Needless to say, athletes are extremely different in terms of their 1) power output (simply put, how fast they are running in absolute terms), 2) metabolic flexibility (i.e., the ability to use different substrates at different intensities) 3) gut tolerance during exercise (i.e., potential issues with nausea, cramping, or worse), 4) relative intensity they race at (which tends to depend on the event duration and athlete level). These factors make any pre-made / generic / standardized fueling plan far from optimal. Guidelines give us a starting point when we don’t know anything about the athlete in question or their event, but as we work with an individual, we must do better and strive to understand what are the individual characteristics that determine the optimal plan for them.

So how do we get there?

As you might know, I’ve developed a great interest in metabolic flexibility and how the interplay of (genetics), training, and diet impacts our ability to rely on different substrates during exercise and, as such, our ability to perform optimally in a long-duration event. I’ve discussed extensively periodized nutrition and my initial experiments (see here), as well as my current approach for my recent races (here), resulting from possibly hundreds of tests since I bought a metabolic cart.

Briefly, my high-carb diet resulted in extremely low fat oxidation rates (0.2 g/min), which combined with my inability to intake high amounts of carbohydrates (low gut tolerance) led to poor(er) than expected performance in events longer than the marathon, until I periodized my diet (some may call it carb-cycling), greately increased my metabolic flexibility, and improved my performance in the 50 km and beyond while typically relying on ~30 grams of carbohydrate per hour. I am of course aware that my rate is on the low side, but protecting my gut is my number one priority, and I had to find ways to make this work. Others might do well at rates that are even 4 times higher, which again speaks to the great individual variability in all things human physiology and performance. I am not here to tell you to eat less, or I’d be the same as the ones telling everyone to eat more (carbs). I’m here to try to elaborate on my thought process when it comes to truly individualizing fueling for an athlete, hoping there are some useful pointers for you in there.

In the past year, I have also had the opportunity to test many of my athletes at my lab, and here below I will summarize my current process and show some of the tools I built for them, so that we can better prepare for the demands of their events, based on their unique physiology, as opposed to generic guidelines that can fail many individuals (just like they have failed me and held me back in the past).

Clearly, I do not have all the answers (and I am always open to learn more and update my thinking), but here is how I work at the moment.

Measurement Protocols for Metabolic Flexibility Assessment

If I had to pick one single test that provides me the most useful data to assess metabolic flexibility and substrate utilization for a long endurance event, it would be a 40-60-minute long incremental test with small increments in intensity (e.g. 0.5 km/h per step) and long steps (e.g. 4-5 minutes each) where we go from a super easy jog to a moderate intensity that is typically below the second threshold (VT2).

This range of intensities covers the spectrum we care about for marathons and ultramarathons, while the long steps and small increments in intensity allow us to reach steady states, which are necessary to derive oxidation rates that are reliable (something that would not be the case with the typical short steps and large increments of VO2max tests).

This is indeed the test that I’ve started doing with all my athletes when they come to the lab:

Additionally, I’ve started doing a second test, exactly the same, but typically shorter, more focused on what we expect race intensities to be, in a fatigued state.

Metabolism is not static, and we know that fat oxidation rates increase over time with exercise, even when we fuel at high rates (e.g., elite athletes can easily get to 1 gram/minute of fat oxidation while intaking 80+ grams of carbohydrates per hour). For this reason, I think it is key to assess how we are doing after an hour or two (or three, all depending on how long the race is). I also think that at times, especially for athletes who have a particularly high-carb diet, fresh-state tests tend to penalize them, or to show fat oxidation rates that are quite low, because they are blunted by e.g. their breakfast. Still, I want to test athletes in realistic race/training conditions, and it would not be meaningful, in my opinion, to test them fasted if that’s not part of their routine.

The problem is easily solved by the ‘double protocol’ in which we do a test in a fresh state, they then go running outside for a number of hours, come back to the lab, and we test again to see if their metabolism has changed and what the impact of that would be in terms of their race nutrition. In my experience, this protocol is not particularly useful if fat oxidation rates at baseline are already high; for example, my own rates don’t change much because I start at 1 gram and that’s already a very high rate that is basically capped. However, if the athlete is moderately low in fat oxidation, e.g., 0.4-0.6 grams/minute, then I think the test can help us understand if they are stuck there or if the body switches to something that spares more glycogen as the training or race goes on. Hence the protocol:

Later we will see some data resulting from these protocols. My main points here are that 1) long tests able to reach steady states at race intensities are necessary, while typical incremental (“VO2max”) tests are of less use, 2) metabolism is dynamic and for long events, measuring fresh and fatigued can be insightful for many athletes.

Assessing Race Demands

Our protocol above allows us to assess metabolic flexibility (fat and carbohydrate oxidation rates at different intensities); however, we also need to know how long we are going to be racing and at what intensity, in order to build our individualized fueling plan.

For this purpose, I built various tools, relying on my athletes' workout data. I’ll use the example of a marathoner here because the math is simpler (with respect to ultratrail, where we have more variables).

First, I log details about their individual workouts every time they do a key session (e.g., intervals, long runs, or pretty much anything that is not an easy run), then I aggregate this data to get a nice overview of the current state of affairs, for example using the density plots below for grade-adjusted pace and heart rate:

From this data, I can estimate Critical Speed, and from that, we can estimate marathon time, for example:

Given that these models are also generic and make lots of assumptions, I built some extra tools that give me more insights into the specific strengths and limiters of an athlete, so that we can adjust the estimate above (if we have poor durability, the estimate is probably optimistic):

This whole process gives us a race expected time and intensity, which we can derive for any event based on historical data of the athlete, and will be an important parameter to build our fueling plan, which is the next step.

Building a Fueling Plan

Why are we fueling?

As discussed earlier, race fueling should depend on a few factors:

1. Your power output: the higher your capacity to produce a certain output, the more carbohydrate you will burn, regardless of your fat oxidation, which can only account for so much. This is why elite athletes’ intake should be higher than, e.g., my own.

2. Your metabolic flexibility in relation to the duration of the event. The more fat you can burn, the less glycogen you will burn, and therefore the less likely you are to “hit the wall”, since glycogen is a limited resource. Contrary to common belief, hitting the wall is not primarily a matter of eating too few carbohydrates during the race but could happen for the opposite reason, e.g., fat oxidation rates are too low, forcing the body to rely heavily on carbohydrates and deplete glycogen quickly. In reality, multiple factors play a role, with the biggest one by far being poor pacing and overly optimistic starts, which not only deplete your glycogen faster, but also speed up neuromuscular fatigue, muscular damage, and the buildup of byproducts that also lead to increased fatigue (e.g., hydrogen ions). Start slower. Finally, keep in mind that apparently no amount of carbohydrate intake can spare muscle glycogen - and only liver glycogen is impacted by carbohydrate intake during exercise (see Matt Carpenter’s excellent video on the topic), which makes a strong argument for the importance of metabolic flexibility even if you have ‘an iron gut’.

3. Maintaining blood glucose levels: there is plenty of new research on this, but in short, this is another important reason why we need a certain (low, about 10-15 grams/hour) level of carbohydrate intake even when we are very metabolically flexible and when the race is short enough for glycogen stores to last.

4. Central effects: the brain likes carbs, to the point that even mouth rinsing improves performance; hence, frequent sips might be a good idea even if the overall rate is not particularly high, unless it leads to nausea.

Based on points 3) and 4) above, we know that we need a somewhat constant intake of carbohydrates, so that we can maintain blood glucose levels and keep the brain happy. This is more important that it might seem at first, because it becomes a reason to ‘protect the gut’ and use a rate that is appropriate for us also in terms of our tolerance: having loads of carbs for 2-3 hours and then being unable to eat for another 2 hours, is not the same as having a constant flow of (fewer) carbs throughout the event, even if the total amount we intake is the same.

Let’s get back to our tests and protocols. Given that the metabolic test I carry out with my athletes covers race intensities, it will provide us with the oxidation rates that the athlete is expected to have on race day, and therefore will take care also of points 1) and 2) above, or the absolute intensity of the effort (again, the reason why elites need more carbs than non-elites, despite their higher fat oxidation rates, is that when you ran at a fast speed in absolute terms, you burn lots of carbs even if it is a relatively low/moderate speed for you because you are an incredible athlete) as well as the relative intensity of the effort (how much fat and carbs we burn relatively to our full profile).

Knowing the athlete and history of intake, we can start with a plan that reflects their past behavior (in terms of intensity for a similar race as well as carbohydrate intake per hour):

We then load their metabolic profile data and look at fat and carbohydrate oxidation rates at race intensity or near race intensity:

We can see that for a sub-3 hour marathon at this athlete’s metabolic profile, we have 2 hours and 42’ of fuel-free range (which of course is not true, just an oversimplification, but it gives us an idea of how much time we need, given fat and oxidation rates, to get close to depleting glycogen), and when adding our 50 grams/hour fueling plan, we have plenty spared:

For a much longer race, 0.4 grams per minute tends to be on the low side and might become a limiter, but for this athlete and marathon goal, we can certainly get away with it without extra interventions.

Below is another example where we look at ultratrail, and specifically a race of 70km with 3000m of vertical gain. In this case, the athlete is expected to take about 9 hours, and here we have both tests mentioned earlier, hence fresh state and fatigued state.

In this case, we can appreciate how metabolism is shifting between tests with fat oxidation rates increasing over time (20% increase in fat oxidation after 3 hours of running):

Given this athlete’s rates, which are higher than the previous example, and also given a lower relative intensity for the race, we have a longer fuel-free range (almost 5 hours), but the race is also much longer, to the point that if we were to target a relatively low intake (e.g. 30 grams per hour, which would be fine for a marathon) we would not meet the demands of the ultra trail race:

While with 50 grams/hour we get there:

Finally, I’ve said earlier how my optimal intake rate is pretty low given high fat oxidation rates, poor gut tolerance, and relatively low race intensity for an ultramarathon.

Looking at my lab data plus the race intensity of the events I can sustain for e.g. 8-9 hours (low Z2), you can see my fuel-free range is 7 hours, and 30 grams per hour of carbohydrate is plenty for me:

This whole process and the tools shown above are always improving, but I feel like at the moment I have a decent framework to start a conversation and work towards an optimal plan for an athlete's goals, as opposed to simply relying on generic guidelines that have little to do with an individual’s metabolism.

A Note on Higher Rates

The exercise above is useful to determine a rate that, with a safe margin, should lead to optimal performance, given that we are not running into premature fatigue because of low glycogen. My preferred way for my individual case is to be metabolically flexible, for the reasons mentioned several times. For others, especially elites burning lots more per unit of time, a higher carbohydrate intake rate might be a solution that leads to the same outcome (i.e., optimal performance). Again, this is about individualization, not about “a universal solution,” as there is no such thing, in my opinion.

Given that we are in the era of “maximizing carb intake”, I will add a few considerations on that as well, especially for non-elites that are unlikely to need such rates. Going very high in terms of intake with respect to our needs increases the risk of self-sabotaging dramatically. Even if we are able to tolerate a certain amount of carbs per hour in training, the race can be a different beast: different temperatures, possibly running at night when the body might be less inclined to absorb food, very different stress levels with all that implies for the gut, a different diet and different training regime in the days preceding the race, etc. - something to be aware of.

On the flip side, despite no established link between particularly high intakes and performance, you can certainly give it a try if you have never experienced gastrointestinal trouble and would like to see if you can perceive a positive effect. It is certainly true that often it takes a while for research to catch up, and also that experimentation and trial and error remain an important part of finding what is optimal for an individual in many things training- and racing-related, in my opinion.

As it often happens, our history can teach us a lot, and if in previous races you tried high rates and consistently had no issues, great, we can keep that going. However, given that we have a clear link between higher rates and higher gastrointestinal issues, we need to be aware of the tradeoffs and risks involved (it would be a pity to DNF because of something our body didn’t really need). In this context, greater metabolic flexibility is simply another tool in the toolbox, which allows us to take fewer risks as well as adjust rates on the go should something unexpected happen in a race (as we can function well at various rates of carbohydrate intake, as opposed to being unable to function without taking in a particularly high rate). Many ways to go about it, as you can see.

To conclude, keep in mind that Yomif Kejelcha ran a sub-2 marathon with a fueling plan of 70 grams/hour, missed 2 bottles, and eventually had about ~55 grams/hour of carbs. Sub-2. Somewhat similarly, Tom Evans, who burns 1.4 grams of fat at his first threshold, reduced his intake after years of gastrointestinal trouble, and eventually won UTMB. They individualized their intake to match their physiology and the demands of the events they cared about, as opposed to “maximizing carbohydrate intake”, which seems to be the common advice these days.

Wrap Up

Alright, I hope this was informative and gave you a decent overview of how I’m thinking about race nutrition and trying to individualize it to the specific demands of the event and physiology of the athletes I’m working with.

The interplay of power output and metabolic flexibility is particularly interesting to me, as it shows how recreational runners can potentially do great with limited or moderate carbohydrate intake, therefore easily protecting their gut (#1 cause of DNFs), if metabolically flexible. At the same time, given that fat oxidation rates tend to be capped at 1-1.4 g/min even for elite athletes, their carbohydrate requirements remain substantial (and will be further impacted by race duration). It follows that if you are not an elite and are not generating a huge power output, fueling like one could be a quick way to sabotage your own race (unnecessarily!).

Please feel free to reach out should you be interested in working with me.

Happy running and thank you for reading!

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

The Book

Our book is out and available for purchase from Human Kinetics’ website, here (or on Amazon).

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified Ultrarunning Coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, Endurance Coach at Destination Unknown, advisor at Oura and augo, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Personal Substack

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

I’ve learned a lot in the process, especially when it comes to aspects of HRV I was less familiar with (e.g., in the context of cognition). Something else I’ve enjoyed learning is the Vagal Tank Theory that Sylvain and Emma have developed (full text of their paper here).

In this blog, I want to cover the basics of their theoretical framework, and link it to practical tools we can use to assess or improve self-regulation, while keeping in mind that there are strong limitations in terms of when and how we measure and interpret HRV (i.e. the association between HRV and the autonomic nervous system response to stress is valid only under certain circumstances, we cannot simply measure HRV all the time and claim it’s a stress signal, as I’ve explained here).

What’s the Vagal Tank Theory then?

The Vagal Tank Theory

The model developed by Sylvain Laborde and Emma Mosley builds on top of the neurovisceral integration model. In particular, the neurovisceral integration model provides a framework to understand resting HRV, i.e. how the brain and the heart communicate, for example, how the regions of the brain governing cognition, emotion, autonomic control (e.g. breathing and heart rate, or things that happen without conscious control, typically) are all deeply interconnected, which eventually reflects on HRV, making it a useful proxy of many self-regulatory mechanisms and our adaptability to stress. The Vagal Tank Theory extends the neurovisceral integration model beyond the resting state, and in particular through the metaphor of a fuel tank, introduces the concepts of Resting, Reactivity and Recovery, which match what happens physiologically as we face a stressor.

I like this framework as a way to think about self-regulation and stress, even more than HRV itself (which is just a tool, albeit a very useful one, in my view).

