“Alexa, do I have atrial fibrillation?” “I’ll listen to your heartbeat now to check for irregular rhythms…” 

That’s a conversation you might have with your smart speaker soon if you have concerns about any type of irregular heart rhythm. Researchers from the University of Washington (UW) have designed technology that allows a smart speaker like Amazon Alexa or Google Home to monitor heart rhythm without physical contact.

Heart rhythm disorders — also called cardiac arrhythmias — can occur when the heart’s electric pulses function abnormally, making heartbeats occur at uneven intervals. Arrhythmias can be too fast, too slow, or just out of sync; examples of common arrhythmias include atrial fibrillation, atrial flutter, and ventricular fibrillation. 

A device that provides easy heart rate monitoring could help prevent dangerous health complications such as stroke, heart failure, and cardiac arrest. A smart speaker equipped with this ability could benefit both users with a known cardiac condition and those experiencing symptoms of heart irregularity. 

Building on UW’s previous work designing smart speaker tools that monitor breathing for signs of health risk, the research team developed a new machine-learning algorithm. Machine-learning allows the tool to identify individual heart rhythm patterns and recognize abnormalities specific to the user. 

To “listen” to a heartbeat, the tool sends inaudible sound waves towards a user sitting directly in front of the speaker. It then measures how the sound changes when reflected back from the user’s body. Machine-learning also helps the system differentiate between the minute differences between sounds of the chest moving during a heartbeat and other sounds, such as breathing.

For now, the tool provides spot checks: monitoring heart rate only when asked. The research team plans to develop a continuous monitoring tool capable of identifying abnormal heart rates in sleeping patients without the use of wires and electrodes. Continuous no-contact monitoring has applications beyond cardiac care, such as sleep studies and remote monitoring of sleep apnea patients.

Tests of a prototype showed the tool identified regular and irregular heart rhythms with similar accuracy to a standard wired heartbeat monitor. The research team published their findings in the March 2020 edition of Nature: Communications Biology.