I have received this question a few times since Suunto released their new feature using HRV (and detrended fluctuation analysis, or DFA) to determine exercise intensity, hence this blog post.
Hi Marco, thanks you so much for the interesting write up. I hope this message still reaches you.
I've looked into using DFA in the past, but to be honest this was more because I found the technology interesting, not necessarily because it gave me information I could not obtain on an easier way ;)
Having said that, do you know of any use cases of measuring HRV ('regular' HRV, so not DFA) during exercise? It would seem easier to measure, less prone to error, than DFA. To be honest I don't know what value it would add or what question it would answer ;). However there seem to many tools around that offer the option of measuring it, for example it has been an option on Garmin watches for years, so I am wondering if I am missing something? ChatGPT had some ideas but I am much more interested in your option ;). To be clear, I am talking about measuring while exercising, with a Polar strap.
thanks Tim! Glad there is still a place for human experts :)
regarding HRV during exercise, other indices have been tested, in similar applications to what you see with DFA (i.e. the goal is always to determine exercise intensity or "thresholds"). In terms of artifacts / issues, true that DFA is really prone, but other HRV indices have also the same problem, a single artifact over a few minutes makes the data useless, and this tends to happen more when we add physical movement.
This being said, the issue is not really the one of data quality when it comes to these other indices, but the "lack of range", meaning that rMSSD would be somewhere between 2-4 milliseconds for most people during exercise, with almost no change during an incremental test, despite dramatically different heart rates (as heart rate gets higher than the internal firing rate of the sinoatrial node, HRV is basically zero, and this is always the case during exercise due to higher sympathetic outflow: https://marcoaltini.substack.com/p/the-physiology-of-heart-rate-variability). Frequency domain features have also been tested, but eventually, they have similar limitations, and have no use if you don't a maximal test frequently to re-calibrate them (because the absolute values are not useful), hence eventually, it seems they don't provide more insights than just looking at heart rate (which is also less noisy, more reliable, etc.).
Also AlphaHRV has a great connect IQ app, where you can even configure the cut-offs for the thresholds (eg. 0.75), so this opens the possibility to test and adjust for future training.
fantastic insight, absolutely love it. I would've 100% fallen for the marketing. This helps tremendously to better understand the areas (e.g. DFA) where I am not knowledgable enough to critique the estimates properly. Thank you! The chain is only as strong as its weakest link. Still following
Very interesting thoughts here. Objective and thorough.
I've been using the ZoneSense for a few runs now and find the data interesting. As a hrv4training user for a number of years now I tend to rely on that app as my "go to" for how I train on a particular day. It's been great, particularly with the large number of metrics available (good work! :-) ) ZoneSense (ZS) is another way to look at my running I guess, in real time, and how my HRV is variously moving from aerobic/anaerobic etc.
However I have found that in some instances, running up hills, and thus generating a high heart rate has ZS still in the aerobic zone and not as I would have thought into anaerobic. Any thoughts on why this would be the case?
thank you Jeff, much appreciated! Regarding the issue you mention, I think it could either be an artifact (DFA is sensitive to a few things, even just changing breathing pattern) or simply a problem linked to e.g. the duration of the change in effort (which might be too short for the math to catch up). I would not take it as a sign that you are not working hard :) (especially considering you report a higher heart rate).
One thing I'm trying to look at is using AlphaHRV to moderate my interval training, more specifically if it can be used to adjust the recovery time and then the cool down phase of the workout. There's a delay to the metric itself and also in your physiology, but this could helpful in understanding when you are ready for your next interval. You don't want to relax too much, but you also don't want to keep your engine too much revved up. Also when doing the cool down after the intervals, I realized that my recovery pace really needs to be low, lower than zone 2, otherwise AlphaHRV will keep low, sometimes even dipping below 0.5, your engine is still running fast, even though you are barely exercising.
Jist of it is not everyone is going to be at 0.75. You take your number from early in the ramp and the mid point from that and 0.5 is a good indicator for intensity. I myself train at 1-1.1 dfa a1. I also find it pretty amazing for detecting fatigue - you notice suppressions immediately on a ramp warm/up even when hrv4training protocol HRV hasn't detected anything. That way you can either lower intensity or shorten duration or both.