Especially in the context of sports performance, where the stressor tends to be somewhat more controllable (at least in terms of the stimulus we provide to the body, i.e. the training plan we design and adjust), looking at the body’s self-regulatory mechanism through the lens of HRV (and this framework) can be quite effective (as research on HRV-guided training has shown us - but I think there’s more to it, as I will try to cover below).

Resting, Reactivity and Recovery

The core of the model are the three Rs, as the authors call them, i.e., Resting, Reactivity and Recovery. Let’s get to them.

Resting: The Morning Measurement

This is likely what you are familiar with the most, i.e., our morning (or night) HRV: our baseline capacity to assimilate stress. Nothing new here as we are looking at a resting state measurement (best practices here), which reflects our ability to assimilate stress on a given day. What matters is that we capture our state 1) at rest and 2) far from stressors, so that we can assess our nervous system response without confounding factors or artifacts, which would happen if we were measuring under different conditions (see here).

We typically aim for a stable resting HRV, which suggests we are responding positively to training and other life stressors. An acute/daily suppression - when feeling well - is something I would not necessarily be worried about, as (I think) a better (and healthier) way to use data is to avoid being overly reactive to the various things we measure. If the suppression remains for a few days though, it is a stronger signal that something is preventing our physiology from re-normalizing, and as such, it is probably a good idea to implement some changes, either reducing training stress, or focusing on other recovery techniques, to aid self-regulation.

Above we can see resting HRV (morning measurements, seated, using HRV4Training and a Polar H10, because there is nothing better than ECG to measure HRV) for the past 60 days, characterized by sickness first, and then a phase of stable to high HRV (with respect to my normal range), which provides very useful feedback as I increase training load near my current limits (about ~100 miles/week): resting physiology remains stable, showing positive capacity to assimilate the load, as opposed to e.g. being overwhelmed and potentially moving towards non-functional overreaching.

Resting HRV is what you typically read about in this blog. Now on to Reactivity.

Reactivity: Acute Response to a Stressor

We can look at reactivity in various ways. For example, in our book we show examples of withdrawal (as in the figure above), i.e., HRV reduces as we face a stressor but we also show replenishment, i.e., HRV increases as we improve recovery with a “positive stressor” (e.g., sleep).

The first response, or the withdrawal / reduction in HRV, is the typical response when we think about a stressor, and needless to say, it is perfectly normal for our physiology to respond this way as we mobilize resources to meet the demands of whatever we are doing (training or else). In fact, a more stable HRV in this case might show a maladaptive response, as a healthy system is stable in the long term, but is not unable to respond and adjust when facing a challenge, a characteristic that is key to both health and performance. Hence, reactivity doesn’t have a optimal signature, but always depends on context. The larger the stressor, the more adaptive a large suppression can be, most likely.

A simple and effective way to look at reactivity without all the issues associated to measuring and interpreting HRV outside of the morning routine (or night data), is to think about the orthostatic stressor, or in other words, changing body position. As we change body position from lying down to sitting (or standing) up, the body needs to quickly adjust so that for example, we don’t faint because blood is pooling in the lower body. I’ve discussed here how this simple orthostatic protocol allows us to capture the stress response in ways that are not visible in night data (night data is a suboptimal protocol for athletes, as shown again in recent research, here). With this protocol, we are exploiting the reactivity of the nervous system and its impact on HRV, while assessing our Resting physiology.

Reactivity can also be seen in relation to “positive stressors”, for example taking an extra nap on a day in which our HRV is a bit lower than our normal, would likely show an increase in HRV in the acute phase.

A slow-breathing exercise could be another option to improve self-regulation and recovery, as I’ve discussed here in the context of using HRV4Biofeedback after high-intensity training (this is just speculative/experimental).

Recovery: Re-Normalizing (or Not)

The final R stands for Recovery, which closes the loop with respect to our initial resting measurement.

Here I want to stress again how we cannot really measure recovery after an acute stressor and can only rely on assessments of resting physiology at specific times, using good protocols (either first thing in the morning, my preference, or during the night). I’m stressing this again because while the Vagal Tank Theory is great at covering how self-regulation works in the context of our stress response, we should keep in mind that this does not mean that we can measure HRV continuously to assess our recovery. There are simply too many things going on for an HRV measurement taken outside of the morning routine or the night, to be of any use in assessing our stress response - it is mostly noise (again, here is an overview of some of those things).

Recently when talking to Dan and Owain for The Physiology of Endurance Running podcast I also covered why, even under realtively controlled conditions, it is basically impossible to get a clean signal to capture post-exercise recovery, something that I tried many times as in the lab it showed to be pretty useful (i.e. to clearly separate athletes by fitness level, or to clearly separate the intensity of the training stimulus, e.g. between low intensiry or below VT1 and higher intensity, as well as the effect on autonomic activity of training duration), but again, too many confounding factors and artifacts to make any of this practically feasible even if you put your best effort (no moving, no muscle contractions of any kind, no swallowing, no talking, no drinking, no eating, self-paced breathing - which is however dynamic after exericise - and all of that, for e.g. half an hour to get anything useful - let alone if you were to collect data passively).

Alright then, what do we make of recovery? In practical terms, I think Recovery is the same as Resting, and is captured every 24 hours either in the morning or in the night, so that we can see if we have re-normalized or not. We all face stressors during the day, some we might impose (e.g. training), some are just coming at us, and the cumulative impact of all of those plus our capacity to assimilate stress in that given day/period, and our ability to re-normalize, will determine tomorrow’s data point, our Recovery is how far that data point is from the previous Resting measurement (to be more precise, how far is it from our normal range). This is very useful when we manipulate training in different ways, for example increasing the density of high-intensity sessions tends to have an impact on the data that lasts many hours, and if the acute suppressions on the following morning measurement keep coming even after repeated exposure (i.e. when we apply the same density over and over, not just the first time), maybe that density is too much for us at this time.

What we do in between morning (or night) measurements to aid recovery can’t be quantified, and that’s fine. Eat well, sleep enough, move around, and possibly, breathe slowly sometimes.

Finally, in the context of both Reactivity and Recovery, keep in mind that acute and chronic responses often differ dramatically (e.g. exercise acutely suppresses HRV, but chronically increases it), and sometimes similar changes in HRV reflect different underlying physiology and stress responses. For example, cold exposure activates pretty much everything in the sympathetic and parasympathetic nervous system → vasoconstriction → higher blood pressure → baroreflex → lower heart rate and higher HRV, which is not really the same as a slow breathing exercise → higher sinus rhythm arrhythmia → higher HRV, a more ‘selective’ approach, so to speak. There is more to the stress response that we cannot “see” through HRV, even if we were able to measure it meaningfully more frequently.

For these reasons, I think our best practical use of HRV is for morning assessments of our Resting physiology, which is linked to our Recovery from previous stressors and provides an indication of our capacity to assimilate additional stress in the current period.

Wrap Up

The Vagal Tank Theory that Sylvain Laborde and Emma Mosley have developed gives us a great framework to understand self-regulation and think of the stress response and its impact on HRV.

We want to keep the vagal tank in our optimal range, i.e., Resting physiology stable over time, while providing timely stressors that will lead to plenty of acute changes (Reactivity) but that shouldn’t go (too often) beyond our current capacity to assimilate them, so that the Recovery and re-normalization is apparent in the following day’s Resting measurement.

While the stress response and self-regulation are rather complex and dynamic, when it comes to useful measurements using HRV as a proxy of those systems, less is more and I’d encourage mostly relying on morning (and/or night) measurements, leaving aside tools built for engagement, more than anything else.

Thank you for reading!

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

The Book

Our book is out and available for purchase from Human Kinetics’ website, here (or on Amazon).

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified Ultrarunning Coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, Endurance Coach at Destination Unknown, advisor at Oura and augo, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Personal Substack

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

Your HRV dropped after a hard session. Most runners take that as proof they trained hard. Marco Altini says it’s proof of the opposite.

Marco is the founder of HRV4Training, an advisor at Oura, and an endurance coach. He came to physiology the long way round — from computer science and early wearable prototypes, back to university to study sports science, because he wanted to understand the bodies he was measuring.

In this conversation we cover what HRV actually measures, why “higher is better” is wrong, how to build your normal range, morning vs night measurement, what to change when your baseline drops, why a falling HRV in taper week is the good news, and why the data should make you need the data less.

Find Marco on Substack, where he writes about HRV, training and coaching.

Audio, here:

Video here:

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

The Book

Our book is out and available for purchase from Human Kinetics’ website, here (or on Amazon).

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified Ultrarunning Coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, Endurance Coach at Destination Unknown, advisor at Oura and augo, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Personal Substack

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

Dr. Marco Altini joins Klaudia and Don for the deepest heart rate variability conversation the show has done yet.

Marco holds a PhD cum laude in applied machine learning, an M.Sc. in human movement sciences and high-performance coaching, has published more than 50 papers and patents at the intersection of physiology and technology, co-founded HRV4Training, advises Oura, guest-lectures at VU Amsterdam, edits for IEEE Pervasive Computing Magazine, and just released his first book, Heart Rate Variability: Science and Strategies for Peak Performance (Human Kinetics, May 2026).

This conversation covers the full arc of HRV, from the foundational science, to how to measure it correctly, to what actually moves the number, to how elite athletes and coaches use it to train smarter, to the surprising links between HRV, the brain, and burnout.

A quick note on this episode: shortly after we started recording, Marco had to turn his video off to make sure we could give you the clearest possible audio. He was joining us from Italy, and the connection wasn’t cooperating on the video side, but the conversation was too good not to bring to you, so we prioritized the story over the picture. Thanks for understanding!

In this episode:

• Marco’s path from machine learning to human movement science to HRV

• The most dangerous hype in wearables and biometrics right now

• The biggest HRV myths even serious athletes and coaches still believe

• Marco’s validity findings on Garmin’s Health Snapshot feature

• The minimum viable daily HRV measurement routine

• Vagally mediated HRV and the Vagal Tank Theory, explained simply

• Chest strap vs. wristband vs. finger sensor vs. phone camera — an honest accuracy ranking

• Whether continuous HRV monitoring is useful or just noise

• What the newest habit-and-HRV research actually shows

• Metabolic flexibility, fueling, and how it shows up in endurance athletes’ data

• How altitude, heat, and travel affect HRV — and what to do about it

• What HRV-guided training looks like day-to-day

• Tapering, race-day HRV, and team-sport applications

• HRV, cognitive function, mental fatigue, and burnout

• The Normal Range concept from HRV4Training Pro

• Where wearable and HRV monitoring is headed in the next 5 years

• The 30-Day HRV Challenge from Marco’s new book

Guest: Dr. Marco Altini — Co-founder, HRV4Training | Advisor, Oura | Guest Lecturer, VU Amsterdam | Editor, IEEE Pervasive Computing Magazine | Co-Author, Heart Rate Variability: Science and Strategies for Peak Performance (Human Kinetics, 2026)

Hosts: Klaudia Balogh (Longevity Journalist, longevityjournalist.com) & Don Moxley (Sports Scientist, trainrecoverwin.com)

🔗 Marco’s Substack: hrv4training.substack.com📖 The book is available now through Human Kinetics and Amazon

🎧 New episodes of the 90% Healthy Podcast drop weekly. Subscribe, comment your HRV questions, and let us know who needs to hear this one.

Watch here:

Or listen here.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

The Book

Our book is out and available for purchase from Human Kinetics’ website, here (or on Amazon).

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified Ultrarunning Coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, Endurance Coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Personal Substack

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

I hope all is well.

This one comes a bit late, as I was pretty sick last month, but that’s life sometimes.

Below is a new recap of articles, research, and updates from May and June.

I hope you’ll find the articles useful, and I would like to take the opportunity to thank you for your support.

Please feel free to comment below or in the articles should you have any questions, and I will follow up soon.

Take care!

Heart Rate Variability (HRV) 🫀

The Physiology of Endurance Running Podcast with Dan and Owain

I had a great conversation with Dan and Owain for their podcast (The Physiology of Endurance Running, which I’d recommend checking out regardless of this interview!).

We discussed pretty much everything from the physiological basis of HRV to applied protocols and limitations of current technologies, including pros and cons of using wearables, issues with stress estimates derived from continuous data, HRV during exercise, and more.

You can find the episode here:

1000 Days of Heart Rate Variability

In this blog, we look at a preprint from a single case study of 1000 days of HRV data, published by Amin Sinichi et al. from the Vrije Universiteit Amsterdam (link here).

His analysis first assessed how 1-minute versus 2-minute recordings showed close agreement, and how nocturnal HRV appeared more blunted and less sensitive to contextual changes (something at this point we’ve stressed plenty in the past, e.g. when discussing the orthostatic stressor or recent research highlighting how night data misses important changes visible in morning data).

In the author’s words: “the current results suggest that seated post-wake-up intentional recordings, rather than passively collected nocturnal averages, align more closely with real-life stressors”.

Validity Assessment of HRV from the Garmin Health Snapshot

A poster shared by Andrew Flatt on LinkedIn caught my eye (see here). The poster covers Kayla Porter’s work in evaluating the accuracy of Garmin’s spot check against ECG. In particular, as reported in the abstract, HRV was recorded simultaneously using ECG and the Forerunner 265’s optical sensor during Garmin’s 2-minute “Health Snapshot.” Measurements were obtained in three standardized positions: supine, seated, and standing. HRV values automatically generated by the Garmin application were compared with ECG-derived HRV.

The authors conclude that HRV estimates from Garmin are unsuitable for research purposes and patient or athlete monitoring based on their poor agreement with ECG, something we had seen already quite a few times in the past.

When in doubt, a good old Polar H10 is what I’d recommend for your morning measurements.

• Monitoring HRV: Why the Orthostatic Stressor is Best and How to Make it Practical

Low Heart Rate Variability (HRV) with Low Heart Rate: What Does it Mean?

I have received the following question:

“For the first time in one of my professional athletes, I’m seeing an HRV / HR trend (nighttime measurements) that seems unusual to me and that I’ve never observed before in other athletes: HRV is decreasing, while resting HR remains stable and/or is actually decreasing further. For example, today I observed one of the lowest HRV values ever for this athlete, but at the same time one of their best resting HR values. How should I interpret this?”

I think we have at least two situations to consider here, with different outcomes: parasympathetic saturation (not a problem) and actual fatigue (a problem). Learn how to distinguish them, here.

Training Talk 🏃🏻‍♂️

Chianti Marathon Trail by UTMB + Tuscany Crossing: The Training (and Racing)

I’ve updated my training (and race) report for this past season, including the second trail race I took part in (Tuscany Crossing), here.

You’ll find the high-volume training block between Chianti and Tuscany Crossing, plus heart rate density plots for the two races, HRV and heart rate between the events (with race day nerves showing quite clearly), nutrition adjustments due to changes in environmental conditions (i.e., warmer weather) and some more.

It’s been a really good season, despite the ending (sickness and no Passatore). I’m grateful for the process and hope you’ll find some useful pointers in here.

CH4 Trail ⛰️

A few weeks ago we organized a trail race here in Brisighella, something I had been working on for the past year together with Alessandra Saviotti and a few others who got us involved after we started spending part of the year back home.

