HRV Explained: The Metric You Are Reading Wrong

HRV has become one of the most talked-about metrics in health and performance, but it’s also one of the most misunderstood.

In this episode, we sit down with Dr Marco Altini, data scientist, endurance coach, and founder of HRV4Training, to explore what heart rate variability actually reflects in the body, and how it should (and shouldn’t) be used.

With a background spanning data science, physiology, and wearable technology, Marco brings a unique perspective on how physiological data can inform stress, recovery, and long-term adaptation. His work has influenced elite athletes, coaches, and increasingly, how HRV is applied in general health and longevity.

We break down the foundations of HRV, how it relates to the autonomic nervous system, and whether it should be viewed as a marker of recovery, resilience, or overall system health.

A key focus of the conversation is interpretation. Marco explains why “higher HRV is always better” is a flawed concept, the importance of understanding your individual baseline, and why daily fluctuations often matter far less than long-term trends.

We also explore the role of wearable technology — from ECG to PPG devices — and whether consumer tools are now accurate enough to guide decision-making. As an advisor to Oura, Marco shares insight into how these technologies are evolving, and the risks of oversimplifying complex physiology into single scores.

Finally, we discuss the practical side: what actually impacts HRV, how lifestyle factors like sleep, stress, and alcohol influence your data, and how to use HRV in a way that genuinely supports healthspan and performance.

🎙️ In This Episode, We Cover

• What HRV actually measures in the body
• HRV and the autonomic nervous system
• HRV as a marker of health, resilience, and longevity
• Why “higher HRV is better” is misleading
• The importance of baseline, normal range, and trends
• Daily vs long-term HRV: what really matters
• Common mistakes when interpreting HRV data
• Which metrics are actually useful (and which aren’t)
• Wearables explained: ECG vs PPG vs camera-based
• Are consumer devices accurate enough?
• Morning vs overnight HRV tracking
• How sleep, stress, and alcohol impact HRV
• Can you meaningfully improve your HRV?
• The risks of over-relying on wearable data
• Practical ways to use HRV for health and performance

🎯 Who This Episode Is For
• Anyone using or considering wearable health technology
• People interested in recovery, stress, and longevity
• Athletes and coaches using HRV for performance
• Individuals looking to better understand their physiology
• Clinicians and practitioners interested in data-driven health

🎙️ Powered by Mayfair Health

At Mayfair Health, we specialise in recovery, performance, and proactive wellness. If you’re navigating ongoing symptoms, optimising your health, or looking to better understand your body, our multidisciplinary team is here to help.

🔗 Website: https://www.mayfairhealth.co.uk
📩 Contact: info@mayfairhealth.co.uk
📞 Phone: 020 3985 1500
📱 Instagram: @mayfairhealth

Transcript

SPEAKER_00 0:00

Welcome to another episode of Move Well, Live Well, Perform Well. Uh I'm Simon Gilchrist, and I'm super excited today to be able to geek out a little bit about physiology, about HRV, about performance data, and about how this relates to human function and recovery and what HRV and some of those parameters actually mean in terms of recovering. Um does it signify recovery, or does it signify stress resilience? And what that does the reading to our body. So I'm joined here by Dr. Marco Altini. He's a scientist, a data expert, and endurance coach. He's best known as the founder of HRV for training, one of the most widely used platforms globally for heart rate variability, which we'll we'll we'll we'll term HRV tracking and recovery analysis. He holds a PhD in data science alongside degrees in computer science, engineering, and human movement sciences, and works at the intersection of physiology, wearable technology, and performance. Over the past decade, Marco has focused on understanding how physiological data reflects stress, recovery, and long-term adaptation, particularly through HRV. His work has been used by elite athletes, Olympic teams, and coaches, and has increasingly influenced how HRV is applied in general health, resilience, and longevity. He also works as an uh an advisor to Aura, one of the leading wearable technology companies, giving him direct insight into how consumer devices measure, interpret, and present physiological data. He's particularly known for challenging simplistic interpretations of wearable metrics, emphasizing that the real value lies not in a single score, but in understanding individual baselines, normal ranges, and long-term trends. Well, Marco, welcome. That's quite a mouthful, all of that. You are clearly a super smart individual.

SPEAKER_01 2:01

Thank you, Simon. And uh yeah, I hope I will uh you know live up to the expectations you're set.

SPEAKER_00 2:08

I I'm I'm I'm sure you will. So if we can just dive right in, Marco. At a physiological level, what is HRV actually capturing in the body?

SPEAKER_01 2:18

Yeah. HRV is technically speaking a measure of the variability between heartbeats. So our heart does not beat constant at a constant frequency, or there is not a constant time between consecutive beats, right? So even if we are sitting here and our heart rate is 60 beats per minute, it's not exactly every second that we have a beat, but there is always some variability. And the more important thing there is that this variability is not random, right? It is due to something, and that something is really what we are after, which is basically the impact of the nervous system in response to stress on our heart rhythm. So as we face a stressor, the nervous system has a response, and that response goes to modulate heart rate. People are familiar probably with feeling maybe their heart rate racing, right? A higher heart rate as they feel particularly stressed for something. And HRV is another way to look at this. HRV typically goes the opposite direction, so it would be reduced in a situation of higher stress, while heart rate would be a bit elevated. And that's what why we look at it. Basically, it's a way to get to look at the nervous system response to stress, non-invasively, just looking at heart rhythm.

SPEAKER_00 3:43

Okay, cool. So is it is HRV really a measure of the autonomic nervous system? And or is it more about a broader system sort of resilience?

