Many people who suffer from autism experience periods when they become aggressive, something that tends to come out of the blue. Those taking care of such folks have a hard time managing such episodes, particularly because there’s no warning and there’s no time to take mitigating actions.
Now, researchers at Northeastern University have developed a system, which relies on a wrist-worn device, that can predict aggressive episodes a minute before they become apparent with nearly 90% accuracy.
The wearable device collects information about the patient’s heart rate, level of sweating, skin temperature, and body movement. This information is collected by a computer and processed to detect the digital biomarkers associated with oncoming emotional outbursts.
In a multi-day study involving 20 kids who frequently experience aggressive periods, the researchers gathered the necessary data to identify signs that arise before the outbursts. This was then used to create a sort-of search algorithm that spots these signs, something that the latest research demonstrated was approximately 84% accurate at a minute before the outburst.
The team is already working to make their technology more accurate and able to predict the outbursts sooner, giving caretakers more time to prepare and help make the situation safer and calmer.