The Apple Watch may be the most popular smartwatch on the planet, but it still lags behind competitors from Fitbit and Samsung when it comes to sensors. However, a new study from Frontiers in Digital Health (via MyHealthyApple) shows that it might not need them.
In a pilot study that tracked 33 mostly white female participants, the researchers were able to use heart rate variability data from the ECG sensor on the Apple Watch Series 6 combined with machine learning techniques to develop a stress prediction tool. The study found that a 30-second ECG reading delivered instant feedback on their stress levels with an accuracy of 52 percent to 64 percent, compared with 60 percent to 80 percent for “state-of-the-art accuracy for stress detection in real-life settings.”
Those are impressive results that relied solely on the ECG sensor. Those results would presumably be higher when using the new temperature sensor on the Apple Watch Series 8 and Apple’s own stress algorithms due to the relationship between stress and skin conductance changes.
Back in 2020, Fitbit released its new Sense watch that had, among other things, a new continuous ElectroDermal Activity (EDA) sensor that tracked tiny electrical changes on your skin to help identify when users were experiencing stress. It was hailed as an advancement over the Apple Watch, which had only just caught up with Fitbit’s blood-oxygen sensor from five years earlier. Using a dedicated EDA app, users need to hold their palms over the screen for at least 2 minutes to get a reading.
The researchers called the results “very promising,” though cautioned that the Apple Watch “lacks the predictive power to accurately predict the ‘stress’ states as of yet.”