It’s been wonderful to see so many old and new friends come together for a lovely morning on our trails (and some good food afterwards!).

Volunteering, directing, crewing my athletes, cheering, running around. What a day. My heart is full, and I cannot wait to start working on the next edition.

Thank you to everyone who joined us and helped make it happen.

🌼

Some more impressions from the race, here.

🌼

p.s. special thanks to Andrea Barra who captured our weekend so beautifully in the video below.

Stay in touch (and join us next year!).

That’s a wrap for today, below you can find other recent updates:

• April 2026

• March 2026

• February 2026

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

The Book

Our book is out and available for purchase from Human Kinetics’ website, here (or on Amazon).

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified Ultrarunning Coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, Endurance Coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Personal Substack

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

We discussed pretty much everything from the physiological basis of HRV to applied protocols and limitations of current technologies, including pros and cons of using wearables, issues with stress estimates derived from continuous data, HRV during exercise, and more.

You can find the episode here:

I hope you will find it interesting. Thank you for listening!

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

The Book

Our book is out and available for purchase from Human Kinetics’ website, here (or on Amazon).

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified Ultrarunning Coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, Endurance Coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Personal Substack

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

“For the first time in one of my professional athletes, I’m seeing an HRV / HR trend (nighttime measurements) that seems unusual to me and that I’ve never observed before in other athletes: HRV is decreasing, while resting HR remains stable and/or is actually decreasing further. For example, today I observed one of the lowest HRV values ever for this athlete, but at the same time one of their best resting HR values. How should I interpret this?”

I think we have at least two situations to consider here, with different outcomes: parasympathetic saturation and actual fatigue.

Parasympathetic Saturation

Parasympathetic saturation refers to a situation in which parasympathetic activity is particularly high, but this is not reflected accurately in HRV data. As Kiviniemi et al. explain, “possible physiological mechanisms underlying saturation could be due to the dose response of the heart to the acetylcholine secreted by vagal nerve ending. The dose response to acetylcholine has been considered to be linear until its concentration reaches the level at which a further increase in acetylcholine concentration does not produce a change in the response” - which means that despite an increase of parasympathetic control on heart rhythm, HRV does not increase. This is a positive state for the athlete, as stressed by my dear friend Daniel Plews: "in some circumstances, such as vagal saturation, decreases in cardiac parasympathetic indices of HRV during this particular training phase can be related to positive performance outcomes and consequently reductions in HRV, so should not be viewed negatively". Parasympathetic saturation is not a problem physiologically; it is a situation in which the measurement does not reflect autonomic nervous system activity.

How do we avoid the issue of HRV data not being representative of parasympathetic activity due to saturation? We can take care of this under most circumstances by using a good protocol (i.e., not using a wearable).

In fact, parasympathetic saturation is one of the reasons why I stopped measuring my HRV while lying down, and started measuring while sitting. If you switched as well, and your HRV is higher when sitting, it is possible that the data was saturated when lying down (as you would otherwise expect the opposite, with higher HRV when lying down with respect to sitting or standing). As sitting adds a little stress, it becomes less likely that you encounter a situation of saturation.

Please see this blog if you’d like to learn more about measurement position and why sitting is in my view the ideal position for HRV measurements.

What can you do about it?

Depending on how you measure your HRV, you could be proactive and collect data that is less likely to be affected by the issue of parasympathetic saturation.

In particular:

• if you measure your HRV during the night, there isn’t anything you can do. Hence you should use the procedure below to determine if parasympathetic saturation is likely in your own case. If there is a weak correlation between resting heart rate (or RR interval length) and HRV, it might be preferable to use a morning measurement.

• If you measure in the morning, you can measure while sitting (or standing), so that you add a little stress on your body and potentially prevent the issue of parasympathetic saturation, as recommended by Andrew Flatt.

Use HRV4Training Pro to determine the likelihood of parasympathetic saturation for you and your athletes

If you are in a period of high training load and HRV is low, together with low heart rate, and therefore the correlation between HRV and the average RR interval length is small or negative, parasympathetic saturation is plausible.

We have developed a feature in HRV4Training Pro to help you analyze this relationship. In the plot below, you would see the darker dots in the lower right corner (low HRV, low heart rate or high RR interval length). In this case, the suppression in HRV should not be interpreted negatively, as reported by ​Plews et al.: “the lack of correlation between the R-R interval and Ln rMSSD indicate that athletes are more likely to undergo parasympathetic saturation”.

​You can find the plot below under Insights / Resting Physiology in HRV4Training Pro, we also report the correlation between HRV and the RR interval length for the past 2 and 6 weeks, so that you can more easily spot any recent changes.

In this first example, you can see a high, positive correlation between HRV and RR interval length, both over 2 and 6 weeks:

In this second example, the correlations are lower, but still positive:

Saturation would likely show no correlation (near zero, as shown in Daniel Plews’ paper).

Actual Fatigue

An alternative scenario, especially if we were measuring in the morning, exploiting the orthostatic stressor (i.e. while seated or standing), and given a certain context, for example, a high training volume/intensity block, could be that we are indeed accumulating fatigue, and we are simply interpreting heart rate data incorrectly.

As a matter of fact, fatigue often manifests as a suppression in our physiology, and when this happens acutely, it should not be interpreted as a positive sign (i.e., while we might want heart rate to slowly reduce over months as our cardiorespiratory fitness improves, because of cardiac remodeling for example, an acute change in resting heart rate is unlikely to happen because fitness suddenly improves, and is often tied to fatigue).

Context will help us here, but there are fatigue patterns that are a good match with what is described in the initial question, i.e., suppressed heart rate, and HRV that remains normal or gets also suppressed. Keep in mind that heart rate and HRV are not the same thing (or the inverse of one another) and HRV is more sensitive to stress than heart rate, hence training stress, which is more subtle than e.g. sickness or other stressors that have a large impact on resting heart rate, might be reflected only in HRV data at times, which is why we go through the trouble of measuring it. This means that we should pay attention to these suppressions in HRV even if heart rate is normal, and possibly we should pay even more attention to suppressions in HRV when heart rate is lower than normal, which might be an additional signal that we are accumulating fatigue.

Normalized HRV in HRV4Training Pro

A good way to look at the data in these cases is by combining HRV and heart rate, using what I have called Normalized HRV, which you can also find in HRV4Training Pro.

As I’ve discussed in a dedicated blog, the most important question for me is this: does normalization of HRV by heart rate improve the actionable insights we derive from HRV data? My current take is that the answer is yes, especially if we have a stable or below normal heart rate, which could be the case in many situations (e.g. dieting, but also cold exposure, accumulated fatigue, or higher parasympathetic activity associated with fatigue, more than recovery).

A recent paper by Bulte et al., looked at Normalized HRV in relation to training load in state and international-level swimmers, concluding that it “may provide a more comprehensive evaluation of an athlete’s physiological status compared with a sole parasympathetic modulation or tone marker“ - given their results in which Normalized HRV was the only parameter that correlated with training overload (while heart rate or HRV didn’t).

Alright, that’s all for today on parasympathetic saturation and fatigue, I hope this was informative. As you can tell, in both cases, additional context is key for data interpretation (e.g., training load, individual responses to previous blocks, other life stressors that are not training related, as well as subjective feel). On top of that, make sure to use good protocols and good tools, and the interpretation of the data will also be easier.

Cheers.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

The Book

Our book is out and available for purchase from Human Kinetics’ website, here (or on Amazon). Thank you for reading!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified Ultrarunning Coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, Endurance Coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Personal Substack

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

Table of Contents

• Foreword
  Introduction. Why This Book? Starting the Journey With HRV
  What Are Our Stories With HRV?
  Why Is HRV Important?
  How Can This Book Help You?
  Final Thoughts to Kick-Start Your HRV Reading Journey

• Part I. The Science Behind HRV
  Chapter 1. Physiology Foundations
  Understanding the Physiological Foundations of HRV
  What Is the Difference Between Heart Rate and HRV?
  The Physiology of HRV
  HRV Indices
  Vagally Mediated HRV
  Cardiac Vagal Activity and Vagal Tank Theory
  The Implications of Vagal Tank Theory
  Myths About HRV: Separating Fact From Fiction When Interpreting HRV
  Interpreting Indices of Non–Vagally Mediated HRV
  The Role of Context in HRV Interpretation

• Chapter 2. Global Factors That Influence HRV
  Population-Level Analyses: Understanding Individual Differences
  Individual-Level Analyses: How Data Change Over Time
  Environmental Stressors

• Chapter 3. Impact of Sleep, Nutrition, and Hydration on HRV
  Sleep and HRV
  Nutrition and HRV
  Hydration and HRV

• Part II. Measurement and Interpretation of HRV
  Chapter 4. HRV Monitoring and Tracking Tools
  Common Methods and Devices for Measuring HRV
  HRV Versus Pulse Rate Variability
  Know Your Application: ECG Versus Photoplethysmography Versus Ballistocardiography
  Protocols and Best Practices: How to Accurately Measure HRV
  When to Measure HRV
  Body Position: Should You Measure Lying Down, Sitting, or Standing?
  Frequency of HRV Measurements
  Duration of HRV Measurements
  Chest Straps, Wristbands, Finger Sensors, and Phones: What Can You Use?
  The Impact of Breathing on Data Collection
  Artifacts in HRV Data: Movement, Ectopic Beats, and Arrhythmias
  Learn to Ask the Right Questions

• Chapter 5. Interpreting HRV Data
  Acute Response to Training Stress and HRV
  Understanding Meaningful Changes in HRV Indices: The Normal Range
  Identifying Changes in HRV and What They Mean for Your Health and Fitness
  Heart Rate and HRV Saturation
  Physiological Adaptations to Training

• Part III. Applications of HRV Monitoring
  Chapter 6. Using HRV to Guide Individualized Training
  Autonomic Regulation and Trainability: What HRV Can Tell Us
  The Beginnings of HRV-Guided Training
  State of the Art in HRV-Guided Training: The Big(ger) Picture
  The Future of Individualized Training: A Combined Approach?
  Expected Outcomes of HRV-Guided Training
  Beyond Baseline Assessments

• Chapter 7. Applying HRV in Team Settings
  Devices for Monitoring HRV in Team Sports
  The Basics of HRV in Team Sports
  Challenges of Monitoring HRV in Team Sports

• Chapter 8. Injury Prevention and Acute Recovery
  Using HRV Monitoring to Detect Early Signs of Overtraining and Injury
  HRV-Guided Training for Injury Prevention
  HRV-Guided Training for Injury Rehabilitation
  HRV for the Management of Brain Injuries and Concussions

• Chapter 9. Health, Disease Prevention, and Longevity
  The Role of Autonomic Function in Health and Disease
  HRV and Fundamental Health Mechanisms
  The Role of HRV in Disease Prevention and Management
  Improving Health Outcomes With HRV Monitoring

• Part IV. The Brain’s Impact on HRV
  Chapter 10. HRV and Cognitive Function
  HRV, Mental Performance, and Function
  Brain Structure and HRV
  The Prefrontal Cortex, Executive Function, and HRV
  Measuring HRV for Cognitive Performance
  Physical Activity and HRV
  HRV and Mental Fatigue
  Collecting HRV Data to Manage Mental Fatigue and Improve Cognitive Functioning
  Strategies for Overcoming Mental Fatigue and Improving Executive Performance

• Chapter 11. Managing Psychological Stress
  Understanding Psychological Stress
  Strategies for Using HRV to Manage Psychological Stress
  Burnout and HRV
  Challenges and Limitations of Monitoring HRV to Manage Psychological Stress

• Chapter 12. HRV and Mental Training
  Psychological Skills: The Foundations
  Breathing Techniques
  Mind–Body Relaxation
  The Low HRV Autopilot
  Inspirational Techniques
  Self-Reflection
  Connection Matters: The Power of Social Support for HRV

• Part V. Maximizing Daily HRV Monitoring for Athletic Performance
  Chapter 13. Your Daily HRV Routine
  The Ease of Capturing HRV Data
  What Could Your HRV Day Look Like?
  HRV Analysis Before and After Exercise
  What About Continuous HRV Monitoring?

• Chapter 14. A 30-Day HRV Training Plan
  Why Complete an HRV Challenge?
  Your HRV-Supporting Morning Routine
  Your HRV-Supporting Evening Routine
  Building Blocks for HRV Awareness and Improvement
  The 30-Day HRV Challenge
  What to Do After the 30-Day HRV Challenge

• Epilogue. Your HRV Journey Is Just Starting
  Summary of Key Takeaways
  Personal Growth and Continuous Learning
  Start Simple and Build a Solid Foundation
  HRV for Athletic Training and Performance
  Interpreting Data and Making Training Decisions
  Overview of HRV Applications Across Sports
  Sharing Your HRV Journey

Backstory

Here is a bit of the backstory on how this unfolded over the past seven years ☺️

Back in 2019, Dan and I started discussing the idea of writing a book on HRV. At the time, there was nothing on the market that offered a truly modern and practical perspective. We began writing and drafted several chapters before and during the pandemic, but eventually lost momentum.

In October 2022, I met Sylvain at the Italian Olympic Committee, where we delivered a workshop on HRV together. We put together a decent show, and afterward, he asked if I would be interested in writing a book. I discussed it with Dan, and we decided to join forces. We also involved Emma, who had previously worked with Sylvain and brought invaluable expertise, especially in HRV and cognitive performance. Cooperation over competition.

This time, Sylvain’s incredible energy set the pace. He helped us find an editor we immediately connected with: Human Kinetics. We signed our agreement in 2023, and from that point on, we spent more than two years writing, revising, and iterating with our editors.

All of us contributed to every chapter. We learned from each other and gradually shaped what we can now confidently call our HRV book, reflecting decades of combined experience in the field. It has been a long journey, and I am truly grateful to my coauthors for making the process genuinely enjoyable and intellectually stimulating.

Our goal has always been to stay between hype and skepticism, so that people can truly benefit from this tool by understanding both the potential and limitations of HRV analysis.

Well, that’s all.

I hope you will enjoy the book!

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

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

Listen here.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

I hope all is well.

Here’s a new recap of articles, research, and updates from the past month. I hope you’ll find the articles useful, and I would like to take the opportunity to thank you for your support.

Please feel free to comment below or in the articles should you have any questions, and I will follow up soon.

Take care!

Heart Rate Variability (HRV) 🫀

The Uphill Athlete: Making Sense of the Data with Dr. Marco Altini

It was a real pleasure to join Steve House - author of one of my favorite endurance training resources at The Uphill Athlete - for a chat about Heart Rate Variability (HRV) and training, a few weeks ago.

You can watch it here:

Or listen here:

I hope you’ll find some useful pointers in there.

Environmental Temperature and HRV

I think one of the most underestimated drivers of change in our resting physiology is simply the environment we live in (and for example, environmental temperature). We’re always quick to point to this or that behavior as the cause of a change in our physiology, but on many occasions, it is not so simple. This is one of the main reasons why I’d argue against using HRV as an outcome measure, especially following any positive behavioral change you might be implementing, as the change you’ll be seeing (or not seeing) in HRV is likely impacted by your environment as well, and not only by the parameter you are manipulating (e.g., diet, exercise, or else).

