Researchers have developed a new tool to monitor people for cardiac arrest contactlessly by detecting agonal breathing using a smart speaker such as a smartphone or devices such as the Amazon Alexa.
Using real-world audio of agonal breathing heard on emergency calls for cardiac arrests, researchers trained a support vector machine (SVM) to accurately classify agonal breathing instances. The researchers also trained the SVM to detect interfering sounds such as air conditioning as well as instances of hypopnoea, central apnoea, obstructive apnoea, snoring and breathing captured during sleep studies.
On average, the proof-of-concept tool detected agonal breathing events 97 per cent of the time from up to 20 feet (6 meters) away.
The team envisions the algorithm could function like an app, or a skill for Alexa that runs passively on a smart speaker or smartphone while people sleep. And they now plan to commercialise the tool.
Writing in npj Digital Medicine, the authors said the increasing adoption of commodity smart speakers in private residences and hospital environments may provide a wide-reaching means to realise the potential of a contactless cardiac arrest detection system.