi said in a statement that their device allows people not only to remove all the sensors that are usually put on a person and make it a more pleasant experience to patients and making it possible to do at home, it also makes the job easier for doctors and sleep technologists because they don’t have to go through data and label it manually.
The researchers from MIT presented their paper on 9th August, at the International Conference on Machine Learning. They said that they intend to use their device to understand further how Parkinson’s disease affects sleep. Among future possible uses of the device are patients with Alzheimer’s, insomnia, and sleep apnea. The sensor is also safe to use on patients that suffer from epileptic seizures that are particularly tough to notice during sleep.