Learn more, here.

Do Healthy Habits Increase HRV?

An interesting new study adds a bit of data to some of the points I’ve tried to detail in the past when discussing factors behind individual differences in HRV (see a paper I wrote with Dan Plews here, and a blog here). In particular, the new study (Saraiva et al., see here), looked at the relationship between everyday lifestyle behaviors and HRV in a group of adults. Participants were classified based on whether they met basic health guidelines across three domains: physical activity, sleep duration, and sedentary time.

Learn more about what the authors found, here.

Training Talk 🏃🏻‍♂️

In April, I raced the second (and last) race of my trail running season. It was a pretty good day at Tuscany Crossing, 53k +1500m of incredible scenery in Val D’Orcia.

Some numbers:

• ⏱️ I finished 16th of about 500 runners, pacing evenly the 4h 43’ it took me to get across the Tuscan hills and to the finish line.

• 🫀I averaged 151 bpm, or ~85% of my maximal heart rate, which is exactly where my first threshold sits. The boundary between Z2 and Z3 (on a 5-zone system), is the highest intensity I can sustain for this duration on a trail race (as in Chianti as well, which I covered in more detail here). See also some useful heart rate density plots here.

• 🍝 I had ~40 g of CHO/hour: 2 flasks (90 g each), and one gel (30 g + 100 mg caffeine). I had planned slightly more, but had to adjust on the go as the warmer day made it a bit more challenging. Given my needs (which I had detailed following a metabolic test on race week), it wasn’t an issue, and I felt strong the whole way.

Last but not least, I caught up with countless old and new friends, which always makes these weekends so special. Thank you.

CoachCorner ⏱️

Bruna’s Trail Marathon Debut at Istria by UTMB

A recap of the past year working with Bruna, building from a stress fracture up to a 6th place in her first trail marathon (Istria by UTMB), in this blog. Grateful to have the opportunity to work with you, Bruna Maia.

CH4 Trail ⛰️

[Italian]

Stiamo organizzando una gara di trail per il weekend del 14 giugno qui in zona Brisighella. Per chi fosse interessata/o, informazioni aggiuntive seguiranno su Instagram. Grazie!

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Recent Updates:

• March 2026

• February 2026

• January 2026

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

Here is where we were in June 2025, alternating a few minutes of running and walking, with cross-training:

while here is Bruna a few weeks ago, finishing 6th at her first trail marathon at Istria by UTMB (42 km +1000 m):

This post is about how we went from A to B, which was mostly about common sense, communication, trust in the process (also called patience), and, of course, a bit of luck.

But let’s start from the beginning.

I met Bruna when I was a guest on the Human Endurance podcast that she is hosting, and while the conversation was mostly on HRV and somewhat data driven, I think we clicked because of the nuance around it, and how eventually we think very similarly about training and how to develop as athletes, getting in tune with perceived effort, building gradually, thinking long term, and sometimes using data to aid the process (Bruna is a running coach as well).

As mentioned above, we started working together after her second stress fracture, rebuilding with a few walks in May 2025, and eventually doing our first continuous run on July 14th, 2025, which was our first 30 minutes of uninterrupted running in over 2 months. By August, we had added a few strides. The number one goal was always not to get injured, hence there was great focus on fueling (something Bruna has talked about in her blog) and on watching any niggles, but also in terms of training prescription, we never tried to push much running frequency or volume, given the past history of stress fractures.

In situations like these, it is key to have a plan, but it is even more important to be willing to adjust that plan should anything prompt more caution or simply should the circumstances around us change (because life happens, which could mean more work stress, travel, etc.). Bruna is also the founder of , (where I’m an advisor), a tool we’ve been using for feedback, communication, and that helped greatly in staying aligned and making all the necessary adjustments while working together remotely. Thanks to Augo, it was easier to stay on top of our nutrition and of any running-related niggles or difficulties that prompted adjustments to make sure we could get to race day as healthy as possible.

In the next section, we’ll look at how we planned the comeback.

High-Level Planning

Building from zero to Bruna’s first trail marathon, I mainly had two phases in mind, broken down as follows:

• Generic Fitness:

  • Cross-training: we used cycling a lot as a low-risk tool to develop aerobic capacity. We also used strength training to put some more mechanical load, possibly impacting positively both injury risk and durability (for trails).

  • Develop running frequency: coming from an injury, we wanted to be able to run 3-4 times per week for a while, before adding more stress on the body. This was our first running goal.

  • Add running intensity: once we had developed frequency, we could start layering on top various intensities, to further develop our physiology (aerobic capacity, fatigue resistance) as well as our structural capacity.

  • Consolidate running frequency and intensity: we were going through big changes, and we could not simply keep adding without risking too much. At a certain point, we needed to decide we had reached a level we were happy with and consolidate it so that the body could make this new level its normal. This will be our base in the future, when we’ll get even better.

• (Trail Marathon) Support + Specificity:

  • Develop durability: in the Generic Fitness phase, we kept long runs relatively short (~20km). Then, given our goal (trail marathon), we needed to work on our ability to sustain efforts twice as long, at least.

  • Develop trail skills: Bruna had mostly raced roads in the past, hence we needed to work on specific trail skills (climbing, hiking/running, descending quickly, etc.).

  • Nutrition: another key aspect for any long race or ultramarathon, as we are staying out for many hours and need to fuel the effort properly, while also making sure our gastrointestinal system can handle the type and amount of nutrition we are consuming.

  • Consolidate all of the above: as in the Generic Fitness phase, Specificity also adds many new stressors, especially on the mechanical side of things (longer runs, vertical gain, etc.) - hence, we cannot keep adding until race day, but need to develop our skills up to a point, and then consolidate.

Training prescription was always done using perceived effort, which I have discussed in more detail here and here. Obviously, we use all data available, pay attention to how our body responds in terms of internal load (e.g., our physiology), keep track of progress in key sessions in terms of external load (e.g., pace for track workouts), and can use both internal and external load as cues, caps or targets for specific sessions, but overall, we mainly aim at mastering our perception of effort at different intensities and for different durations, as we develop at athletes.

Generic Fitness: details

Quite importantly, there was no timeline attached to building Generic Fitness. We went on for as long as it was needed to reach a capacity that we thought would allow us to move forward without risking breaking down.

This is not an easy task, as we always have some event coming up, and at times, we need to simply recognize that we need a little more time, and patiently keep doing the work, consistently, until we have built that basic, generic fitness capacity required to race successfully. For example, Bruna was initially supposed to run the Chicago marathon in October, but having just done only a few strides and 20-30 km / week of running up to August, we talked it through and decided it was too risky to push a marathon block in September. This seems an obvious choice after the fact, but it’s not an easy one to make.

Looking in a bit more detail at the various aspects discussed above, we can first see that cross-training was a key aspect that allowed us to build and maintain fitness when running was either not possible or simply limited:

From the graph above, we can also see that the volume of cycling reduces over time, as we are able to run more. Below is our running volume, which we kept increasing, but only up to a point, as we noticed we could not go beyond 4 runs / week without setbacks, for now. There’s no golden rule for this; sometimes we had to try things (e.g., a slightly denser block), and see what would happen:

When looking at all workouts, we can see how weekly load was rather even across the year, even if we gradually reduced cycling and increased running as our capacity improved.

We found our sweetspot at 4 runs + 2 rides per week, 1 day off, 1 strength session, typically with one hard workout and one longer run, as part of our 4 weekly sessions:

In terms of intensity, we were in a phase (rebuilding fitness) where I think variability in the stimulus was more important than the specifics of any single session. The priority was not optimization, as it might be when preparing for a road marathon, but rather reintroducing and developing capacities that had not been stressed for a long time. This ranged from aerobic power to fatigue resistance, and eventually durability at race pace.

As a result, we touched on multiple systems and intensities, alternating different types of workouts across weeks and months, without dedicating extended blocks to a single stimulus. The goal was broad development rather than maximizing any one aspect of our physiology.

We ran short intervals, longer intervals, tempo blocks, strides, all with a specific purpose and logical progression of course, but without stressing any single aspect in particular.

By the end of the year, training was rather stable in terms of frequency and intensity, and we found a good balance and consistency, as well as improved performance, and as such, it was time to move to the next phase, picking a race and getting specific for trail running.

(Trail Marathon) Support + Specificity: details

I typically split more specific work into a Support block where we start e.g. doing trail runs, getting used to different terrain, vertical gain, and so on, but maintain ‘standard’ hard workouts, and an actual Specific block where workouts are also moved to race-terrain, and often get combined with long runs, so that we really stress durability and assess race capacity (e.g. long runs with tempo blocks or hard climbs/quick descents, etc.).

Depending on the athlete, this block might be very long or very short. For example, I discussed in the past how durability is a big limiter for me, and as such, how I’ve dedicated half a year recently to getting better at the longer distance from a muscular point of view (it worked). For Bruna, durability is a strength, and as such, we could limit our specific work to 2 blocks (4+4 weeks):

In terms of vertical gain, living in Zurich, we had some exposure throughout the year, even in less specific blocks, but eventually pushed it quite a bit in the past 2 blocks (again 4+4 weeks):

During these last 2 specific blocks, we kept running 4 times per week but often added a second long run on trails mid-week(about 90 minutes). Here are some of our key specific sessions:

• Long runs on trails: 3h, 3h 30’, 4h, over three weeks (as intensity, RPE 2-3, or 10-15 bpm below the first threshold, hence quite a bit easier than race pace).

• Hard workouts on trails or Long runs with race intensity (or slightly harder) blocks, e.g., 3 x 10’ @ threshold uphill, or 2 x 45’ moderate on trails. Once we had developed the capacity to run long and to run hard, we combined them.

To limit risks when adding very long runs or long runs with intensity, we often dropped intensity from the rest of the week, and typically added an indoor bike workout as the only high-intensity stimulus. In this final phase, we weren’t chasing fitness as much as we were trying to get ready for the specific demands of the event, learning to feel comfortable at race intensity, sorting out nutrition, shoes, and so on.

It’s always a balancing act.

Finally, it wasn’t all smooth sailing of course: we had a few more niggles than we’d liked (e.g. had to switch back to flat running during the specific block as we were having some issues when running downhill), and we had a few trips in which there was nothing but a treadmill to use for training, but we stayed focused on the long term goal, which was really not about doing ‘the perfect week’ or ‘the perfect workout’, but mostly about keeping doing the work consistently, day after day, maintaining good frequency, and adding layers of specificity as we got closer to the race.

The Race

For the race itself, we discussed strategy in terms of pacing/managing intensity using perceived effort as well as a heart rate cap for the first climbs (see also here), which was executed perfectly:

and which led to a fun race, overtaking the ones that probably weren’t as wise at the start, so to speak:

There is, of course, more to it, but I will leave you to Bruna’s blog for all the details about her race, which she’s covered brilliantly here.

Wrap up

It’s been a good year, one made of small steps toward a bigger goal, where common sense, patience, and close communication played the biggest role, overshadowing the things I often talk about here (a lot of talking, not much data!).

Of course, it’s easy to praise ourselves when things go well. We shouldn’t forget that, despite common sense (building gradually, keeping running frequency low, using cross-training a lot, etc.), patience (canceling races and taking more time to rebuild), and well-thought-out nutrition (pre, during, and post-workout) and training (from long-term periodization to weekly details), randomness always plays a role. We were certainly lucky that everything came together and that no small issue turned into something more serious over the past year.

We’ll see what the future holds, but for now, congratulations Bruna, and thank you for your trust.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

In this episode of Simple Science, I’m joined by Dr. Marco Altini — runner, researcher, and expert in heart rate variability (HRV) and metabolic flexibility.

We dive into how endurance athletes can optimize fueling strategies, improve fat oxidation, and better understand their physiology through both data and intuition. Marco shares how he has periodized his carbohydrate intake based on training demands, allowing him to become more metabolically flexible and efficient during long races.

We also explore the role of HRV, how lifestyle factors influence recovery, and why the best athletes learn to balance objective data with subjective feel.

This is a valuable conversation for endurance athletes, coaches, and anyone interested in performance, recovery, and long-term training sustainability.

Topics include:
• What metabolic flexibility actually means for endurance athletes
• Periodizing carbohydrate intake based on training demands
• Training the body to rely more on fat during long efforts
• The role of HRV in monitoring recovery and readiness
• Data vs intuition in endurance training
• How sleep, stress, and lifestyle impact performance
• The future of wearable technology in human performance

Watch here:

Or listen here:

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

From the show notes:

“Steve House sits down with Dr. Marco Altini — data scientist, physiologist, and founder of HRV4Training — for a deep-dive into one of the most misunderstood tools in endurance athletics: heart rate variability. With over 50 peer-reviewed papers to his name, Dr. Altini brings rare scientific rigor to a space crowded with oversimplified wearable metrics. The conversation opens by challenging the industry's obsession with single-number readiness scores, which Dr. Altini argues dilute nuanced physiological signals into deterministic assumptions that don't reflect the complexity of the human body.

The episode unpacks the fundamentals of HRV — what it actually measures, why morning is the optimal measurement window for endurance athletes, and how to interpret it as a marker of stress response rather than performance capacity. Steve and Marco also cover hardware accuracy across chest straps, optical sensors, and phone cameras; HRV research at altitude; and the broader question of how biometric data should support — not replace — the athlete's own sense of feel.”

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

Most runners track HRV, but very few actually know how to use it correctly. In this conversation, Dr. Marco Altini explains what HRV can tell you, what it can’t, and how to use it to make better training decisions without becoming a slave to the data.

Dr. Marco Altini is a physiological data scientist, HRV researcher, coach, and the founder of HRV4Training. Over the last 15 years, he has worked at the intersection of physiology, wearable technology, machine learning, and endurance performance to help athletes better understand stress, recovery, and training readiness.

In this episode, we go far beyond the usual “high HRV is good, low HRV is bad” conversation. We break down how HRV should actually be interpreted in the real world, why resting heart rate still matters, when HRV is useful for adjusting training, and why context matters more than any single number on your watch. We also discuss how HRV changes with age, how life stress compares to training stress, whether HRV can predict overtraining, how tapering affects HRV before a race, and whether HRV is really a meaningful marker of long-term health and longevity.

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

I hope all is well.

Here’s a new recap of articles, research, and updates from the past month. I hope you’ll find the articles useful, and I would like to take the opportunity to thank you for your support.

Please feel free to comment below or in the articles should you have any questions, and I will follow up soon.

Take care!

Heart Rate Variability (HRV) 🫀

HRV, Training Load and Cardiorespiratory Fitness (VO2max)

In the past month, I covered a recent study by Kaimusik et al, titled “Effects of seasonal training on physical fitness and heart rate variability in amateur Futsal players“, as I think there are some interesting pointers in the context of the relationship between training load, fitness, and resting physiology (heart rate and HRV).