SPEAKER_01 4:00

Yeah, it's a great question, and I think there's no easy answer. Meaning that when we look at HRV, sure, it is impacted by the nervous system. So it is eventually the vagus nerve and cardiac vagal activity that we capture, as that's what is modulating heart rhythm. But at the same time, the nervous system is responding to anything, any sort of stressor that we face, right? Internal stresses, psychological stresses, the environment and you know, physical activity, anything that we do. So when we look at HRV over time and we measure it in a certain way, which allows us to interpret it in a certain way, and we will discuss later what that means. But this is to say that as we look at it, measure it, measuring it at the right time, then it gives us also an overview of our overall capacity to handle the stresses that we are facing, and also to assimilate additional stresses in the future. So it does give us that kind of bigger picture perspective on our stress response, but that is only if we look at it under certain circumstances.

SPEAKER_00 5:14

And we'll probably touch on what those circumstances are in a little bit in terms of the time frames, especially. But so is HRV a measure of our parasympathetic nervous system? Is it a reliable indicator to, you know, as we know that the higher your HRV state, um perhaps the more relaxed the system is, or the lower the HRV measure, the the less your parasympathetic nervous system is is heightened and your sympathetic nervous system is is more active. Is that an accurate way of looking at it?

SPEAKER_01 5:59

Yeah, I think uh we look in particular at the parasympathetic nervous system when talking about HRV, just because the parasympathetic nervous system acts at a high frequency that is captured by the way we compute HRV when measuring it over a short time frame. Then in the past, it was believed that by looking at HRV in different ways, you could capture how the two branches of the nervous system impact us, so the parasympathetic and the sympathetic. So that view is not current anymore, meaning that we now understand that some of those parameters do not reflect the activity of one branch or the other, and it's not so simple, they are more of a global overview of how the different branches impact heart rhythm. And this is all to say that still it doesn't mean that we are looking only at something like the parasympathetic nervous system because, of course, if we are more active and you know we are moving around or exercising, and our heart rate is much higher, there is higher sympathetic activity, and HRV at the same time is quite suppressed, and it's not only due to maybe parasympathetic withdrawal during that time. So it's always acting all at the same time. And in most cases, the two branches are you know acting in opposite directions, but not always, and as such, we cannot really oversimplify that way, but still, typically when looking at HRV, we say we are looking more closely at the activity of the parasympathetic nervous system, and as such, when we are in a more relaxed or rested state, which typically means the parasympathetic nervous system is more active, then we will have also higher HRV with respect to what are our typical values and our historical values. While when we are more stressed, the parasympathetic nervous system will be less active and our heart rate will be a bit higher and our HRV a bit lower.

SPEAKER_00 8:06

Yeah, okay, I think that makes sense. Okay, good. Marco, in some scientific journals, they often r reference um an abbreviation called RMSSD, which I believe is the root measurement or root mean of the the difference between your heartbeats. Is that essentially a proxy for HRV, or is that um uh is some of the data cleaned up by wearables to get an HRV score?

SPEAKER_01 8:45

Right. So indeed HRV, meaning heart rate variability or just you know the variability in heartbeats can be computed in many different ways. And these are you know what we could call HRV indices, and RMSSD is one of those and is the most commonly used at the moment. So the both the both I would say in the scientific literature and in consumer products, there has been over the years an agreement more or less on which metrics are the most useful, and RMSSD became the most useful. So we use it now as to mean HRV. It is the same thing, meaning it is a way to compute HRV, and it is also what we think is the most useful because because of the math, it looks at these high frequency changes representative of parasympathetic activity, and so the math and the physiology go well together because indeed the parasympathetic nervous system has this high-frequency impact on heart rhythm, and as such, we basically use that metric across the board now to look at HIV and both wearables or other devices that even look at HIV the way we should look at it, so to speak, meaning the actual electrical activity of the heart and not maybe an optical proxy, still would use RMSSD, but they would compute it from bit-to-bit data of an electrocardiogram as opposed to the peak-to-peak data of an optical signal that shows typically changes in blood volume across the cardiac cycle. So still, RMSSD is the most useful metric when we look at HRV, and yeah, we just use it to mean HRV basically.

SPEAKER_00 10:28

Yeah. Surely they've got to come up with a better name for it. Yeah, because it's it's RMSSD is just not doesn't flow off the tongue very well. But um anyway, okay, so thank you for clarifying that. So how relevant do you feel HRV is for general health and longevity beyond just the sporting field? I know you've worked a lot across performance, but you um also discussed sort of longevity as well. And a lot of our listeners um look at trying to optimize their health span. So how could they use that as one of the markers um for optimizing their well-being?

SPEAKER_01 11:13

Yep, so the way I look at it, HRV is just a marker of our stress response. And as such, it doesn't really matter if we care about performance or if we exercise or anything like that, because we all experience stress, right? So we all have stressors that impact us. It could be our jobs, it could be our family, it could be sports, it could be our environment, it could be many different things, but we all experience stress, and HRV is just an objective marker of our stress response, which I think can be helpful to look at for people in different situations. And the wide application into sports, I think, is due to the fact that, of course, the stressor, which is exercise, is much easier to manipulate. So if we are in a situation in which maybe our stress response is not great, we can just you know scale down the intensity of a session. But if that is the situation because we have a lot of trouble at work, or maybe you know someone is not doing well with their health in the long term, like all sorts of different things that are much more complex than to act on or to change in the short term, then of course uh the data still reflects that situation, but it we cannot close the feedback loop right in an easy way like we do with exercise, which is probably just why that application works quite well.

