Apple could soon estimate your breathing rate using an accessory you might not have thought of: your AirPods. This could be good news to monitor the occurrence of certain diseases and improve the monitoring of sports activity.
It was estimated with sufficient precision the respiratory rate by exploiting the microphone of AirPods or an Apple Watch with adequate accuracy. A study conducted by its Machine Learning Research section shows that the firm can measure this vital sign by simply analyzing the “sound of breath obtained thanks to microphones integrated into wearables.”
Your AirPods could soon make sure you breathe well
Thanks to the proximity of its microphones to the mouth, the AirPods seem ideal to engage in this analysis, even if the Apple Watch could also be used thanks to its built-in microphone. In this case, listening to the breath is done in adults before and after physical exertion. It is limited only to a few short audio excerpts dedicated to analysis.
This analysis, which AI should ultimately carry out, would allow, for example, to detect the appearance of heart or respiratory diseases in time. It could also be used to improve the monitoring of sports activity over the long term, a fortiori among users of the Apple Watch, which already provides many health data such as heart rate or blood oxygen level (with the Apple Watch Series 6 and its new sensors) especially during physical activity.
A technique for measuring respiratory rate still under study
As MyHealthyApple reports, this measure is still limited to the experimental stage. In this case, the study conducted by Apple was conducted among 21 users. Their breathing frequencies were measured in a fairly rudimentary way by counting audible breaths and exhalations cycles. Suppose the device used for recording the breath is unfortunately not specified. In that case, we learn, on the other hand, that a sophisticated neural network has been used to improve, among other things, the clarity of the signal.
Review of the operation? The study shows that the respiratory rate can be estimated through these eavesdropping with a correlation coefficient of concordance (CCC) of 0.76 and a mean square error (EQM) of 0.2, it reads. In other words, analysis based on the breath recorded by the microphones of a simple wearable is viable to estimate the respiratory rate with sufficient accuracy.
Towards more efficient mics on the next AirPods?
However, there is no assurance that the encouraging findings of this study will lead to a feature being added later in Apple’s software toolkit. That said, this research is interesting in more ways than one first of all because they establish that it is possible to calculate a respiratory rate using a simple microphone efficiently and under customary conditions of use, both indoors and outdoors. On the other hand, they allow the test of Apple’s neural networks to improve the sharpness of a signal picked up in situations sometimes not optimal.
One can quite imagine that Apple will do something from this research. The brand may later add better microphones to its AirPods to improve breathing listening before offering a potential feature on iOS or watchOS. It will also be necessary to find an algorithm capable of accurately interpreting the respiratory eavesdropping performed.