The measurement protocol was aligned with what I recommend, i.e., measuring seated and while breathing naturally, self-paced, in the morning. This type of protocol is designed to minimize external influences on autonomic activity while amplifying the physiological response thanks to the orthostatic stressor, something we miss when measuring lying down (or sleeping!). See also here for a broader discussion on this point.

Keep reading, here.

HRV and Seasonal Allergies

A few weeks ago I was chatting with an athlete I coach who reported a very large drop in HRV due to seasonal allergies, hence, I figured we could expand a bit on the topic for today’s newsletter.

Seasonal allergies can, of course, become an important issue for many people, but are not often discussed in the context of endurance training. From an HRV perspective, however, on most occasions, we can see clear changes due to allergies, as HRV reflects overall stress on the body, and as such, it can help us back up how we are feeling, and potentially make meaningful adjustments. While there is somewhat limited research on the topic, the nervous system response in the context of allergies is quite interesting and not necessarily intuitive.

Learn more, here.

Publications

HRV Coefficient of Variation

Recently, I was involved in a nice study, titled “Heart rate variability coefficient of variation during sleep as a digital biomarker that reflects behavior and varies by age and sex“ (thank you, Greg), looking at the HRV Coefficient of Variation, in relation to certain behaviors and other health-related aspects in a large population. In particular, as the study shows, a higher HRV Coefficient of Variation is associated with greater alcohol consumption, lower activity, shorter and less consistent sleep, older age, and higher BMI.

More info, here.

HRV-Guided Training

Finally, a new narrative review was just published covering HRV-guided training, a topic we’ve discussed a few times in the past (no involvement in this article). In particular, the review looked not only at RMSSD (our marker of parasympathetic activity) but also its coefficient of variation (see week 10 of this newsletter) in relation to interventions using HRV to adjust training at the individual level, based on their own physiological response to the imposed load.

The authors report how “HRV-guided training improves aerobic capacity and endurance more efficiently than standard programs, even with fewer high-intensity sessions. It allows training during optimal readiness, prioritizes recovery when needed, reduces non-responders, and facilitates precise load regulation. Regular HRV monitoring with individualized interpretation of RMSSD and CV supports safe, personalized training. HRV-guided approaches optimize adaptation, manage fatigue, and minimize overtraining risk in recreational athletes.”

The full text of the article is available here.

If you are curious to incorporate this method in your own training, please see the two articles below for guidance:

• A brief history of HRV-guided training

• When to hold back and when to PUSH

Training Talk 🏃🏻‍♂️

Chianti Marathon Trail by UTMB: The Training (and Racing)

I’m just back from a lovely weekend meeting old and new friends while racing what was probably my best trail race at Chianti, and have a few things to say about how differently I prepared for it, which I hope you’ll find useful.

In particular, in my latest training report, I go through the reasoning behind my long-term periodization before Chianti, data on metabolic flexibility, how I’ve addressed muscular limiters and improved durability with strength training and hard running on hills, the use of cross-training and passive heat, race-day fueling, and more.

See here.

Building 🛠️

Joining Augo as an advisor

I’m excited to join Augo as an advisor, a female-founded Swiss startup seeking to establish a new standard for endurance sports coaching: one where coaches and athletes feel more connected, communicate fluidly, and share a deeper understanding of training, performance, and progress.

I first met Bruna and Fabi quite some time ago when chatting for their podcast, and loved their vision to build a better tool to support coaches. Augo has just launched last week, and you can learn more here.

Augo is also running this survey with athletes to understand why we choose to work with a coach or not, and what could be better about our experience in either scenario. Runners, cyclists, and triathletes: you can win $200 USD from On by filling in the 6-minute questionnaire.

Interviews and Podcasts🎙️

I was in The New York Times a few days ago as the only skeptic in the room, as it often happens. Thanks Noam for the chat and you can find the article here.

CH4 Trail ⛰️

[Italian]

Stiamo organizzando una gara di trail per il weekend del 14 giugno qui in zona Brisighella. Per chi fosse interessata/o, informazioni aggiuntive seguiranno su Instagram. Grazie mille.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Recent Updates:

• February 2026

• January 2026

• December 2025

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

• Chianti Marathon Trail, 46 km +1900m, 95th / 1400, 9th age group.

• Tuscany Crossing, 52 km +1500m, 16th / 500, 3rd age group.

My last race report was after the Winschoten 50 km, 6 months ago, which was arguably my best road race, running five nearly identical laps at the limit of my ability. While my A goal remains a road race (did I tell you about Passatore yet?), I’ve changed my mindset and goals since last summer, and I’m gradually shifting to racing trails. I think this is partially due to having reached my limits (and the goals I had set for myself) on the road - and therefore being less motivated - and partially due to the way racing trails adds layers of complexity that challenge me (technical terrain, steepness, fueling, temperature, etc.) while also amplifying aspects of exercise that I value (being in nature, the focus on perceived effort, being more aware and in tune with my body, variability in effort and race strategy, being less obsessed by hitting an exact pace every kilometer, etc.).

Just like for Winschoten, this blog is not really about the race, as much as it is about the training, which I think is where I might have something more interesting to say, given my average performance level. In particular, I’ve prepared quite differently from the past, given how I’ve grown as an athlete and the different demands of the race with respect to road running.

Hence, I will go through the reasoning behind my long-term periodization, data on metabolic flexibility, how I’ve addressed muscular limiters with strength training and hard running, the use of cross-training and passive heat, race-day fueling, and more.

I’ll add a bit of race analysis in terms of heart rate and pacing as well, because that’s relevant too.

Big Picture: What’s Missing and Where Can I Improve?

Some context here is needed. I love to run the whole day, but not beyond that. Say 4-10 hours, which could be on roads or trails, and ideally include a good scenery. I dislike running in loops, especially for long events, and I’ve learned somewhat surprisingly that one of my strengths is climbing, so a fair amount of that is something I now enjoy and chase, as long as it’s runnable (i.e. not in the mountains). We are talking ultratrail, say 40-60 km with ~1000-2000m of vertical gain up to Passatore (road, 100km of mostly rolling hills).

I’ve mentioned in the past how the work I’ve done on metabolic flexibility has improved greatly my durability for these events, from a metabolic point of view, of course. Durability is multifaceted, and while now I can rely a lot more on fat oxidation, save glycogen, limit depletion, delay fatigue, and race personal bests, there is one more aspect of durability (which we could define as our ability to maintain a desired output for a looong time), limiting my performance in these events, i.e., muscular durability.

Having 1) a relatively high VO2max for my performance level (~65 ml/min/kg) 2) the ability to hold an intensity quite close to it for about 80’ minutes (i.e. a strong Threshold as well), 3) pathetic running economy (~230 ml/min/km), 4) a highly flexible metabolism able to burn over 1 gram of fat per minute above race intensity, and 5) having trained high volume for many years, I believe there is very little margin for me to race faster by addressing any of these 5 aspects (my VO2max does not change anymore, my threshold is already as good as it gets given my ceiling, my economy also seems not to move - unless I wear supershoes, 5-10% better at that point - my fat oxidation is basically elite level because of the perfect mix of training volume, avocados and genetics, and I cannot possibly train more than ~20 hours per week without burning out). What’s left out in the context of improving performance for the events I’m interested in is muscular durability, which is ‘so bad’ (frequent cramps and other difficulties) that it feels like even a reasonable improvement could lead to meaningful race performance gains.

So that’s what it is: to get better at the races I enjoy, I needed to condition my muscles to be more durable.

How do we do that? Below is the approach I’ve used in the past 6-7 months.

Periodization: Building Muscular Durability

As it often happens in training, especially in the era of social media and algorithmic engagement, different schools of thought tend to get polarized, so that we have the ones claiming you only have to run, and the ones claiming you can do it all in the gym, when it comes to improving muscular durability (or muscular endurance).

Fortunately, I am not a religious man, and I have friends on both sides of this ‘fight’. Also fortunately, my training is not a scientific study where you need to isolate a single component, and so I can throw lots of things in the mix and only care about the outcome (am I more durable?). The ingredients might change (forms of cross-training, running workouts, etc.), depending on a number of factors (neural capacity was a big one for me in these months, while it wasn’t in the previous build), which is the beauty of the art of training and coaching (get coached?).

Needless to say, I’ve used both the gym and hard running workouts to improve my muscular durability. In particular, this was my reasoning:

• Gym: it should be possible to improve the maximal strength of my (leg) muscles. By doing so, I might be able to delay fatigue when racing at a relatively lower percentage of my (then higher) maximal capacity. Think maximal strength.

• Gym: it should be possible to improve the durability of my (leg) muscles by training their endurance capacity. Think strength endurance, just more training or more stress for specific muscles.

• Run: it should be possible to do the two things above with more volume of hard running on hills. Yes, running hills is a good strength workout, but when you are very well conditioned, train a lot, run a lot of vert on a weekly basis, and still have muscular durability issues, then the principle of supercompensation still applies: you have to get beyond your capacity to get better. This means that ‘running hills’ isn’t good enough, but you might have to run harder and steeper. Cook those muscles (sometimes!).

In terms of periodization, I’ve always done little in the gym in the past, or used it more for health (mixed sessions upper and lower body without a specific focus or periodization or anything that would cause fatigue during running), while this time I decided to give the gym the number one priority in my training. It was a bit of a bet and certainly the biggest change in my training in years, in particular because I did not have the capacity to just add it, but to do so, I had to drop completely any hard running in my training for several weeks.

Hence, the high-level periodization of these past months:

• October: closing racing season.

• Capacity. November - mid-December: High Volume of easy running (peak week 180km), 2 x week in the gym.

• Maximal Strength. mid-December - February: bringing back my Threshold, 1 x week in the gym (sometimes 2 x, but mostly unable to do it well that frequently).

• Endurance. March: Tempo for Race Specificity and fitness, 1 x week gym maintenance.

I’ve also done 2 x extra core sessions in the gym, those were always the same and always carried out across the 6 months, not periodized, as they don’t create collateral damage (hence won’t be discussed below).

In the next section, we look at this periodization in more detail.

Details: Strength in the Gym and Hard Running Workouts

October: closing racing season

After Winschoten in September (50 km +0m, road), I raced Big Rock (47 km +1000m, trail) and Trail del Cinghiale (43 km +2600m, trail), underperforming in both races. At Big Rock I misjudged my ability to sustain a certain intensity on the trails. As I was coming from summer in Amsterdam, I thought I could race closer to Winschoten intensity, and faded badly at 30 km. At Trail del Cinghiale, I learned my lesson and paced well, but all my (muscular) limits were really obvious; I cramped after 20km already in both adductors, and in general struggled a lot in a race where you basically never run (too steep and technical for me). Again, coming from summer in the lowlands, that’s okay, but the whole experience reinforced the need to address this specific performance limiter, and that training volume alone wasn’t cutting it for me (I had done multiple 150km+ weeks with plenty of vert in that build).

Capacity. November - mid-December: High Volume of easy running, 2 x week in the gym

In November, I planned my Strength training as follows:

• Base Block: develop basic capacity for something new. I kept it very simple to begin with, 4-6 exercises (Squat(x1-2), Lunge(x1-2), Hinge, Calf). To load the exercises I used a weighted vest for practical reasons (it works for all exercises, I can increase the weight easily, etc.). For this block, I used 3-4 sets of ~8 reps with moderate weight.

In terms of running, I had to pause high-intensity workouts because I never had the required freshness (due to the gym). This break lasted about 5 weeks, then in mid-December I was able to run hard again (while maintaining 2 x gym sessions).

Maximal Strength. mid-December - February: bringing back my Threshold

Between mid-December and January, the gym was the same, but I also did 6 Threshold sessions (one per week tops, sometimes one every 2 weeks):

• 25’, rolling hills (<3% grade)

• 2 x 12’, one hard downhill, one hard uphill (6-12% grade)

• 3 x 10’ indoor, flat

• 2 x 12’ uphill on trails (10-20% grade)

• 20’ + 2 x 10’ indoor, flat

• 2 x 12’ uphill (6-12% grade)

After this month, my threshold was as good as it gets for me (about 3’45”/km), despite the previous block without any running intensity. The lesson there is that our strengths come back really fast; there’s no need to stress about them.

In terms of gym work:

• Support Block: develop ‘maximal’ strength. Similar structure to the base block, with 4-6 exercises (Squat(x1-2), Lunge(x1-2), Hinge, Calf) but with higher weight and lower reps (I used a bit more than 20% of my body weight for the vest, 3-4 sets of 4-6 reps).

Endurance. March: Tempo for Race Specificity and Fitness, 1 x week gym maintenance

Phase three was about training hard and for long (what I call Tempo in my training prescription) which of course is all about specificity for long (but not extremely long) events as well as pushing race fitness even higher. In this phase, I had to drop the gym to 1/week as the long and hard runs would leave me too tired for 2 sessions per week. The gym done well was still the highest priority, but training hard was more and more important given that we are getting closer to racing season.

Running workouts in this phase, all Tempo, over a month and a half more or less:

• 25 km at marathon pace on rolling hills (4’14”/km)

• 2 x 30’ of road climbing (6-10% grade)

• 2 x 40’ of road climbing and descending (4-8%)

• 2 x 20’ of climbing on trails (15-25%)

For both Threshold and Tempo sessions, I used hills a lot, and often went on trails to increase the steepness of climbs, while sporadically pushing hard also on the downhills, to add stress on the muscles and improve their durability.

Gym work:

• Specific Block: strength endurance. Similar structure with 4-6 exercises (Squat(x1-2), Lunge(x1-2), Hinge, Calf) but with lower weight and more reps, e.g., 3-4 sets of 12-15 reps.

Here are all the key workouts I’ve done in these 5 months between races, which is only 14 (8 Threshold and 5 Tempo + 1 race simulation for Passatore). Quality sessions done well and at the right time:

Between races: High Volume Block

I’ve discussed elsewhere how I thrive on volume and keep my taper really short, hence since in this phase I felt structurally solid (i.e. no niggles or other reasons to limit volume), I took the time I needed to feel good again after Chianti, and then pushed volume near my limit before doing another short taper ahead of Tuscany Crossing.

Eventually both races were raced at the same intensity, and took me a very similar time, reflecting what was my capacity those months.

Other training notes

During the training blocks up to Chianti, I’ve added plenty of vert in my weeks, with a reduction in January because the weather was quite poor and I couldn’t get on the trails (which were just rivers of mud), and as such, I trained on the road or indoors that month.

I’ve also had the usual niggles, and managed them with the bike (see below). I was always able to do what I thought mattered the most: gym, hard running workouts and long runs, but I had to often reduce weekly volume by replacing easy runs with easy rides, until after Chianti, when I started to be able to tolerate more load (up to 235 km/week after Tuscany Crossing).

Most importantly, the weekly vert was ‘more focused’ than in the past. I tried to run faster on average, outside of workouts, for specific climbs or descents. The goal was to add intensity just enough to add muscular stress that could lead to positive adaptations without compromising the rest of my training. Thin line, of course, but that’s what pushing our limits is about - wherever they are.

