Apple looking to boost HRV features

This story points out where Apple need to go next with HRV-related features on the Apple Watch.

More: HRV4Training

Marco Altini (sports science researcher and creator of market-leading HRV tool HRV4Training) has noticed that several Apple staff have been poking around his work and so he tweeted some helpful suggestions to them on what to do next. Here they are, with my thoughts against each point and the original tweet, further below.

1. Move from SDNN calculation to industry standard rMSSD – this is mostly a suggestin as the SDNN and rMSSD calculations give broadly the same types of results. I always thought that SDNN was marginally better in how it handled more extreme values but would definitely make the HRV world easier if everyone, everywhere is talking about the same thing when they refer to HRV
2. Random Sampling – I don’t think Apple’s samples are random as such, although they appear that way. They are certainly saved more frequently at night than during the day, however, the period between readings does indeed vary and therefore might appear random. My understanding was that Apple only took HRV readings when there was less wrist movement and that they also might exclude certain other readings for quality reasons. So Marco might want these ‘bad’ readings included? I’m not sure about that but if more frequent readings could be taken that would be great despite the battery hit. (I’m assuming the reader has aFib History enabled as this boosts the frequency of recorded results)
3. Manual Spot checks – As Marco points out you can already do this with the Breathe app but I’d certainly agree that virtually no one knows that the Breathe app actually creates an HRV data point. It could be a bit clearer.
4. Introduce a normal range – Yes this would be great. If your HRV and mine today are both ’50’, then whose is best? The idea of a normal range is used by HRV4Training, Garmin and others and it contextualises the 50 in respect to the normal. The range is determined by a moving average. (50 could be above my normal range, below yours or within someone else’s normal range.)
5. Provide signal quality – As I alluded to above, I believe that Apple already assesses the signal quality and the likelihood of getting good signal quality. However, I would agree that it would be useful for app developers to also know this information.