SPEAKER_00 12:38

Yeah, okay. So putting it into context is probably really critical. Like, you know, I have an Apple Watch, I have a Whoop, I know you work with Aura, um, I've used your app in the past, um HRV for training. Um but I think uh getting a measure of HRV by its own needs to be interpreted and we'll touch on this along more a normative sort of data framework and but also putting it into context of what else has been happening. You know, your data doesn't know that you've had a couple glasses of alcohol or had a shitty sleep or had stress um previously. So is there a need to ensure that we're putting these values within the context of everything else that's going on within our lifestyle?

SPEAKER_01 13:34

Yeah, yeah, for sure. I think you raised two important points there. Um the normative data is really key and is something that is less obvious with respect to some other parameters. I think everybody is familiar with you know taking some measurements and checking if they fall within what is a range that is expected to be healthy. That could be for blood pressure, for blood glucose, many different parameters, right? That's a bit how we do things, also for many other markers. But then with HRV, when we take a measurement, we don't really have that frame of reference. We can only build our own by collecting data over time and then try to see when things change with respect to what is normal for us, not even for people similar to us, because there could be important differences also there. And then once we have that frame of reference, we need all the context we can gather, as you mentioned, because all various you know behavioral factors, but also environmental factors, all sorts of things will impact our data. So, even coming back to the simplistic example of using HRV in the context of training and exercise. If you, you know, sometimes people start using HRV, mean thinking that it will only reflect that stressor that is training that they care about, and forgetting that maybe you know their lifestyle is impacting the data a lot more than their training does. So we always need to keep in mind it's just an overall marker of stress. It's sensitive to any stressor, but it's not specific of anything. So you cannot, you know, looking at the data, derive which stressor is leading to a change. That's why you need you know to have the context, and that typically means that for your own data, you can kind of do that, right? You you're living your life, you know what is happening, even though it's not always obvious, you have the most context. But that's also why, if you look at data for another person, you know, the data will not tell you anything except for okay, we have maybe this stress response, but what is the root cause that you cannot know without all the context.

SPEAKER_00 15:39

Yeah. Okay. Cool. So we'll touch on a couple of points that you raised there in a little bit, Marco, the the time frames for um for capturing a HRV and how what sort of time frames you feel are ideal to create that normal band or normal range of of HR V markers. But in terms of context, should we all be tracking what other aspects and I know you're you tag in your um in your app, you tag a lot of you can tag a lot of sort of um other maybe not comorbidities, but other factors that might be adding layers of stress to to things. Um so is that critical to actually make sure that you're doing that if you're trying to understand what factors are helping you and what factors are hindering you.

SPEAKER_01 16:35

Yeah, yeah, for sure. I think um it's important as we want to look at this data and see what is leading to certain changes, maybe in a more repeatable way, tagging aspects of our behavior, simply, you know, when we have some alcohol, when we are sick, uh, how we feel we are sleeping. Like subjective data also very important, like how fatigued we feel. Maybe you know it's there's a change in environmental temperature, and something I was looking at today in terms of some of the data I collected, and then you see, you know, the season is changing, so to speak, and you know, it's getting warmer, and we maybe we feel a bit of fatigue, and that could also be reflected in the data, regardless of our training or other things that we are doing. So if we can collect some context around our data, then it will help us explore a bit of these relationships and see what is impacting it.

SPEAKER_00 17:27

Now you've got what 200,000 users on your platform or something similar? Um I'm sure you'd be have more up-to-date figures than I would. Um so what are the big factors that uh you've been able to identify that uh uh either poorly or positively impact people's HRV? You know, we know sleep is a is a massive a critically underlooked sort of tool for recovery, but um alcohol is a very, very negative has a neg negative impact on on HRV for most people as well. So uh are there any really interesting uh things that you found to address um improvements in HIV or to um really deteriorate HRV scores?

SPEAKER_01 18:28

Yeah, I would say what we see is uh in line with what we know from the scientific literature, we typically is that the obvious things help and that are sleep, as you mentioned, and then also typically a healthy diet uh and physical exercise. Um and of course, our sample is skewed towards people that exercise plenty, so that is sort of the baseline there. Um, but then we can also look at how exercise of different intensities impact HRV and how other stressors impact HRV. Something interesting we saw, for example, is that you know the impact of alcohol could be much worse than the impact of the hardest training we do. So, from that point of view, of just the magnitude of the suppression, you can see that you know there are behavioral aspects that definitely override even the hardest training sessions. So that's all context that we need to keep in mind when we uh look at the data. And similarly, sickness would be very aligned with those changes, which I think we all understand. The sickness is one of the strongest stressors that we can face, but maybe we don't often think of alcohol the same way. Um, and in terms of the positive ones, again, also at the large scale, I think sleep can be seen as a positive restorative stressor, so to speak, and that's also one of the reasons why I think in measurement protocols it could be interesting to measure HIV after sleep as opposed to during something again we can address more later, but you know, that way you see it as a positive stressor if you sleep well, or just something that was not as restorative as we were hoping, and that would impact you know your capacity to assimilate stress on a given day.

SPEAKER_00 20:27

Okay, interesting. So I think there's sort of known examples that are talked about quite a lot within the literature. Do you feel that um HRV can be used as an early warning signal for I know, maybe 24, 48 hours before people get b notice symptoms of being unwell that their HRV and their resting heart their HRV can plummet and their resting heart rate can elevate. Is that right? And can you see it change or predict a chronic illness or even metabolic dysfunction, for example?

