Shownotes

Marco's background

03:38 -

• I have a mixed background in technology and physiology.
• I have degrees in computer science and engineering, a PhD in data science and a master's in sports science in human performance coaching.
• I have developed technology to measure physiology, heart rate variability, and exercise with different methods. We use simple sensors like a camera to measure these parameters accurately without sensors.
• But it is better to use sensors that you can pay and use in different apps to access this data.
• I run a small business called HRV4Training, which started almost ten years ago, and that is what I do now with some teaching in Amsterdam. I am also an advisor to another company that measures physiology. Oura Ring allows us to measure HRV during the night and day.
• I love to run, and I sometimes cycle as well. I do cycling because if I ran all the time, I might get injured too quickly. (a break from running)
• Despite the abundance of bike lanes in the Netherlands, the weather is not great for cycling outdoors.
• At the moment, I have been working on increasing my running mileage to do a 100k race.
• I do not have much talent, but I do love the sport.

The concept of HRV

06:03 -

• Heart rate variability is the variation between heartbeats, and we measure this variation.
• We care about this parameter because heart rate variability happens from the autonomic responses to stress.
• The heart would beat at the same pace if we measured it compared to its pacemaker. But we know that heart rate beats slower at rest because of the inhibition of the autonomic nervous system and the activity of the parasympathetic system. The heart rate is lower at rest, increasing the variability between beats.
• This change in the autonomic nervous system responds to stress. That is why it has become relevant.
• As we stress, we get a response from the autonomic nervous system reflected in a change in the heartbeat. Therefore, HRV correlates with stress and the autonomic nervous system.
• We cannot measure this system, but we can measure this parameter that informs us how the system responds to different stresses.
• We can use it to understand how we respond to different stresses acutely and chronically.
• Under stress, we have inhibition in parasympathetic activity and reduced variability (lower HRV). So, heart rate becomes more constant and higher.

Application of HRV for acute vs chronic measurements

08:30 -

• We can think of acute stress on different scales. Usually, we consider more substantial acute stresses to things like hard sessions or getting sick or intercontinental travel. (Anything that has an impact of 24 hours)
• When we measure physiology correctly, we can capture those stress types, which is crucial. We can measure physiology all the time, but there are moments in which we can better capture the resting physiology, which is typically first thing in the morning or during the night.
• The contextualise measurements we can take daily do not capture these acute responses. The reason is that anything that affects the autonomic system would constantly change. Even the most ridiculous things affect the autonomic system. Sneezing, having a coffee or going upstairs create acute stresses in a way.
• However, the time scale is different. (only a few seconds)
• You do not want to capture those stresses, but you want to capture the actual stresses that make an impact on your body.
• The stresses that impact your body at rest will impact your performance.
• Therefore, it is essential when we measure but also what we measure.
• Chronically, we would look at the data in the long term. And in these, it is more challenging to understand which stresses impact more if we talk of a long training block or psychological stresses that will affect our physiology.

Examples of chronic stresses that will affect your HRV

11:43 -

• Even with sickness (e.g. covid), we have seen suppressions that are acute and long-lasting. So, you have an acute change that remains a trend over weeks or months.
• There could be more chronic issues, but we can spot them precisely by looking at the data. (looking at your running average)
• We look at how the baseline changes over time concerning your historical data. In this way, we can develop a model that represents your physiology and that changes over time in the longer term.
• And I believe it is essential. The physiology will change even if nothing happens.
• For example, resting physiology differs from the winter to the summer. This kind of data we can interpret with simple measurements allows you to quantify the normal ranges and analyse how the data is behaving concerning physiologic abnormalities. (abnormalities are usually negative responses to different stresses we face)

Pros and cons of different methods for measuring HRV

14:20 -

• As we have more sensors, we have to understand better the differences between the methods, which provides reliable data for the user and provides more noise and is more problematic for the user.
• There is a distinction between a sensor that presents accurate or meaningful data.
• You can use an accurate sensor, but that provides no meaningful data because we take samples at the wrong time.
• For example, we have the Apple watch. This device is accurate because it can measure HRV as good as an ECG.
• However, if you use this sensor throughout the night, it will give you some sporadic measurements. And that is not useful because we have different components that impact HRV during the night.
• Many people think it is better to use the sensors at night because you cannot influence the measurements. However, your autonomic system is still very active when you are unconscious. (e.g. in rem sleep, the activity is as high as when you are awake)
• If your sensors take a sample during the sleep phase, you capture data that does not represent your baseline physiology.
• It is a bit problematic. It can also take samples at different hours, and this does not consider the physiological changes that happen during the night.
• Usually, your resting heart rate goes down throughout the night, and your HRV goes up. So, if you measure it once at 2 am and another at 5 am, the values will differ.
• It leads to problems because you introduce variability during the sleep stage that you do not have if you measure it during the morning.
• We are more or less at the same "sleep stage" at that period.
• We should make some considerations, but we should avoid complicating things. The solution to solve this is straightforward: average the data of the whole night.
• In this case, the circadian rhythm and the sleep stage components will have the same impact throughout the measurements. Therefore, the average values of the night become a good marker of resting physiology.
• Some sensors provide that information. Oura Ring is one example, and the Polar Advantage watch takes four-hour samples during the night to give an average of these four hours.
• I believe that is a reliable method as well. (if you take four to five hours of data) You can take reliable values from both HR and HRV.
• A couple of weeks ago, a paper compared morning measurements and night data. And these measurements had a strong correlation with each other.
• Applying this according to the best practices makes these methods similar and provide the same insights over time.
• The absolute values will differ because we are measuring during different periods of the night/day.
• You cannot alternate between different measuring methods. However, it is more important to understand how the data changes over time. (you can capture this well with both, and you can use which methods you prefer the most)

Measuring different parameters with the same devices

20:10 -

• We have more and more wearables and devices that can measure many different parameters, but that can be problematic.
• First, we mix behaviour and physiology. If you sleep less or exercise more, the model will penalise you.
• The results will present that you will need more recovery. However, your physiology may not reflect that, and you may be excellent.
• Nevertheless, you have a lower score and do not know if your physiology is stressed or if your behaviours have changed.
• The results are more informative when you look at them together, but you cannot understand the causes of different scores changes.