Also note how little hard running was needed to be the best version of myself: maximum one hard run per week. This is not a universal truth, it’s just my capacity. Better athletes can do more, but copying their training won’t do us any good if we cannot assimilate the work. Always ask yourslef, do I have the capacity to do ____ (whatever is hip today)? Or should I just stick to the basics and do less, more consistently, until my capacity has increased? Often, the answer is the latter. For me personally, in terms of intensity, I kept it very simple these past 6 months: I did a block of threshold to raise my capacity for sessions that will be both high volume and high intensity (Tempo work), and then I did those sessions, which resemble a race or are harder than a race (but a bit shorter), eventually getting me ready for the big day. Nothing particularly fancy, but everything carefully planned.

Cross-training: The Bike and Passive Heat

My original plan for these 6 months was not to bike, if anything, just because winter cycling isn’t really my thing. That being said, when my body doesn’t cooperate (e.g., running niggles), I’m quick at this point to switch risky runs (which might turn a niggle into an actual injury) for rides, and as such, I ended up cycling 3-9 hours per week most weeks since January:

All my cycling was low intensity. This was the exact opposite of what I did July-September, when almost all my cycling was high intensity. Main reason: this time I had no neural capacity to go hard on the bike once the gym and hard running were done well, and those were way more important for my goals.

Running was modulated accordingly, with lower volume when I had more niggles (and therefore biked more):

As a result, this year I trained more time (210 hours, run+bike), but ran less than any other year since 2021 (I ran 1200 km up to March 24th, 75-80% of what I had done in the past).

Total load, if that’s a thing:

An important takeaway here is that the quality of what we do matters a lot. It seems obvious, but especially when thinking about trail running and ultra distances, it’s hard not to start thinking that volume is what matters the most. Volume matters, of course, but there’s a point when the quality of our work, either running or in the gym, is worth more than large volumes of running done with less intent (‘pascolate’, a technical term I learned from my colleagues at ).

Once I run about 100-120 km in a week, adding some high-quality running and a good session in the gym is worth more than increasing that running volume by shuffling around many hours. There are more muscle fibers to use and gains to make when we have the capacity to tolerate non-easy running.

In a way, I was glad to be forced to run less (by the various niggles), as it allowed me to test something otherwise I would have unlikely tested: focusing on key sessions, keeping volume moderate (as opposed to very high), and then using the bike more, to avoid physical issues that are so common when running more.

Again, durability is more about quality and the intent of each session than overall running volume.

You can see below that my weekly running volume is quite a bit lower than last year (+19% time), but includes plenty of very long runs with more intentional climbing and descending (elevation gain is indeed much higher despite less running, +38% in meters), sometimes with quality (hard climbs and hard downhills, to work both on fitness and muscular durability). Overall training volume is higher (+28% time) because of cycling. The data below does not include the block between Chianti and Tuscany Crossing.

Passive heat

Big fan of passive heat.

I started years ago in preparation for the Florence furnace (i.e., 3pm race start at the end of May for Passatore), but then it grew on me, and I started doing it basically every day when I’m in Italy. Since then, quite a few studies have shown interesting results on the benefits of passive heat also for racing in temperate environments, hence why not (a reason not to would be that all stress matters, so if adding passive heat makes you sleep poorly, train poorly, recover poorly, etc. - you might have more to lose by adding it, but if it doesn’t, then I think it can be an incredible tool). As someone who has always struggled in the heat, I feel like a bit of chronic exposure can do me good, hence the strategy.

My protocol is 20-22’ after training pretty much daily, maintaining temperature at about 42C.

Adaptation: How am I Responding?

You probably know I find it particularly fascinating to try to capture how we respond to the various things we put ourselves through.

One way to do so, is through the lens of heart rate variability (HRV), see my guide here, and best practices on how to measure it, here (spoiler: not with wearables).

From a nervous system point of view, it’s been a good couple of months, with things rather stable apart from when I had a hip issue and maybe some associated inflammation. Reintroducing intensity in December-January worked well(I maintaned stability), and similarly, peaking with volume as well as introducing long and hard tempo runs (which, together with the previous threshold block, would be the 'hardest to digest’, from a nervous system point of view), led to no suppressions in HRV, which I took as a good sign that I was balancing load well, recovery well, and assimilating the work well.

With the introduction of the bike, maybe at a certain point I overcooked it, as I started routinely training more than 3 hours per day, or at least I reached the limit of my capacity to sustain a certain load (which at that point was over 20 hours / week of training for 5 weeks in a row).

We can see this in a sustained suppression that I could not turn around even with a taper in which I trained 3 hours in 6 days, even though I suspect race nerves got in there as well in the last week, which is very typical of me (the ‘passion for stress’ is rooted in an inability to deal with it well).

Heart rate data shows nothing noteworthy, as it tends to be a less sensitive marker of stress, apart from race morning, when things went bananas:

All things considered, I think I pushed it a bit too much with overall volume when I was also doing long tempo runs, we are talking very high volumes + very long and hard runs. I am a regular person in the great scheme of athletic performance after all, not an outlier.

My thinking was that I was going to force myself to almost stop training for 2 weeks (1 week of taper, race at Chianti, then one week of recovery, so basically 2 very easy weeks with a big effort in the middle), hence I pushed the volume a bit more than I would have otherwise done, given the context.

No issues eventually with the race, where I felt strong and fresh, but maybe this is what we are seeing in the data. Between races we can see a good pattern with stability in HRV data, and a positive trend even when pushing the load quite high (about 200 km for the week):

and then another heart rate peak on race day, I clearly have my issues with race nerves:

Testing: VO2max, Thresholds, Metabolic Flexibility

As you might know, I’ve put together a small private lab which I use for myself and the athletes I coach whenever they can get to Brisighella, which allowed me to test a few things during this whole process.

In particular, I was curious to see the impact of stopping high intensity running on my fresh state physiology (VO2max and Thresholds, VT1/VT2) as well as how those parameters would change when reintroducing high intensity running and during this training build, which was quite different from my previous ones.

VO2max and Thresholds

I first tested in September after Winschoten, hence arguably in my best shape. Then I tested again after my ~5 weeks without high-intensity running. During this phase, I reintroduced passive heat training, which was missing in summer as I don’t have a bathtub in Amsterdam. Hence, this latter factor might have contributed to VO2max staying the same between these two tests. On the other hand, both my thresholds were slightly worse. This is in line with a recent study showing that passive heat improves VO2max but doesn’t improve Threshold.

I then tested again mid-February, when I had already done all my threshold training and some tempo runs, which led to (maybe surprisingly) some of my best testing: VO2max was again exactly the same, but vVO2max (the velocity reached) was 1km/h faster, and both my thresholds were higher (in terms of velocity). Hence, despite this type of fresh state profiling capturing nothing of durability (which was the focus of my training), I had a better profile than back in September:

It could be that the strength work enabled greater capacity for uphill climbing and, as such, a greater training stimulus on my hard sessions, which led to greater improvements also in my fresh state profile. Speculations.

Metabolic Flexibility

My little project of last year.

I think at this point I’ve understood quite well how it all works, which took me a while, and lots of testing (fresh state, before / after long runs, while changing macros in my daily diet for several days or weeks, etc.). Some learning here, and a podcast with where I think I finally articulated it decently, here.

During these 6 months my diet has been quite heavily periodized (the way I discussed here), which means typically I do 2-3 days per week in which I stay quite low carb (e.g. omelette for breakfast, greek yogurt, nuts, peanut butter and berries as snacks, a salad with avocado and some protein for lunch, more vegetables and lean meat or fish for dinner), 2 days in which I stay high(er) in carbs (i.e. the same as the low carb day but replacing the low carb breakfast with oats and fruit, then during the day more fruit and possibly pasta either for dinner or lunch), and 2-3 days in which I stay moderate (which means low but less anal about it). The periodization is driven by training of course, and since I tend to do 1 hard run per week, and maybe a longer run with some quality, I end up doing 2 higher carb days to support those, and stay lower for the rest of the week (and easy training). This mix works wonderfully for me, and I also enjoy it very much at this point.

Before Chianti, I did a short check starting from an easy jog to the highest intensities I would think to touch on a race like this (that is, just around my first threshold, as going harder would not be sustainable for the time it takes me to run that distance. Plus, my first threshold is pretty high at this point and I normally train way below it, in Zone 1).

The data showed what I had seen a few times in the previous months, i.e., I’m burning 1 gram of fat per minute, with no degradation with increased intensity (a steep drop would then happen if I had kept increasing intensity up to the second Threshold):

The same can be seen also below:

The last image shows how the higher the fat oxidation, the lower carbohydrate burning becomes, but at the same time, fat oxidation is limited in its maximum value and can only help so much when you increase intensity / power output.

What does this mean in practical terms for the fueling of athletes of different levels?

This is a key consideration to understand. It explains why an elite athlete can still need plenty of carbohydrates per hour despite having a very high fat oxidation rate (their fat oxidation in absolute terms is not unlike mine, but their power output is much higher, meaning they burn more carbohydrates at what is still a ‘low’ or ultramarathon intensity for them, despite high efficiency).

Somewhat paradoxically, a less good athlete would benefit even more from having high fat oxidation rates because they could sustain their power output indefinitely at their race intensity, and could therefore have more freedom (than the elite) in terms of how to deal with their race nutrition.

I repeated these tests also before Tuscany Crossing, with very similar results given that my diet and training hadn’t changed:

Some data from the last test:

• Body mass = 73 kg. Estimated usable glycogen at full capacity: ~550 g

• CHO oxidation at race pace (average between the last two stages shown in the image): 1.7 g/min, given about 1 gram/minute of fat oxidation at race pace.

• Total carbohydrate need = 1.7 g/min x 300 min (5 hours race) = ~510 grams

• Exogenous carbohydrate need given stored glycogen and oxidative capacity: 510 grams (CHO oxidation rate at race pace) - 550 grams (stored in the body) = -40 grams, i.e., 0 grams needed of exogenous carbohydrates. Obviously, I’m messing with you, but see the nutrition section below to see how all of this plays a role.

Durability: Is it Getting Somewhere?

Million-dollar question.

Despite the positive results in my lab testing, it remained a challenge to understand if the whole point of these 6 months (i.e., improving muscular durability) was getting somewhere or not, before the first race.

Durability is something we work on, but we put it to the test only on race day, when the system is stressed to its limit. It is not something we can easily see improving day by day or in our hard workouts or even in our longer runs, at least not in the same way we can see progress during, e.g., a marathon build or when progressing from a state of limited or moderate training to high volume training.

Personally, I found this quite challenging mentally. I mean, I was totally fine, but still, I was wondering, is this helping me make another breakthrough, or should I do something else? At the end of the day, I want to get to Faenza before midnight, IYKYK.

Thinking about these months, here are some cues that I noticed during this process, and that gave me confidence that something positive was happening:

• After 5 weeks of no intensity, I ran all out for 5 km uphill, at what we could consider Threshold or a bit above Threshold intensity. I ran a personal best on that occasion for a segment that I had raced many times before, feeling like, for once, the limit was my cardiorespiratory system and not my legs’ inability to generate the required power against gravity. It was a good first sign that gym work was transferring to running (a strong point against the ones arguing for ‘only running’, something you wouldn’t argue in any other sport).

• On the bike, I could ride out of the saddle on climbs for ‘a long time’, while normally it’d be a 30-second activity for me before my legs would start burning. I found this interesting as ‘the work was transferring’ even if not running related.

• On my runs, I could run faster on the downhills, without even trying. It was like what was a normal subjective feeling of going downhill easy while ‘saving the legs’, now was reset to a different level, which meant running quicker. The difference wasn’t enormous, but enough to notice.

• On my runs, I could climb better when the grade went higher than 10%. I couldn’t really notice much difference on 6-8% grades - where I was already efficient and running well earlier - but 2 months in, I started seeing I could run 20% grades and stay in zone 1, which I was never able to do before.

As you can see, none of these are direct assessments of durability, and I did not really see particular differences during my long and easy runs, but there were a number of indications that made me think maybe I was getting somewhere with the plan I put together for this year.

The final answer was going to be at Chianti, the first race of the season.

The Races: Planning and Execution

Tapering

For both events I kept going with my training until the Sunday prior to the race, and then tapered heavily for a week, which meant one day off, and a few days of pretending to run (Thursday to Saturday on the race course with friends).

Given the whole focus of these 6 months, I decided also not to do any hard running but to do one strength session instead, as a final ‘quality stimulus’ before the race. Race week before Chianti:

Race week before Tuscany Crossing:

Pacing and race intensity

For both events, I planned to race just below my first threshold for most of the time, getting around it or slightly above on the climbs, but then making sure I’d get my heart rate a bit lower on the downhills or flatter parts. All of this was driven by perceived effort of course, but with a heart rate cap as a limiter to make sure I wasn’t overcooking myself too early. I was expecting a 5-hour effort for both races (given that one is longer but has less vert), or maybe 4h 45’ on a very good day, hence the intensity needed to be managed accordingly.

While grade-adjusted pace as its limits (type of terrain of course matters and can’t be accounted for, i.e., how technical it is), we can see the pacing was as good as it gets for Chianti:

My heart rate at Chianti was very stable. I expected more drift, but we were extremely lucky with the weather, which was a bit chilly, but not humid, and even when the sun came out, there was the right amount of breeze not to overheat. Again, just very lucky on that front.

Eventually, my average heart rate was 85% of my recent maximal heart rate, hence I think I’ve managed it very well this time, combining perceived effort and a heart rate cap (and of course, dropping the damn ego since the start) to stay around my first threshold, which after many years of good training, has reached a relatively high intensity.

Somewhat similarly for Tuscany Crossing:

which was, however, a month later and quite a bit warmer in the second half of the race, leading to higher cardiac drift (and other issues I’ll discuss in the nutrition session below):

Environmental conditions aside, I’ve felt pretty much the same for both races, and kept the intensity where I thought I could maintain it for the required time. I’ve also put together the plots below to show heart rate distributions across race quarters, which I think is quite informative to show us potential issues with pacing.

Here is Chianti:

For Tuscany Crossing, we have a similar situation, but with a higher heart rate in the last quarter, due to higher temperatures. I was able to maintain the same power output but definitely struggled more and felt that I had to dig quite a bit more, mentally, in order not to fade in the last part of the race. The higher internal load was definitely perceived as such:

For context, before is a race that I paced poorly, which shows the typical profile you’ll find in most athletes that start too enthusiastic, so to speak:

Fueling

As discussed earlier, race fueling should depend on 1) your power output (the higher your capacity to produce a certain output, the more carbohydrate you will burn, regardless of your fat oxidation, which can only account for so much). This is why elite athletes’ intake should be higher than my own, 2) your metabolic flexibility in relation to the duration of the event. The more fat you can burn, the less glycogen you burn, and therefore the less likely you are to hit the wall (that’s right, you don’t hit the wall because of not eating enough carbohydrates, but you do because your body can’t function without burning carbohydrates due to poor metabolic flexibility and poor fat oxidation rates, resulting in quick depletion of glycogen - since no amount of carbohydrate intake can spare muscle glycogen - and only liver glycogen is impacted by carbohydrate intake during exercise, see ’s excellent video on the topic). Now, I added “in relation to the duration of the event” because it’s overly simplistic to ignore the fact that for 3-5 hours, we can certainly do very well with limited carbohydrate intake, but if you start skipping meals because you are running the whole day, then a higher hourly intake is important regardless of the macros. At that point it’s just calories, and we need the calories for such a long event, we cannot only focus on the next point, which is 3) maintaining blood glucose levels, or yet another imporant reason why we need a certain (very low) level of carbohydrate intake even when we are very metabolically flexible and when the race is short enought for glycogen stores to last. Finally, we should remember that 4) the brain likes carbs, and even mouth rinsing improves performance, hence frequent sips might be a good idea even if the overall rate is not particularly high.