SPEAKER_01 21:06

So I would say that in terms of uh, let's say more acute sickness or something like that, um it depends. Like let's say that there has been at least anecdotes of that. Uh people maybe not realizing that they were sick, not having other symptoms, and then seeing that in the data first. Um but obviously it's not that the data is predicting anything, right? It is just detecting a change that has already happened in your body, like you're already sick and you know your physiology has already changed. And then it could be that maybe you haven't realized yet, or there is still not a very strong symptom there. Um, so that is part of the picture. Like, as we get sick, our physiology changes in ways that are quite obvious because the change is quite large with respect to our normal, our heart rate is. Quite a bit higher, or HRV is quite a bit lower, and that can be captured. I think most cases people realize it when they also see it in the data, so it's kind of happening at the same time. But sometimes it could also happen that you see it a bit earlier in the data. And chronically, I think it's more complex because from one angle we said we don't really have frame of references as values. So if you were measuring your HIV for years and then you have developed some kind of chronic health condition, it is very likely that you've seen that change. So the data captures it.

SPEAKER_00 22:38

But if you start measuring potentially see a shift down from your normal range, is that what you're suggesting?

SPEAKER_01 22:48

Yeah, even though the normal range then is always computed over a shorter time frame to make it useful for you in the day-to-day. So eventually that will become your normal. So you really have to zoom out, yeah, to see the change. And then because even when you're in a chronic negative chronic health state, you will still have you know good days and bad days, so to speak. So I think it makes sense that your normal is again adjusted to your current condition, but then that needs to be managed, and that is probably different from where you would be without that metabolic state or other conditions. So for sure it is captured, but we have much less data there because it requires a much longer longitudinal assessment over years, and you know, we are just probably starting to see more of that as people have been using wearables now for a few years at least, while before there was basically nothing but maybe some studies doing spot checks, you know, every top month so that we know there is so much day-to-day variability that that is not particularly helpful.

SPEAKER_00 23:58

So you just touched on a point there, Marco, um that HR your HRV will vary significantly within the day. So my HRV this morning when I got up compared to being measured on my whoop in the middle of the night, compared to what it's like through the day, I've rushed into clinic, I've treated patients, I've dashed out, had lunch, rushed over here. That's going to change and fluctuate. So is there an ideal time to take that measurement? So I know wearables generally take it through the night, well, or and whoop do I think, is that correct? That's right. And you're I know you've often advocated for with you using your HRV for training tool to take it first thing in the morning, because then it's representing, okay, a positive outcome of okay, a restful sleep from the previous night. So is it just important that you're reliable in how you're measuring it, or is there a more ideal time to measure it over this longer time frame?

SPEAKER_01 25:16

Yeah, so I think uh that first we need to think about why we are measuring it and what we are trying to capture. So in this context, what we want to look at is really our stress response. Our which that means how we respond to the various stresses that we face and we have faced, let's say, in the past 24-48 hours, so the ones that have a bit of a stronger impact on our physiology, while we don't want to capture all the almost irrelevant stressors that we face during the day and that impact our physiology, but not in a way that if we're looking at it, would represent our overall state. And to simplify, I could say if I have coffee and then I measure, I will just see the effect of that coffee, that excitement, but that is really not representative, you know, my state as a stressed person or my inability to capture you know yesterday's workout and things like that. So we want to avoid those acute transitory stressors, and we want to capture the long-lasting ones, right? To do that, I think we have only two ways to do that, which is to measure in a reproducible context, so pretty much the same every day, and in a situation that is of rest and far from stressors, and that means the night or the morning. So the two cases that you have um brought up. And with the wearables, of course, you wear them and then they measure possibly through the night, so that is a reliable way to assess your resting physiology. While with an app, you could measure first thing in the morning after sleep, and that is also a similar context every day, so before you exercise, before you have breakfast, and things like that. Both work fine. There are differences that I think it's just important to understand because it's not the same thing. So obviously, if we measure in the night, that happens before the morning, and if we have some stressors later in the evening, again, could be alcohol, it could be that we exercise in the evening because we like it that way, or we don't have time earlier in the day, or many other different things, or we have just you know, maybe a good night socializing with friends, and that also leaves some excitement that you know you need some time for the body to renormalize. So as you measure earlier, because you measure the night instead of the morning, the effect of those things will be stronger. While if you measure in the morning, you might have renormalized because maybe those things were not so impactful after all, if you just let the body rest for a few hours. Yeah, exactly. And so, what does that mean in practice is that with the wearable, you might capture better your behavior, so what you did before, while with the morning measurement, you might capture better your capacity to assimilate additional stress for the day ahead. So it might be better for adjusting training because you know exactly what is your state at that point, while maybe the night and the wearable gives you some information about also your behavior and its impact on your physiology, and then maybe from that point of view could still be useful for your behavioral change and for things like that. So I think both are useful in slightly different ways if we understand also how your physiology changes in response to the stressors and over time.

SPEAKER_00 28:43

Yeah, okay. Okay, so people should possibly be aware of when they are measuring because that and just be consistent across their measuring time frame because you're going to get probably different results if you compare raw data from during the night compared to early morning uh after you've woken up measurements. Um why what what are some of the biggest mistakes people often make when interpreting HRV? Is it that higher is always better? That is probably the common Yeah, exactly.