PS - AlphaHRV on Garmin is free - the author has some forthcoming papers on it and the same app is going to be used with a continuos lactate monitor that's releasing soon. So soon we'll truly know if dfa lines up with lactate! https://x.com/inigo_tolosa_12/status/1848455884126433441
thanks Dush, I’m glad Rogers eventually came to his senses :)
regarding lactate, I think (and Thomas would agree) that this whole story is pointless: we are looking at cardiac activity, and not metabolic activity. Everything is correlated (heart rate, HRV, blood pressure, lactate, intensity, etc.) but this doesn’t mean that one should be a proxy for the other, unless re-calibrated every few weeks. As I mention in the blog, the interesting part is how lactate and heart rate (or HRV) change in different ways based on how we train (and detrain), and therefore it is meaningless to use one to guess the other (how would I know that my heart rate at LT1 is now higher, if I wasn’t measuring both my heart rate and lactate? - rhetorical question). If HRV during exercises, tuned to your individual thresholds, helps you manage load / fatigue, that’s great, but that’s all there is, it’s not a marker of metabolic thresholds as lactate, which is also fine, just a different way to look at various aspects that relate to fatigue and performance.
100% agree - for me it's an autonomic load tool. "How far can I push my autonomic system today", "how much training does my autonomic system let me do today".
Btw - after using hrv4training for a number of months it finally gave me the confidence to ditch wearing my Garmin continuously & stopping relying on all day stress/overnight HRV. My mind is much freer & I'm not so obsessed with the metrics. I moved to the AW Ultra 2 which I like because of the ability to do morning measurement with hrv4training & the vitals app which answers the simple question "Am I brewing an illness that I can't feel, so should I take it easy today?". Much easier than trying to read the tea leaves with all of Garmins metrics.
Hi Marco, thanks you so much for the interesting write up. I hope this message still reaches you.
I've looked into using DFA in the past, but to be honest this was more because I found the technology interesting, not necessarily because it gave me information I could not obtain on an easier way ;)
Having said that, do you know of any use cases of measuring HRV ('regular' HRV, so not DFA) during exercise? It would seem easier to measure, less prone to error, than DFA. To be honest I don't know what value it would add or what question it would answer ;). However there seem to many tools around that offer the option of measuring it, for example it has been an option on Garmin watches for years, so I am wondering if I am missing something? ChatGPT had some ideas but I am much more interested in your option ;). To be clear, I am talking about measuring while exercising, with a Polar strap.
Thanks in advance for your time to read, thanks,
Tim
thanks Tim! Glad there is still a place for human experts :)
regarding HRV during exercise, other indices have been tested, in similar applications to what you see with DFA (i.e. the goal is always to determine exercise intensity or "thresholds"). In terms of artifacts / issues, true that DFA is really prone, but other HRV indices have also the same problem, a single artifact over a few minutes makes the data useless, and this tends to happen more when we add physical movement.
This being said, the issue is not really the one of data quality when it comes to these other indices, but the "lack of range", meaning that rMSSD would be somewhere between 2-4 milliseconds for most people during exercise, with almost no change during an incremental test, despite dramatically different heart rates (as heart rate gets higher than the internal firing rate of the sinoatrial node, HRV is basically zero, and this is always the case during exercise due to higher sympathetic outflow: https://marcoaltini.substack.com/p/the-physiology-of-heart-rate-variability). Frequency domain features have also been tested, but eventually, they have similar limitations, and have no use if you don't a maximal test frequently to re-calibrate them (because the absolute values are not useful), hence eventually, it seems they don't provide more insights than just looking at heart rate (which is also less noisy, more reliable, etc.).
Thanks Marco! That makes sense, much appreciated ;)
Also AlphaHRV has a great connect IQ app, where you can even configure the cut-offs for the thresholds (eg. 0.75), so this opens the possibility to test and adjust for future training.
great to hear, thanks Pedro!
