Sleep tracking isn’t something that researchers experiment in the lab anymore. With apps, wearables, and peripherals it has moved into your bedroom. However, those devices use movement or your device’s movement to monitor whether you’re sleeping or awake, not your sleep stages. Now, researchers at MIT have used an artificial intelligence algorithm and radio waves to analyze someone’s sleep stages without requiring any physical sensor. The AI analyzes the radio waves around the person and translates the assessment into sleep stages, such as light, deep, or rapid eye movement (REM).
“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,” said study leader Dina Katabi in a statement. “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.”
First to be more accurate
Although there are other systems that monitor sleep using radio signals, this MIT system is the first to claim a high rate of accuracy (80 percent) as measured against electroencephalography (EEG) recordings. While tracking sleep, the RF signals congregate some incoherent information that may confuse the existing algorithm. So, the researchers came up with new algorithms known as deep neural networks that help extract and analyze information from complex datasets. Using the MIT-written AI algorithm this new sleep monitoring system analyzes the data and translates them to valuable sleep data. The researchers will present their study at the International Conference on Machine Learning on August 9.
Opening a new door to study sleeping
More than 50 million people in the USA have sleeping disorders and diseases like Alzheimer’s, Parkinson’s, as well as epilepsy can have sleep disruptions that are difficult to detect. Diagnosing those patients using a traditional method includes lots of sensors that can further disrupt their sleep. As this new method doesn’t require any sensors, it can easily be used to diagnose those patients without disturbing them. The researchers are already thinking about using this new technique to study how Parkinson’s disease affects sleep.