SPEAKER_01 29:25

As you were saying, that's probably the most common misconception. Um you cannot blame people because that's what most other people that talk about HRV are saying is that you know it should be higher. So you hear that a lot. Um the way I think we should interpret it is a bit different, is more similar also to how we interpret other physiological signals, coming back to blood pressure or blood glucose, you don't want it the highest or the lowest. It's more about staying in that range that is optimal for you. And HRV is the same. And of course, this data changes a lot, as you were mentioning earlier during the day, because it responds to the demands of the various things that we do. But then when I measure tomorrow morning or tonight, I want things to renormalize because it means that I could respond positively to the stresses that I faced, and then my physiology bounced back to its normal. It's not really about going anywhere else, going higher. It's more about okay, let's bounce back to what is our normal. And then, of course, there are situations in which maybe our behavior could use some improvements in terms of I don't know, we have never exercised in our life and we start exercising. And that brings many different benefits to us. And I would say any marker that you measure is better three months later, when you once you go from inactive to having exercised. And among those things, right? It could be that your resting heart rate is a bit lower and your HRV is also a bit higher. But again, it's more about the behavior than trying to get HRV somewhere else.

SPEAKER_00 31:07

So what we're really looking for is the trends in if we're looking at HRV, we're looking at trends over a longer period of time. Um is that right?

SPEAKER_01 31:21

I would say what we want to avoid typically is maybe that we start spending too much time below what is our normal range. So we like stability, we like we like to bounce back. Sometimes it's fine to have an acute suppression. We don't sh we shouldn't be, I think, overly reactive. Things happen, you know, life is complex, but then we want to bounce back, you know, the day after, three days after. We don't want to start to spend too much time in a negative chronic state. Yeah, that's right. Because then we might get into those, you know, long-term negative stress responses like burnout or overtraining for an athlete and things like that.

SPEAKER_00 32:02

Yeah. Now I know I I use Whoop, and I think Aura the same, and they build a baseline normal data over about a 30-day period for uh your HRV. Now, I know you're an advocate for perhaps a longer time frame because you know, for example, a 60-day time frame, you know, it's pretty easy to go on holiday for a couple of weeks and suddenly you're in a very relaxed state and your HRV will probably climb substantially. And that becomes your new normal when you get back into work and perhaps training mode, and and so it it adapts in your your whole baseline is having to shift back again. So what's your ideal in in of that sort of normal range? How long should we be looking at it to create that normal range?

SPEAKER_01 32:58

Yeah, yeah, I think you explained it very well. Um I would say anything between one month and two months is good, is acceptable, uh, meaning that that is also what we see in research. Typically they keep it a bit short, and I think that is mostly because they're always in a hurry, right? People need otherwise to measure for a month longer just to get started with the study. So it doesn't mean that that's the way I think we should necessarily do it in uh for wearables, our consumer products where people don't have any issues in using them for several months or years. Um if we keep a shorter window, uh as in your example, then our reference range changes quite quickly. And as in your example, you know, then our frame of reference changes and our data might not represent where we are actually in a given time, because maybe even you know the opposite would be also maybe we get sick and we stay sick for a bit longer, two weeks, let's say, and our normal becomes that low state very quickly because then it's only four weeks and two have been of sickness. Like those situations, I think we want to avoid to be too reactive in adjusting the normal range so much. At the same time, we cannot have it too long. In the past, that's was actually what was done. You would start collecting some data and determine what is normal, and then you just hold it there. And that is basically for all the time that you collect data, you are using a reference normal range that you collected maybe even three months ago, six months ago, and so on, in which you try maybe to minimize stress or things like that. But that also doesn't work because physiology changes even just because of seasonality, right? So it's summer and it's winter, and our physiology is different, right? It doesn't make sense to be stuck into something so old. And so I use two months because I think it's a good trade-off, but you know, if you use six weeks to do the same or another tool uses something slightly different, I think it's all fine as long as we understand how things are changing. And in some cases it could be more reactive, in other cases it could be a bit less reactive. But as long as we have our own frame of reference, I think then it helps a lot in interpreting our data.

SPEAKER_00 35:24

Is is there a case? I know, Marco, you are an endurance coach and you work with a lot of runners. Is there a case for looking at it and and athletes generally have sort of standard sort of mesocycles over years over a year? They'll have, you know, standard training blocks that that will happen at different times of the year and if they're competing, it's generally sort of the same time frames. Is there a sense of you know, we we often compare training data points from year to year? Would you ever uh compare HRV and physiological markers uh over that previous year to their current year? Is that something that you do, or you would still stick with that sort of 60-day normal window of uh their normal range?

SPEAKER_01 36:15

Yeah, that's a great question. I think that in general we try to use the data always in the current context and time frame more than to look at where we were in the past, just because so many other things are happening that impact our physiology that we cannot always contextualize it or make sense of that outside of what is happening as your stress response. So maybe you were your range was a bit different, but then you're still your response was stable, which is typically what we see with athletes and people that are used to deal with certain stresses that are basically their day-to-day stresses like training. We see that their data is much more stable than for other people. And so that's really what we are looking at more than anything else, that things remain stable, regardless really of where they are uh in terms of the absolute values, even of the range.

SPEAKER_00 37:14

Yep. I mean one of the biggest things that I see with a lot of our patients when they share some some of their data and and uh data markers with us is that they are often very worried about uh daily drops in their scores. And most wearables give you a recovery score. I know Roop gives you a recovery score, I think, Aura is a is it a readiness score?

SPEAKER_01 37:39

Yeah.

SPEAKER_00 37:39

Um maybe it's a recovery score, and Garmin gives you a readiness score, and and these are uh you know an algorithm of various markers. But they often worry about drops in these scores, which are sometimes quite biased to HRV, um and and sleep. That's anyway, that's what I feel from looking at and interpreting my WHOP data. But people need to take that in put that into context again, don't they? And understand that actually, you know, a d a drop put into context because you've had some alcohol, you've had a poor night's sleep, or you've trained very hard, actually that's quite normal. And that's the body's r responding to a little bit of stress. And then if it extends beyond what, three days, four days, five days, maybe that's a sign that something else could be going on.