Put all of that together, and for Chianti, I know that given my race intensity (85% of my maximal heart rate, and a bit over 12 km/h grade-adjusted speed), and my fat oxidation rates at those intensities (i.e. 1 gram/minute), plus full glycogen stores given the limited training of that week plus 2-3 days of high-carb intake prior to the race, I can last about 5-6 hours without any intake at all (what? yes, true story).

However, this doesn’t account for maintaining blood glucose stability (which requires 10-15 g/carbohydrate per hour), and brain-mediated performance effects of carbohydrate intake (again, see research on carbohydrate mouth rinsing), hence in training, or previous races of similar duration, including some of my best events since becoming metabolically flexible, I gravitated around ~30 grams/hour, which allowed me to perform at my best without gastrointenstinal problems (or with limited but somewhat manageable gastrointenstinal problems).

At Chianti I decided to experiment more as I’m still experiencing issues with gels, and as such, I’ve been drinking carbohydrate solutions to see if I could do better in terms of gastrointestinal discomfort (the reasoning is rather simple, if I sip every few minutes, my gastrointestinal system can deal with a little food each time, as opposed to ‘a lot’ of food in a single go). Eventually, I had about 80g + 80g in 2 flasks, plus 1 gel with caffeine (100mg) and 30g of carbohydrate (+200 mg sodium). Half an hour to the finish, as fatigue was more present, and for once I had no nausea or any other gastrointestinal issues (which might mean drinking carbs is working better for me), I decided to take another gel, hence I had about 220g CHO total in 4 hours and a half, or about 50 g/hour, which sits more or less where recommendations were before most people went nuts (e.g., providing intake rates without accounting for power output, which seems the unfortunate norm).

At Tuscany Crossing, I opted for the exact same strategy I had used at Chianti, but eventually had ~40 g of CHO/hour: 2 flasks (90 g each), and one gel (30 g + 100 mg caffeine). I had planned slightly more, but had to adjust on the go as the warmer day made it a bit more challenging; I got the usual nausea, and eventually decided to skip a few gels. Knowing that your metabolism can handle a lot of different situations - not only chronic high-carb intake - helps a lot when you need to make adjustments because of unforeseen issues. Less fuel, but no drop in pace or other issues, because of metabolic flexibility.

To recap, given my power output and metabolism (fat oxidation rates), I can run 4-5 hours at moderate intensity (as in a race like Chianti or Tuscany Crossing) without running out of glycogen, hence the intake is there to maintain blood sugar, delay hunger, and for other neural benefits. More carbohydrate at that point means an increased likelihood of more gastrointestinal distress (no, vomiting during a race is not normal), without any performance gains. Running faster might benefit from higher intake, but I am not capable of running faster for that long, regardless of what I eat (obviously!).

That’s a Wrap

These two races were a lot of fun.

Never in an ultramarathon I felt like my muscles could sustain what the rest of my body is capable of, allowing me to race well, while being considerate, but still being able to push, overtake, and just have some fun instead of simply having to manage cramps and muscular issues for half of the race until the finish line, for what is typically just a very long mental exercise for me.

Needless to say, I like it more this way. Everything needs to be managed, and it’s always a thin line when it comes to an ultramarathon raced at the limits of our abilities, but the muscular durability I was trying to improve has indeed improved and allowed me to live one of my best running days. Two races out of two done well seems a good signal.

In the past, I used a more standard and generic build towards ultramarathons, in which the focus was on getting as fit as possible, and then trying to extend that fitness (e.g., prepare for a half-marathon, then a marathon, then an ultra, during a 6-month build). While the approach makes sense, and was what I needed while still developing my capacity over those distances, this was no longer what I needed now to make another step forward. In particular, in the past two years, I went quite deep into my limiters to find more room for improvement, first with metabolic flexibility (and dietary changes) and now with muscular durability (by combining gym sessions and hard workouts on steep hills).

Eventually, I ran very similar races. The level in the two events is quite different, some of the best runners in the world were at Chianti, while Tuscany Crossing is a smaller and lower profile event. Obviously, at my level, I can only race for myself, and the way I felt during the race (i.e., strong) is what I was really chasing with this training block. This being said, the competitive aspect does remain an important part of the enjoyment of racing for me, and I look forward to a few more of these days.

• Chianti Marathon Trail, 46 km +1900m, 95th / 1400, 9th age group.

• Tuscany Crossing, 52 km +1500m, 16th / 500, 3rd age group.

Thank you for reading, and happy training!

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

I hope all is well.

Here’s a new recap of articles, research, and updates from the past month. I hope you’ll find the articles useful, and I would like to take the opportunity to thank you for your support.

Please feel free to comment below or in the articles should you have any questions, and I will follow up soon.

Take care!

Heart Rate Variability (HRV) 🫀

Signal vs Noise

More than fifteen years ago, I worked on some of the first systems that could stream ECG data from the body to a smartphone (see here). At the time, I also wore one of those systems myself for more than a year. That experience shaped much of what I later focused on in my work, but it also shaped what I decided not to pursue.

When optical sensing came out, it became clear very quickly that without an ECG, most physiological data collected continuously in daily life is of limited value for HRV analysis. Outside of very specific moments, such as a quiet measurement in the morning or sometimes during sleep, the signal is simply too unstable and noisy to support meaningful interpretation. Movement, posture, muscle activity (even tiny contractions like the ones I’m generating in this very moment as I type) are sufficient to make PPG data collected at the wrist or finger completely useless, and typing or scrolling a phone is clearly no exercise.

This month, we look at a preprint validating HRV data collected in ambulatory settings using wearables (full text here). [Spoiler: it’s not any good, but you knew this already].

Full blog here.

The Heart Rate Variability Book will be out in May

We finally have a publication date for our book, Heart Rate Variability: Science and Strategies for Peak Performance.

Here is a bit of the backstory on how this unfolded over the past seven years ☺️

Back in 2019, Dan and I started discussing the idea of writing a book on HRV. At the time, there was nothing on the market that offered a truly modern and practical perspective. We began writing and drafted several chapters before and during the pandemic, but eventually lost momentum.

In October 2022, I met Sylvain at the Italian Olympic Committee, where we delivered a workshop on HRV together. We put together a decent show, and afterward, he asked if I would be interested in writing a book. I discussed it with Dan, and we decided to join forces. We also involved Emma, who had previously worked with Sylvain and brought invaluable expertise, especially in HRV and cognitive performance. Cooperation over competition.

This time, Sylvain’s incredible energy set the pace. He helped us find an editor we immediately connected with: Human Kinetics. We signed our agreement in 2023, and from that point on, we spent more than two years writing, revising, and iterating with our editors.

All of us contributed to every chapter. We learned from each other and gradually shaped what we can now confidently call our HRV book, reflecting decades of combined experience in the field. It has been a long journey, and I am truly grateful to my coauthors for making the process genuinely enjoyable and intellectually stimulating.

Our goal has always been to stay between hype and skepticism, so that people can truly benefit from this tool by understanding both the potential and limitations of HRV analysis.

I hope you will enjoy the book.

Publisher website here, or pre-order on Amazon, here.

How To Get Started with Biofeedback

I received a question from someone approaching breathing exercises and wondering how to practically get started, and as such, this blog.

In particular, I will refer to the work of Paul Lehrer and colleagues, who developed a structured protocol, grounded in physiology and refined over years of clinical and experimental work. The goal is to use respiration in a way that can directly impact heart rate (variability) and blood pressure (or improve autonomic regulation), which can benefit us from a physiological and psychological point of view, as documented in the scientific literature.

CoachCorner ⏱️

HRV-Guided Training: an example

An example of HRV-guided training in real life, from two of my athletes. We can see how some stressors are planned or expected, e.g., traveling, and typically our periodization accounts for such stressors when planning weekly sessions (i.e. we had already scheduled lower intensity training).

Other stressors are however unexpected, and might require adjusting when looking at the physiological response, as we can see in the other annotated cases.

The essence of it is quite simple: start with a plan, and adjust based on the individual response to training and other stressors.

• Get HRV4Training.

• Learn more about HRV-guided training.

• and apply to get coached.

Improving Ultra Marathon Performance through Metabolic Flexibility with Marco Altini

In the words of , “While everybody has spent 2025 increasing their carbohydrate intake, Marco spent his year improving his metabolic flexibility and fat oxidation through periodised nutrition. Although he agrees that carbohydrate availability is essential for performance and recovery, his approach challenges the assumption that simply increasing carbohydrate intake is always the primary lever for improvement.

In this episode, I speak to Marco Altini about what metabolic flexibility actually is, what it is not, his metabolic flexibility and fat oxidation journey, along with practical tips and suggestions on how athletes should approach the topic of metabolic flexibility.

Marco has a wealth of knowledge and I am very happy that he has decided to share some of it with my audience.“

Born on the Trail

Improving Ultra Marathon Performance through Metabolic Flexibility with Marco Altini

While everybody has spent 2025 increasing their carbohydrate intake, Marco spent his year improving his metabolic flexibility and fat oxidation through periodised nutrition. Although he agrees that carbohydrate availability is essential for performance and recovery, his approach challenges the assumption that simply…

Listen now

6 months ago · 5 likes · Niki Micallef

Personally, I’ve enjoyed this chat, and I think I was able to elaborate on the topic a bit better than other times in the past. I hope you will find it useful.

CH4 Trail

[Italian]

Stiamo organizzando una gara di trail per il weekend del 14 giugno qui in zona. Per chi fosse interessata/o, informazioni aggiuntive seguiranno su Instagram.

Grazie mille!

Interviews and Podcasts🎙️

The Australian Financial Review

A few words from me and Dr. Daniel Plews on The Australian Financial Review. Thank you Euan for the interview, much appreciated.

Full article if you are a subscriber, here.

Recent Updates:

• January 2026

• December 2025

• November 2025

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

In particular, I will refer to the work of Paul Lehrer and colleagues, who developed a structured protocol, grounded in physiology and refined over years of clinical and experimental work. The goal is to use respiration in a way that can directly impact heart rate (variability) and blood pressure (or improve autonomic regulation), which can benefit us from a physiological and psychological point of view, as documented in the scientific literature.

Learning the breathing pattern first

The protocol begins by familiarizing the person with diaphragmatic (abdominal) breathing. Early sessions focus on learning to breathe in a smooth, continuous way, with minimal chest movement and no breath holds. At this stage, the emphasis is on comfort and awareness rather than on changing HRV or following any specific pattern.

Biofeedback is already introduced here, not to “optimize” anything, but to help the person see how their heart rate naturally responds to breathing. Simply observing the heart rate speeding up during inhalation and slowing during exhalation is an important first step, as it links an abstract concept to a concrete physiological signal. As you know, I’ve developed the HRV4Biofeedback app, which can be used for this purpose.

Finding the individual resonance frequency

Once the breathing pattern is stable, the protocol moves to identifying the individual’s resonance frequency, the breathing rate at which heart rate oscillations become largest and most regular. This is a key element of Lehrer’s work and one that is often skipped in simplified approaches.

The person breathes at a series of slow, paced rates, typically between about 4.5 and 6.5 breaths per minute, for short periods at each rate. The rate that produces the largest, cleanest oscillations in heart rate is selected as the training frequency. This rate differs slightly between individuals.

Importantly, this step makes clear that HRV biofeedback is not about breathing as slowly as possible, but about breathing at the rate that best matches one’s physiology.

(Supervised) Practice

After lerning diaphragmatic breathing and finding your resonance frequency, the original protocol includes supervised practice, in which several sessions are dedicated to practicing breathing at this rate with continuous feedback. The person follows a pacing signal while observing their heart rate waveform in real time. Instruction focuses on keeping the breath smooth, avoiding exaggerated inhalations, and maintaining a slightly longer exhalation without strain. You can for example set the inhale to exhale ratio to 40:60 in the HRV4Biofeedback app, in order to be guided with this protocol.

At this stage, biofeedback serves a corrective role. If the person over-breathes or becomes tense, the heart rate pattern immediately reflects that. Over time, people learn how the breathing should feel when it is effective, rather than forcing the pattern mechanically.

Daily home practice: the core of the protocol

A central component of the Lehrer protocol is regular home practice, and this is where most of the training effect is expected to occur, and also what is applicable the most to the case of someone approaching biofeedback without taking a course or participating in a study. The classic recommendation is two sessions per day, each lasting about 20 minutes, practiced consistently over several weeks. This dose reflects what was used in early clinical applications and research studies, even though shorter protocols have also been used by other groups recently, down to a few minutes per day. Still, if you’d like to give the original protocol a go, two twenty-minute sessions per day are what is recommended.

During home practice, pacing and feedback are still used whenever possible, so that the person continues to train at their individual resonant frequency. HRV biofeedback is not a quick intervention. In the original work, training typically unfolds over six to eight weeks or longer. Improvements are evaluated by broader changes in physiological and psychological regulation.

As described by Lehrer, HRV biofeedback is a form of physiological skill training. It is not about chasing high HRV numbers, inducing relaxation, or using breathing as a substitute for medical care. It is a structured way that could train autonomic regulation using controlled breathing and feedback, which can benefit us from a physiological and psychological point of view, as documented in the scientific literature.

Modern tools, such as dedicated HRV biofeedback apps, make it easier to apply this protocol outside the lab. However, the underlying logic remains the same as in the original work: learn the breathing pattern, find the right rate, and practice regularly.

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

Despite metabolic flexibility being well supported physiologically and in the scientific literature, the practical implementation of these methods is often misunderstood and reduced to fasted training or chronic carbohydrate restriction, neither of which I would recommend to endurance athletes.

In this blog, I’ll cover the case study of Matteo, an ultratrail runner I’ve been working with, whose goals and training profile made metabolic flexibility a relevant area to attempt to improve, and with whom I have worked towards that goal, using a more nuanced and sustainable approach.

Let’s get to it.

Baseline Testing: Identifying Limiters

One key point before embarking on this journey is that rather than assuming metabolic flexibility is a limitation, we need to test it objectively. Even if we work towards a sustainable approach (details later), we do need to make some changes that are somewhat in conflict with our current habits and cultural context (as these have likely led to our poor metabolic flexibility!), and as such, it wouldn’t make much sense to do so if we weren’t lacking metabolic flexibility (unless we had other reasons to go there, e.g. health related, but I’ll spare you the lecture on glycemic control for today).

Matteo lives nearby, and as such, I could test him at my lab. As I typically do, he underwent laboratory testing, including a VO2max test and a submaximal test to assess substrate utilization, running economy, ventilatory thresholds, and heart-rate responses at various intensities. What we care about here is the submaximal test mostly, as in this test, we use long steps to reach steady state and can assess the metabolic response more accurately than we would be able to do during a shorter maximal test.

