Hi Marco, great content—thanks for sharing! Do you apply any transformations to rMSSD besides the logarithm? I've seen some authors compute relative rMSSD as rMSSD/HR, but I'm unsure how to interpret it or which HR value to use (basal vs. instantaneous). Additionally, it would be more intuitive to work with the inverse of HRV so that higher values indicate greater stress. Is this approach recommended? I've come across log(1/rMSSD) in a couple of papers, but I wonder if there's another method I'm missing. Thanks again!
thank you Raimundo. What we call HRV is simply a logarithmic transformation of rMSSD but recently I got more interested as well in the normalized version, as I've seen quite a few instances in which it provides more useful insights (e.g. if HRV is normal but there is a large suppression in HR, then HRV isn't really normal, but suppressed - this can be common when losing weight or changing diet, in my experience). The normalization should be done in the same units, so rMSSD / average RR intervals, which would track similarly to HRV (higher as more parasympathetic). See here: https://marcoaltini.substack.com/p/should-we-normalize-hrv-by-heart - this is now part of HRV4Training Pro (there, for simplicity I called it HRV:HR).
Thanks a lot for your article! I really learnt a lot. I have one question: When I take the LN of 250, this is 5.5 and not 10. Where are my thoughts going in the wrong direction here? Thanks a lot for your answer!
thank you Richard, multiplied by two plus a small adjustment to make it more human friendly (without altering relative changes within individual over time). rMSSD is also available in HRV4Training if you prefer
Thank you very much for your response, Marco. I also looked at the rMSSD and HRV values in your HRV4Training app, and I was wondering about a similar question. When my HRV was at 10, I was wondering about the corresponding rMSSD of 200 instead of the 250, you wrote in this article. Is there any individual or user-specific adjustment for these values, aside from the general transformation of 2*ln(rMSSD)?
correct, in the earlier days we also added an age correction factor, that aimed at making this HRV score more comparable between people. At this point I wouldn't do it this way (as the last thing we should do is compare HRV with others), but it's still there as it doesn't really matter for the typical within individual analysis.
Hi Marco, great content—thanks for sharing! Do you apply any transformations to rMSSD besides the logarithm? I've seen some authors compute relative rMSSD as rMSSD/HR, but I'm unsure how to interpret it or which HR value to use (basal vs. instantaneous). Additionally, it would be more intuitive to work with the inverse of HRV so that higher values indicate greater stress. Is this approach recommended? I've come across log(1/rMSSD) in a couple of papers, but I wonder if there's another method I'm missing. Thanks again!
thank you Raimundo. What we call HRV is simply a logarithmic transformation of rMSSD but recently I got more interested as well in the normalized version, as I've seen quite a few instances in which it provides more useful insights (e.g. if HRV is normal but there is a large suppression in HR, then HRV isn't really normal, but suppressed - this can be common when losing weight or changing diet, in my experience). The normalization should be done in the same units, so rMSSD / average RR intervals, which would track similarly to HRV (higher as more parasympathetic). See here: https://marcoaltini.substack.com/p/should-we-normalize-hrv-by-heart - this is now part of HRV4Training Pro (there, for simplicity I called it HRV:HR).
pretty clear, thanks for the quick reply, saluti!
Thanks a lot for your article! I really learnt a lot. I have one question: When I take the LN of 250, this is 5.5 and not 10. Where are my thoughts going in the wrong direction here? Thanks a lot for your answer!
thank you Richard, multiplied by two plus a small adjustment to make it more human friendly (without altering relative changes within individual over time). rMSSD is also available in HRV4Training if you prefer
Thank you very much for your response, Marco. I also looked at the rMSSD and HRV values in your HRV4Training app, and I was wondering about a similar question. When my HRV was at 10, I was wondering about the corresponding rMSSD of 200 instead of the 250, you wrote in this article. Is there any individual or user-specific adjustment for these values, aside from the general transformation of 2*ln(rMSSD)?
correct, in the earlier days we also added an age correction factor, that aimed at making this HRV score more comparable between people. At this point I wouldn't do it this way (as the last thing we should do is compare HRV with others), but it's still there as it doesn't really matter for the typical within individual analysis.
Awesome Marco! This article answered a lot of questions we had on the subject.