SPEAKER_01 38:35

That's right. I would always recommend trying not to be overly reactive and always balancing out what you see with how you feel. I think that is really key because you know we cannot say also you should just not you know ignore it or when it's you know particularly low or anything like that, because that is in uh, for example, in research, we went from being overly reactive and having this approach where the suppression leads to the change, even uh for training adjustments, to an approach that is more chronic, so to speak. So we wait for the baseline, the seven days moving average to go below the normal range, and then we implement a change. And of course, that happens when you have three, four bad days, as you mentioned. So it makes a lot of sense for a study. But in real life, if you have suppression but you also feel terrible because you're actually sick, then you cannot just wait three, four days, right, before you implement a change. So I think we always need to balance, right? Yeah, you need to balance how you feel and what the data is showing. But if you feel okay, I don't think there is any reason to stress too much about the data, just give it another day, another two days to see if everything goes back to normal and if you keep feeling good. And also always remember that a suppression does not mean that you are unable to perform on that given day. If anything, what we've learned from the literature is that when you are in a suppression or in a phase in which your HIV is also more chronically low, what is impacted is your ability to assimilate additional stress, not to perform. So if you are in need to perform, whatever that means for you, you know, in your profession or in your sport and things like that, a low score, assuming you're healthy, does not mean anything. It just means that you might not respond or assimilate positively that stressure and maybe further grow, right? Uh as in your performance or whatever else you're looking at. But it does not mean that you cannot perform. So that that is, I think, also a common misconception that people see, you know, a low value, and then they are afraid maybe they cannot do this or that, or the other way around, they see a low value and they are still able to perform, and then they think that the data is not useful, but that's not what the data is telling you. It's not really about your ability to perform, it's more about your ability to turn that stress that you're facing in the future into something positive for your performance or for your growth. And that might be impaired if your physiology is not ready to assimilate the stress. And that means, you know, it happens when your physiology is a bit suppressed.

unknown 41:21

Yeah.

SPEAKER_00 41:22

Because I think a lot of people m misconstrue that actually stress is is critical for the body's growth and development. You know, little bits of stress, be it exercise, be it uh uh a sauna or be you know, a little bit of cold plunge, actually can be incredibly powerful for for generating a response. But right, uh you know, most people when they look at their recovery score or readiness to score and they'll see that it's in the red zone or something like that, they will be alarmed that actually you know, or or maybe I can't perform today. So they're they're sort of re too reactive to that moment. So that's a super key point that people need to understand that actually it's perhaps more just uh telling you that actually you may not grow from this next uh stressor that you're about to apply on the body, but it's more about how you're going to um recover from the next stressor moving forwards. Is that have I summarized that? Yeah. So you mentioned before that there's a rolling baseline over a seven-day period. Is that sort of what we should be comparing to the to the um to your normal range over that 30 or 60 day window?

SPEAKER_01 42:47

Exactly. So we look at the normal range and then we look at our baseline as our current state and response to the stresses that we are facing. And then of course there is also the daily score that still acutely represents something that can be informative, as we said, for example, in case of sickness or else.

SPEAKER_00 43:09

Yeah, okay, okay, perfect. So there's so many metrics out there. There's so many different markers. We measure our sleep, we measure our stress levels, we measure HRV, we measure our heart rate, we uh uh measure RPE rating of perceived exertion if we're exercising, we we m we'll subjectively measure things. You know, in in my field we measure strength markers and and we measure range of motion and all sorts of things. So in terms of optimizing health, what are the key metrics that really matter, do you think? You know, you're a wealth of knowledge in this area. Is it a combination of those factors? Are these algorithms that the wearables have created? And look, I tested an aura ring for a period of time and a whoop ring, and I was like, what? They're giving me totally different recovery scores. How is that? And and so clearly they're interpreting the same raw data, but they're just interpreting it in a different way. Um So what are the key metrics that we should be really concerned about?

SPEAKER_01 44:30

I think that if we look at wearables data and the data we can collect with most sensors that people are wearing these days. A way to start reading a bit through the noise, so to speak, or through the many metrics that the devices provide us, is to distinguish between what is the device actually measuring and what is not measuring and is therefore estimating or sometimes making up based on some other signals that are somewhat related to what they want to measure. For example, recovery or readiness, they're not actual constructs that we can measure, right? So they are coming up with a model that tries to guess what is your recovery based on in part your physiology, but also in part what they estimate as your behavior, because also your behavior is not really measured. It's more like okay, maybe you've been active for this amount of time, or you know, you did this sort of exercise, and then maybe you slept for these hours, but also there, there's a margin of error, right? It is common for me that if I'm reading in bed, it will think that I'm sleeping. So there's error there, there is error in the sleep stages. There's many errors when we try to estimate things that we are not measuring, because again, to measure sleep, we will need to measure brain waves and other things that the wearable cannot measure. So if we really need to focus on a few, I will stick to what is actually measured, and that is typically just heart rate and heart rate variability, sometimes also temperature. So those can be useful signals to look at because it's actually what is happening in your body. And of course, it doesn't tell you everything you need to know, but then typically it tells you something that is more important than the aggregate scores, where there are still guesses at the end of the day and could be more wrong than other things. And indeed, if you use multiple wearables and you look at how your physiology changes over time, it is likely to be extremely similar across devices. But if you look at how your scores change over time, as you were mentioning, they could be all over the place because they try to estimate that entity that could be your sleep quality or your um readiness or recovery, but they do it in different ways and it doesn't really track well. So to me, it is just more useful if we stick to the actual data and look at that and learn to interpret that maybe ourselves sometimes.