The results showed a clear pattern: fat oxidation rates were very low, even at slow speeds, around ~0.3 g/min, with a strong reliance on carbohydrates across all intensities and no crossover point (i.e., at no intensity, he was burning more fat than carbohydrates):

For an ultratrail runner, this profile can be problematic. Heavy reliance on carbohydrates at low and moderate intensities (i.e., poor metabolic flexibility) increases the likelihood of early glycogen depletion, compromising performance. Finally, poor metabolic flexibility complicates fueling strategies during long events, as it limits our options (more considerations on this, here).

Below, I plotted the breakdown of total energy in terms of fat and carbohydrates, and we can also see here that he’s running on carbs only:

The lab data provided a clear rationale to intervene, while also giving us a baseline to evaluate whether any changes would be meaningful.

Matteo had already finished various ultramarathons, but struggled a lot in the past year to finish some of his key races, or to simply feel good during them, often in relation to his nutrition.

Together with my work on HRV, the recent work I’ve done on metabolic flexibility was one of the reasons he approached me, meaning that he was already highly motivated in attempting a different avenue, which was quite different from what he had tried in the past (i.e., maximizing carbohydrate intake).

This is why we decided to give it a go.

Dietary Changes: Shifting Emphasis Without Restricting Energy

As we started the nutritional intervention, we were intentionally conservative, but at the same time, we wanted to give ourselves a good chance to succeed. In practice, this means that there was no caloric restriction, no prolonged fasted training, and no blanket removal of carbohydrates. Instead, we selectively shifted macronutrient emphasis based on session demands.

Easy runs were often preceded by protein- and fat-based meals (eggs, avocados, Greek yogurt, nuts, nut butters), while harder workouts and long runs were appropriately fueled with carbohydrates before and during the run. This approach mirrors my periodized nutrition framework, where fuel availability is matched to the expected demands of the session rather than applied uniformly. This is how we can improve fat metabolism without compromising our top-end capacity, or in other words, how we become metabolically flexible, which is key to endurance performance success.

I have asked Matteo to log everything he ate for about 10 days before we started manipulating his diet, and then again three months later for another 10 days. While these are just approximate breakdowns (we logged meals without weighing anything or counting calories), we can see the (again, approximate) shift in macronutrients from his previous diet to his current diet:

As we reduce carbohydrate intake on certain occasions, it is rather normal that both protein and fats go up a bit to account for the missing calories.

I will add that when testing again, over roughly three months, Matteo gained about one kilogram. This is worth stressing, because improving fat oxidation is often mistakenly associated with weight loss or energy deficit. In practice, metabolic flexibility can improve substantially even with adequate energy intake or slight weight gain, provided carbohydrate intake is reduced selectively, not chronically.

Above, we can see a qualitative shift from a carbohydrate-dominant diet toward higher protein and fat contributions, while carbohydrates remain present. This reflects a change in dietary signaling rather than restriction.

Training progression: Volume, Intensity, and Progress

Importantly, the work on metabolic flexibility did not slow down training progress as we were shifting from a Base training block to a Support training block for Matteo’s spring ultramarathon (Chianti 73k).

Over this same period, training volume increased gradually and structure improved, with the inclusion of more threshold work and longer runs that over time started to get a bit more specific. This shows how metabolic work did not require reducing load or did not impact negatively Matteo’s ability to tolerate load. We can also see that his key sessions have improved in quality, with faster and longer threshold workouts, for example:

Follow-up Testing: Metabolic and Cardiovascular Adaptations

When Matteo returned to the lab after about 10 weeks, the changes were clear. Fat oxidation rates were substantially higher across all submaximal speeds, with meaningful contributions even at moderate intensities (which I consider what matters the most).

Additionally, heart rate was lower at the same speeds, suggesting that improvements in metabolic flexibility occurred alongside improvements in cardiorespiratory fitness:

Most importantly, Matteo has been feeling great during training, and was able to run long, multiple times per week, and all while being very flexible on fueling (i.e., feeling very good for several hours of running even with limited carbohydrate intake, which is a valuable skill to have in long races where gastrointestinal issues tend to happen).

Key Takeaways

Matteo’s case shows that metabolic flexibility and performance are not competing goals; quite the contrary. When nutritional changes are selective, sustainable, and aligned with training demands, it is possible to improve substrate utilization while maintaining or improving training quality, workload, and fitness.

Once again, metabolic flexibility is not about fasted training, weight loss, or carbohydrate avoidance. It is about improving the athlete’s ability to match fuel use to intensity, which is particularly relevant for long-duration endurance events. On the flipside, you can have poor fat oxidation rates even when training fasted on a daily basis (maybe for practical reasons), if, for example, your diet (and genetics) do not support metabolic flexibility adequately.

Needless to say, this is a single-athlete case study and should be interpreted accordingly. Individual responses to dietary and training interventions vary based on genetics, training history, and baseline metabolic profile, even though we could fairly say that for most people, and contrary to popular belief, 75% of metabolic flexibility comes down to diet, and 25% to training (as Bob Seebohar would say). Some of the figures presented (e.g., dietary macronutrient distribution and training progression plots) are illustrative summaries derived from logs rather than precise nutrient quantification. In case you’d like to experiment with nutrition, work with a professional, ideally one who has the required knowledge and tools to truly assess your metabolism and any changes over time.

Finally, this case study aligns with broader observations discussed previously on periodized nutrition and endurance performance. It provides a concrete example of how metabolic flexibility can be addressed pragmatically, without compromising training or long-term sustainability and as such, I hope you have found it useful.

Happy training!

FAQ – Common questions on Metabolic Flexibility in Practice

Do I need to train fasted to improve fat oxidation?
No. In this case, improvements in fat oxidation occurred without systematic fasted training. Selectively reducing carbohydrate intake before (and after) easy sessions is sufficient, while harder workouts and long runs are properly fueled. Fasted training is one possible tool, but it is neither required nor universally appropriate, and the rest of our daily diet will have a greater impact typically.

Does working on metabolic flexibility require a caloric deficit?
No. This intervention did not involve intentional energy restriction, and body weight increased slightly over the intervention period. Improving fat oxidation is not about eating less, but about changing substrate availability and signaling in specific contexts.

Were carbohydrates eliminated or minimized?
No. Carbohydrates remained a regular part of the diet, particularly before long runs and high-intensity sessions. The change was in when carbohydrates were emphasized, not in their complete removal. Obviously, on some occasions, we have fewer carbs (yes, you will survive).

Can this approach harm high-intensity performance?
No. When applied appropriately, there is no reason it should. In this case, training quality, volume, and intensity were maintained or improved alongside metabolic changes. The key is to ensure adequate carbohydrate availability for demanding sessions. I love my carbs and would never go keto, but I do enjoy periodizing my nutrition more than my previous diet. It does take a few weeks to get to this level of enjoyment, and just like any other change, the first weeks won’t be easy.

Is this approach suitable for all endurance athletes?
Not necessarily. Individual metabolic profiles, training background, and event demands matter. This approach is particularly relevant for long-duration endurance athletes, but it should always be guided by context and objective data.

How long does it take to see meaningful changes?
In this case, substantial improvements were observed over roughly ten weeks. The timeline will vary depending on baseline metabolic flexibility, training consistency, and dietary habits, but typically, 6-8 weeks should be sufficient.

Is fat oxidation a performance metric to maximize?
No. Improving metabolic flexibility and in particular, fat oxidation at low and moderate intensities is a means to support performance and sustainability, not an outcome to chase in isolation. The goal is to improve fuel use in a way that complements training and racing demands. There’s no prize for the highest fat oxidation rate, just like there’s no prize for the highest carbohydrate intake during a race (and most importantly, the two can go really well together if you have no gastrointestinal distress!).

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

The Book is Coming Soon

Keep an eye on my Substack for updates.

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training

I hope all is well.

Here’s a new recap of articles, research, and updates from the past month. I hope you’ll find the articles useful, and I would like to take the opportunity to thank you for your support.

Please feel free to comment below or in the articles should you have any questions, and I will follow up soon.

Take care!

Heart Rate Variability (HRV) 🫀

(Another) Independent Validation of our Phone Camera-Based Heart Rate Variability (HRV) Measurements

Heart rate variability (HRV) is only as useful as the quality of the data it is based on. While these days HRV seems to be everywhere, only very few devices have been validated and have shown acceptable accuracy with respect to electrocardiography.

In the early days of my work on phone camera-based HRV analysis, I met plenty of skepticism, and rightfully so, since I was introducing a new method that was quite far from what was current practice (this was before wearables, more than a decade ago, and the alternative back then was only the chest strap or ad hoc sensors with sticky gel electrodes).

As HRV has moved from the lab to chest straps to phone-based solutions, access has improved dramatically, and proving we could do what we were claiming we could do was always my priority, as shown in early blogs and our own scientific validation, eventually leading to the success of HRV4Training.

Yet, independent validation is always another key step. While internal testing and engineering benchmarks are necessary, they are not sufficient. What ultimately matters is whether a method holds up when evaluated by external researchers, using independent protocols and reference standards, and using the tools I’ve built just like any user would.

Recently, this question was addressed again in a paper published in Frontiers in Physiology, where phone camera-based HRV measurements were compared directly against ECG.

The Science of Maximal Athletic Development

It was a great pleasure to sit down with Alan Couzens and Iñaki de la Parra and talk training after learning so much from them over the years.

I really enjoyed the conversation, where we covered both HRV and metabolic flexibility, thank you for having me!

Link below.

The Science of Maximal Athletic Development

MSMRL10: HRV, Readiness & Metabolic Realities w/Marco Altini

Heart Rate Variability has gone mainstream: Rings, watches, readiness scores, green lights, and red lights everywhere…

Listen now

7 months ago · 31 likes · 1 comment · Alan Couzens and Iñaki de la Parra

HRV in 2025

Below is a bit of context around the past 12 months of normalized Heart Rate Variability (HRV) data, showing responses in relation to dietary changes, travel, sickness, heat, overload, recovery from races, and more.

Let’s see what this year will bring.

• Try HRV4Training Pro here.

• More about Normalized HRV, here.

• Long term trends in HRV, here.

Building 🛠️

Biofeedback data in HRV4Training Pro

Last year, we moved our HRV4Biofeedback app to the same backend as HRV4Training, in order to allow you to see your data all in one place (which is HRV4Training Pro). I have resolved a few issues with the integration in Pro, which should now allow you to visualize time spent practicing slow breathing exercises with the HRV4Biofeedback app, in relation to your morning resting physiology collected with HRV4Training.

In particular, you can now pick “Biofeedback Time” as a Tag in the Overview page of HRV4Training Pro, as shown in the example below.

To recap, here is how I use the apps:

• HRV4Training: morning HRV measurement, while seated (see reasons behind this protocol here) + analysis with respect to my normal range to assess my response to acute and chronic stressors (see article above about how to use this data in the context of managing training load).

• HRV4Biofeedback: slow breathing exercises, which could possibly impact recovery, e.g. by quickening a return to normal in terms of nervous system activity after high-intensity exercise, apart from the various benefits that have been shown in the scientific literature (e.g. reduced anxiety, better sleep, etc.). While I measure with the iPhone camera in HRV4Training for my morning assessment (see data and validation here), I prefer to use a Polar H10 for longer, biofeedback sessions.

Please see my other articles below if you are new to biofeedback and interested in learning more:

• Heart Rate Variability (HRV) Biofeedback and Athletic Performance

• Slow-paced breathing after high-intensity exercise

• Case study: morning vs night HRV and using low-intensity exercise and slow breathing to (possibly!) aid recovery

I hope you’ll find the update in HRV4Training Pro useful. Thank you again for supporting my work.

CoachCorner ⏱️

In the past 10 weeks, I’ve been working with an athlete on his metabolic flexibility, using a periodized nutrition approach. We went this way given his goals (ultratrail), and metabolic profile (poor fat oxidation even at low intensities, no crossover point). Then last week we tested again, and the data could not look any better: large improvements in fat oxidation, with no decrement even at moderate intensities.

In this phase, we made sure to focus on shifting macronutrients, while being extra careful in fueling for the work required: no caloric deficit, and a higher carbohydrate intake before long runs and workouts. Given that we are building towards spring races, we also wanted to make sure the work on metabolic flexibility wouldn’t get in the way of our training, which aimed at slightly increasing volume and adding good structure to improve cardiorespiratory fitness.

Eventually, we gained a kilogram during the ten weeks. I want to stress this because it is an often misunderstood aspect: working on fat oxidation is not about creating an energy deficit, and is not even about training fasted (!). We can simply shift macros on some occasions, e.g., having an omelette or some Greek yogurt with nuts and peanut butter before an easy run. You can have lots of calories and still greatly improve fat oxidation rates if you limit carbohydrate intake on certain occasions, and this is the type of work we did. Once again: no need for energy deficits or training fasted, and no need to be too strict either. We want to build habits that are sustainable for the athlete long term and can benefit their health and performance.

The data shows that not only have fat oxidation rates greatly improved, eventually leading to glycogen sparing and limiting depletion during a long endurance event, but we were able to do so while also improving fitness, as highlighted by a lower heart rate at different speeds.

So far so good, I’d say. More on this in an upcoming blog post.

Interviews and Podcasts🎙️

Inside Sports Nutrition: Dr. Marco Altini’s Metabolic Efficiency Experiment - Ep. #215

Last year, when I got more interested in periodized nutrition in the context of substrate utilization and endurance performance, I took Bob Seebohar’s nutrition course on UESCA.

At that time, I didn’t know Bob’s work, and I was expecting the usual “eat lots of carbs all the time” advice that these days seems the default regardless of individual characteristics and circumstances. Instead, Bob’s course went deep into metabolic flexibility and periodized nutrition, in ways that resembled a lot the work I had done in the previous months on my own diet.

For me, it was a great ‘discovery’, and later I’ve also enjoyed reading Bob’s books (I recommend in particular “Metabolic Efficiency Training”, which shows examples similar to my own, in which metabolic flexibility and substrate utilization at different intensities are greatly improved by a periodized diet, as opposed to a diet that is either high carb or low carb).

Needless to say, it was a real pleasure then to finally meet Bob and discuss my own case study with him and Dina for their show. Thank you again for the invite and for sharing your knowledge.

Link here.

Recent Updates:

• December 2025

• November 2025

• October 2025

How to Show Your Support

No paywalls here. All my content is and will remain free.

As a HRV4Training user, the best way to help is to sign up for HRV4Training Pro.

Thank you for supporting my work.

Personal Coaching for Runners

Please apply here should you be interested in working with me.

You can also learn more about my coaching, here.

Thank you!

Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching. He is a certified ultrarunning coach.

Marco has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.

He is co-founder of HRV4Training, endurance coach at Destination Unknown, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.

Social:

• Substack

  • Personal

  • HRV4Training Newsletter

• Strava

  • Personal

  • HRV4Training Club

• Instagram

  • Personal

  • HRV4Training