Researchers at the University of Washington have developed a new AI tool to monitor people’s heart condition. The tool is especially designed to prevent increasing cases of cardiac arrest, most commonly observed in out-of-hospital settings.
The study focuses on smart speaker devices, like Google Home or Amazon Alexa or even smartphones that will detect agonal breathing and will thus call for help. Agonal breathing is a common phenomenon among people suffering a cardiac arrest; it is a case when a person stops breathing or gasps for breath due to low oxygen levels.
The proof- of- concept tool was developed by detecting real gasping sounds of agonal breathing instances collected from 911 calls. According to reports, the tool could identify 97% agonal breathing events from up to 20 feet away from the individual. The uniqueness of the guttural gasping sound that an individual makes during agonal breathing is one of the reasons why the breathing type is recognized on the audio device.
Using a database of 911 calls from Seattle’s Emergency Medical Services, bystanders recorded sounds of agonal breathing by putting their phones close to a patient’s mouth. The research team collected records from about 162 calls between 2009 and 2017. Moreover, they extracted 2.5 seconds of audio at the start of each agonal breath which produced a total of 236 clips. The result was captured on different smart speaker devices like Amazon Alexa, iPhones and Samsung Galaxy S4. Moreover, the team also produced an algorithm to make sure that any other kind of breathing apart from agonal breathing was not detected.
“We don’t want to alert either emergency services or loved ones unnecessarily, so it’s important that we reduce our false positive rate,” said a spokesperson for the research.
Furthermore, the team envisions to use the algorithm as an app for smart speaker devices like Alexa. “This could run locally on the processors contained in the Alexa. It’s running in real time, so you don’t need to store anything or send anything to the cloud,” as reported by a research expert. “Right now, this is a good proof of concept using the 911 calls in the Seattle metropolitan area,” he said. “But we need to get access to more 911 calls related to cardiac arrest so that we can improve the accuracy of the algorithm further and ensure that it generalizes across a larger population.”
The study was published on July 19 in npj Digital Medicine. According to reports, the study will be further commercialized through UW spinout, Sound Life Sciences, Inc.