SPEAKER_00 47:14

Cool. Okay, that makes a lot of sense. Um so diving into how these devices and your systems, you've been a big advocate for sort of camera-based measurements, um, which is what you can use off of the back of your phone. Um but uh uh these uh wearables generally use PPG and then also you've got ECGs, which you mentioned previously, as probably the most accurate uh direct measurement. So how how do they compare in terms of levels of accuracy and do we need to worry about that? Um should we just be reliable in what we're measuring and stick with that because there's there's an error in there, but actually it's reliable because we're continuing to measure the same to measure it in the same way.

SPEAKER_01 48:08

Yes, so um in general and under certain circumstances which are the ones in which we really care about measuring this, which is during sleep or first thing in the morning at rest. The data is rather similar between looking at NECG, so the actual electrical activity of the heart, and looking at blood volume changes at the peripheries, at the with a ring or with a wristband or with a watch. And that is because, of course, the blood is flowing when the heart is beating, and so if we measure at the heart directly, or you measure blood flow at the periphery, they are very similar. And what we really care about again is not really the absolute value, where we certainly can have differences, but it's more how those values change over time. So if my values measured with ECG and PPG are slightly different between themselves, but over the days they change in the exact same way and same pattern, then I can capture my stress response in the same way, and that's what that is what we typically see, even though, even in terms of absolute values, they can be very similar when measured at rest. There's always some differences, or maybe people that might have you know poor perfusion, right? Just blood flow at the finger maybe is not as easily measurable. And there are cases in which the optical technology might not work as well as the chest strap, for example, which is like a polar strap linked to an app would be another way to measure in the morning using ECG. So different ways to do this, but in general, I think there is enough um evidence that these methods can all be used at rest conditions, right? So the during exercise or during movement, the optical methods can become very quickly inaccurate, and therefore they are not really the best way to look at this. But wearables measure HRB typically when you sleep, and when you sleep um there is no movement, and the data can be accurate even when measured in different positions on the body, again, breast, finger, um, things like that.

SPEAKER_00 50:28

Okay, cool. So so the wearables are pretty good in terms of their level of accuracy for what we're trying to achieve here. That's sort of what you're saying, isn't it? Yeah, for sure. Um so we've we've touched on this on sort of lifestyle and behavioural sort of factors around sort of HRV. But stress and alcohol, are they probably the two biggest negative contributors to to impacting your HRV?

SPEAKER_01 51:01

Yeah, well, probably. Um, together with sickness, of course, uh which illness, yes. Yeah, yeah, hopefully we experience that less. Uh but then um yeah, I would say those are some of the factors that have uh the biggest impact uh or that we could define as negative impact, then of course some of the positives can turn into negatives if you know we don't um yeah, prioritize them or or have the right dose of them. I'm thinking about sleep, right? If you have very disruptive sleep or very short sleep, then sleep becomes a problem, but then by acting on that we could try to see if we can also lead to changes.

SPEAKER_00 51:45

Yeah, okay. And so I know you've done some work with aura around a s the sleep algorithm. Is that is that correct? That's right, yeah. But how how accurate are the wearables at measuring sleep then? You know, I uh I understand that uh okay heart rate and HRV can be relatively accurate um when taken during the night or or or in the morning after a sort of restful state. Do they generally get sleep relatively right or is there a big variability between a a proper sleep study?

SPEAKER_01 52:30

Yeah, yeah, I think that maybe with the sleep time and as such also sleep efficiency, which is just how much time you are actually sleeping against how much time we are awake during the night, we get quite good, right? So that's a two-stages classification, basically, just asleep and awake. As we start looking at deep sleep or REM sleep and four stages, the results have been getting better and better over the years, but it is a difficult application. I think that there is a main issue there, which is that even the reference is not perfect, right? So because you train your model and algorithm and whatever you're building against a reference, which should be the truth, like what is this stage at this time? But that is derived from experts that annotate the data and that agree maybe 80-85% of the time. So the best you can get is that level of agreement in which we already know there isn't a perfect agreement when we look at those things. So it is a bit of um difficult application to reach particularly high levels of accuracy. Um that's also why we see some inconsistencies between wearables. To me, that's the easiest way to understand if a metric is reliable or not is to look at different devices and how they behave for you. Because if we were able to get that information reliably, there is no doubt that the smart people at VUP, the smart people at Aura, the smart people at Garmin, at Apple, they would have figured out the way to do it. Just like for the other parameters that you can see track very well. You know, if you look at heart rate, it's gonna be the same across the board. But then if you look at RAM sleep, it's gonna be a bit all over the place. So that tells us that maybe the level of complexity or the input data that we look at is just not enough to get extremely reliable for these signals. Yeah.

SPEAKER_00 54:35

Do you you know you work in this field, you're a data scientist, you probably love looking at the the data. Do you think that as a as humans and as a society we're getting too obsessed with data? And we're losing that ability to just go, hey, you know what? I feel good today. I feel okay. I don't need to look at my Apple Watch, my O-ring. Is there a case of still just checking in with yourself and going, great, had a good sleep, ready to go?