This is awesome - thank you! I've been reading about DFA from AI Endurance, who seem to have similar perspectives to you :)
fantastic insight, absolutely love it. I would've 100% fallen for the marketing. This helps tremendously to better understand the areas (e.g. DFA) where I am not knowledgable enough to critique the estimates properly. Thank you! The chain is only as strong as its weakest link. Still following
Hi Marco,
Very interesting thoughts here. Objective and thorough.
I've been using the ZoneSense for a few runs now and find the data interesting. As a hrv4training user for a number of years now I tend to rely on that app as my "go to" for how I train on a particular day. It's been great, particularly with the large number of metrics available (good work! :-) ) ZoneSense (ZS) is another way to look at my running I guess, in real time, and how my HRV is variously moving from aerobic/anaerobic etc.
However I have found that in some instances, running up hills, and thus generating a high heart rate has ZS still in the aerobic zone and not as I would have thought into anaerobic. Any thoughts on why this would be the case?
Regards,
Jeff
thank you Jeff, much appreciated! Regarding the issue you mention, I think it could either be an artifact (DFA is sensitive to a few things, even just changing breathing pattern) or simply a problem linked to e.g. the duration of the change in effort (which might be too short for the math to catch up). I would not take it as a sign that you are not working hard :) (especially considering you report a higher heart rate).
One thing I'm trying to look at is using AlphaHRV to moderate my interval training, more specifically if it can be used to adjust the recovery time and then the cool down phase of the workout. There's a delay to the metric itself and also in your physiology, but this could helpful in understanding when you are ready for your next interval. You don't want to relax too much, but you also don't want to keep your engine too much revved up. Also when doing the cool down after the intervals, I realized that my recovery pace really needs to be low, lower than zone 2, otherwise AlphaHRV will keep low, sometimes even dipping below 0.5, your engine is still running fast, even though you are barely exercising.
Marco have you seen Bruce Rogers team's latest paper? It's about customising the DFA a1 thresholds. http://www.muscleoxygentraining.com/2024/09/improving-hrvt1-agreement-ijspp-102024.html
Jist of it is not everyone is going to be at 0.75. You take your number from early in the ramp and the mid point from that and 0.5 is a good indicator for intensity. I myself train at 1-1.1 dfa a1. I also find it pretty amazing for detecting fatigue - you notice suppressions immediately on a ramp warm/up even when hrv4training protocol HRV hasn't detected anything. That way you can either lower intensity or shorten duration or both.
PS - AlphaHRV on Garmin is free - the author has some forthcoming papers on it and the same app is going to be used with a continuos lactate monitor that's releasing soon. So soon we'll truly know if dfa lines up with lactate! https://x.com/inigo_tolosa_12/status/1848455884126433441
thanks Dush, I’m glad Rogers eventually came to his senses :)
regarding lactate, I think (and Thomas would agree) that this whole story is pointless: we are looking at cardiac activity, and not metabolic activity. Everything is correlated (heart rate, HRV, blood pressure, lactate, intensity, etc.) but this doesn’t mean that one should be a proxy for the other, unless re-calibrated every few weeks. As I mention in the blog, the interesting part is how lactate and heart rate (or HRV) change in different ways based on how we train (and detrain), and therefore it is meaningless to use one to guess the other (how would I know that my heart rate at LT1 is now higher, if I wasn’t measuring both my heart rate and lactate? - rhetorical question). If HRV during exercises, tuned to your individual thresholds, helps you manage load / fatigue, that’s great, but that’s all there is, it’s not a marker of metabolic thresholds as lactate, which is also fine, just a different way to look at various aspects that relate to fatigue and performance.
100% agree - for me it's an autonomic load tool. "How far can I push my autonomic system today", "how much training does my autonomic system let me do today".
Btw - after using hrv4training for a number of months it finally gave me the confidence to ditch wearing my Garmin continuously & stopping relying on all day stress/overnight HRV. My mind is much freer & I'm not so obsessed with the metrics. I moved to the AW Ultra 2 which I like because of the ability to do morning measurement with hrv4training & the vitals app which answers the simple question "Am I brewing an illness that I can't feel, so should I take it easy today?". Much easier than trying to read the tea leaves with all of Garmins metrics.
thank you so much Dush, I am really glad to read this 🙏