SPEAKER_01 55:13

Yeah, yeah, for sure. I think actually that is the so the that aspect of self-awareness for me would be the only reason why you would use the data. So it would be to get more self-aware, but we see clearly that that is not how the data is used by a lot of people, which either become a bit too obsessed with the data or that decide how they feel based on the data. So I think there's a lot there that went wrong, so to speak, and that maybe we should rethink a bit in terms of how you know the data is presented or used by the people that use devices like this. Um, but yeah, in my opinion, you know, the way you feel, and also in terms of, you know, we said I do some coaching for runners, and actually the way I prescribe training is just by perceived effort. Like I don't say, you know, you should get this space or you should get this heart rate or this power. It's always by perceived effort. It's like the core of the whole system, is to learn how you know you feel at different intensities. And sometimes that is not an innate ability. So it's not that the data is useless and it's just okay, let's forget about the data because we can just use our feel, our perceived effort, and all of that. Like sometimes we need to develop that, and the data can be really helpful. So maybe we need to look at our heart rate when we go running and understand that we need to go a bit lower intensity, a bit easier, and then later we start to be able to do that without looking at heart rate. And with resting physiology, it's the same. Maybe sometimes we have some difficulties understanding, you know, how we are responding to the various things that we do, what is our fatigue, and so on. And by looking at the data and using the data, we can improve that self-awareness piece. But we need to work, I think, proactively towards that, as opposed to, you know, just looking at the data and relying on the data.

SPEAKER_00 57:12

And so you're you're talking about just measuring RPE, rating of perceived exertion. Yeah, yeah. Because the standard scale is is sort of zero to twenty, but do you use a zero to ten scale or do you use the the the standard scale?

SPEAKER_01 57:29

Yeah, for training I use zero to ten uh just because I think it's a bit more intuitive for people. Yeah, yeah.

SPEAKER_00 57:37

Yeah, I mean that's interesting. You've got an amazing lab over in Italy. You train a bunch of elite athletes, you have all this data at your fingertips. But one of the best, quickest ways that you optimize their training zone is to get them to, okay, I want you to train, so it's about a seven or eight out of ten. That's right. And but but as you mentioned, that a lot of people who haven't had an athletic background, they actually struggle. And certainly with some of our patients who don't come from an athletic background will will sometimes match exercises against an RPE score, but they struggle to be able to calculate that until they've you know got into exercise for a period of time, which I suppose is where the data tracking to match that can become really, really useful.

SPEAKER_01 58:29

Yeah, yeah, I agree. I think that process is key. Uh, even for my own training over the years. I mean, I got better when I started looking at data and understanding, for example, that my heart rate was always way too high when I was running and I had to slow down, and it was all of a learning process, and you know, now I can go out and I don't need to track my heart rate. But again, I wasn't just able to do this. So the data is part of the process, but it shouldn't be it shouldn't be, you know, the only outcome we we look at, you know.

SPEAKER_00 58:58

Yeah, okay, amazing. So if a couple of little practical takeaways. So I think you we've sort of probably touched on lots of aspects of it. But one of the simplest ways for people to start using or measuring HRV might be to use, if they don't have a wearable device, would be to use your app. Um which and they can simply take that and it records it for them first thing in the morning. That's what you're advocating.

SPEAKER_01 59:29

That's right. So with the app sitting first thing in the morning, the only thing to do before the measurement is if you need to go go to the bathroom and then come back, sit up and take your measurement. I think that's the easiest way, just for a minute, every day in the same context more or less, and then you will start building that normal range that we discussed so that your physiology can be interpreted with respect to what is normal for you.

SPEAKER_00 59:56

Okay, perfect. And if someone wants to improve their health span, we need to look at multiple metrics, really. HRV is not the the only metric that we need to measure looking at that. Um are there any particular what are the other metrics that we want to do? We know that okay, VO2 max is generally correlated. A higher VO2 max suggests that you might uh have a better, longer health span, as in you may be more mobile and agile and stay well for a longer period of time um as you're older. It may not increase your your longevity. So that's an important metric. Your resting heart rate is probably one other key metric. HRV is another one uh very useful one to track. What else am I missing here?

SPEAKER_01 1:00:56

Yeah, I would think what you're mentioning are all in some way integrative measures of certain capacities, right? From our cardiorespiratory fitness level to maybe our stress response, uh which we could capture with HIV or SIN heart rate. Maybe we can look at also our basically our metabolism in different ways, right? So we could do that today also in a way with wearables, right, with CGMs and continuous glucose monitors and other devices that maybe come up in the future able to do that even in a more non-invasive way. I think that is for sure a relevant aspect that is linking a bit of everything, right? Our stress response, our activity, and also how frequently we are active, as well as our diet. So I think that is probably one signal that um we should start to pay a bit more attention to as well, given also the prevalence, right, in of metabolic diseases and other conditions.

SPEAKER_00 1:02:04

Yeah, definitely. And it's interesting that there's a big um CGM maker who I've just forgotten their name, but uh Whoop have just raised, I think, five hundred and fifty million um dollars just recently and uh they've had a big investment uh come in from one of those big CGM makers. So whether whether whether that's uh you know a line of travel that Whoop is trying to head to to be able to develop a non-invasive aspect that will monitor uh you know blood sugar levels, I think that might be yeah, might be interesting as another market to be able to track uh fairly easily from a health span perspective. And clearly, you know, measuring measuring muscle uh your your muscle bulk and strength markers and all of these add weight to optimising your health span, but they're uh a little more complicated to measure rather than uh through a simple device. Um so Marco, it's been I've loved chatting to you and loved being able to um geek out a little bit around HRV, the physiology of it, and how it represents sort of stress in the system. Uh and you're clearly someone who is at the top of the game and very, very no knowledgeable in this field. So really thank you for your time on this and um would love to pick your brain again in in in the future at some stage.

SPEAKER_01 1:03:44

Thank you. I really appreciate it. And uh yeah, my pleasure to be here and and shot. I hope this was useful. Excellent. Thanks very much, Marco.