Different devices for measuring HRV

22:03 -

• The original method to measure these parameters was chest straps, which are still an option.
• Chest straps are accurate, and we recommend the Polar and most recent Garmin chest straps.
• The strap will be accurate if you take care of it. It needs to be humid in the morning or not have good contact with the skin.
• I would say it is handy to take more extended measurements.
• Otherwise, if you use the camera on the phone, we recommend taking shorter measurements for our features to evaluate the autonomic nervous system. (through HRV or RMSS)
• Studies show that measurements of only 60 seconds are equivalent to what you would get with five minutes. It is handy for most people. If you use the camera on the phone, it is also easier. (holding a finger on the camera becomes impractical for five minutes)
• When using a camera, you cannot move (or try not to). It is why sensors only work when you sleep or take a short measurement.
• Optical sensors are not suitable for measuring HRV or HR when you exercise. There is a lot of movement involved, and so the sensors make many averages of the values and do not get a stable value.
• There is only one sensor that you can use for resting measurements and exercise. However, it has different modalities. There is one mode (HRV), and you cannot move. Even if you contract or relax your muscles, it will generate many artefacts. Optical data is sensitive to movement or disruption in the signal.
• It looks at blood volume changes in your capillaries. Anything that messes with that will create noise with the data.
• But if you are still for a minute in the morning, these methods can work very well.

Laying down vs sitting up to do the measurements

25:19 -

• I had many conversations about this with Andrew Flatt, an expert in these areas.
• He stands up to do these measurements because when you stand up, you cause some stress. If you measure within a minute from that, you might get slightly better physiology that day.
• You are analysing the response to the stress and not just the physiology at complete rest.
• I started playing around with this more. I would measure myself always laying down, but now I measure it also sitting.
• I have six months of data from the two options, and they agree with each other in my case.
• It is not a definite answer because I am the only sample from the data. For me, it captures both acute and chronic stresses over time with differences. The day-to-day variability is higher when sitting, which I find surprising. However, it could explain the stress trigger you measure. When you sit, you create stress and have a broader range of responses to that stimulus.
• I think there is a good case for not laying down. Athletes with a low heart rate may consider using other positions. There could be a situation where you are in a period of parasympathetic saturation, which I think is quite rare. It means that even though the parasympathetic system is high, your HRV does not reflect it well.
• It can happen if your heart rate is in the low 30s and you have a high training volume.
• You might feel great, your heart rate is not increasing or even reducing, and your HRV stays the same. The reason for this could be this saturation issue.
• Nevertheless, laying down is suitable for 99.99 % of the cases. And we also have to consider trade-offs. If you lay down, it is easier not to mess it up. If you do other things, you might perform measurements wrongly most of the time.
• When you stand up, you might measure HRV too soon. Thirty seconds in the morning might feel like five hours when you are only waiting.
• Everything might feel rushed, so we need to consider these things. That is why I still recommend laying down as the easiest way to capture high-quality data.

Marco's new study on HRV

Summary of the study

28:59 -

• In this work, we looked at primarily two things: population level (30 000 people), and we analysed how HRV and HR changes concerning personal characteristics. (genre, BMI, physical activity level and age)
• We wanted to understand the factors behind differences between people's HRV and HR and the relationships between all these parameters.
• In the second part of the paper, we looked at each person and analysed one year's worth of data.
• We looked at the relationship between different stresses like alcohol intake, sickness, training at different intensities and the menstrual cycle.
• The goal was to improve the ability to inform better on how we can use the data. (a tool at a population level and as a tool to measure stress at the individual level)
• We had about nine million measurements.
• In the past 15 years, we had studies looking at HRV data in the lab. However, I feel that is not a realistic setting. You have to ask people to come to a lab without eating or wait two hours before doing something.
• You make them sit or lay down on a bed and measure the physiology for 30 minutes. And we expect that to represent a sample of the resting physiology.
• Luckily, thanks to the technology we developed, we can perform those tests naturally. However, the stresses and the reference points are something we have to deal with while studying.
• However, with the size of the population, we can accept some level of noise in the training or sickness reported. When you have so much data, you can better analyse different responses even when the data is not perfect.

Results at a demographical level

33:18 -

• We found some of the apparent relationships we expected to see at a population level. For example, women have slightly higher HR than men, sub-optimal resting physiology (lower HRV) correlates with a higher HR.
• People with a normal BMI have the highest HRV and lowest HR. We see a less ideal profile from the average values (under or overweight).
• There is no change in resting HR when comparing people from different age groups. We are not saying HR is the same, as there is much fluctuation from individual to individual.
• We saw stable HR for people ranging from 20 to 60 years old. However, HRV is much higher in the lower 20s than in people in their 60s.
• Then, we started to break down these relationships by physical activity level. This part is where things get more interesting.
• First, we looked at the relation between physical activity level and HR. A higher level of physical activity correlates with a lower HR.
• When you look at this correlation across different age groups, this relation remains precisely the same.
• People who exercise more in their 20s or 60s have the same HR.
• For HRV, the relation with physical activity level is weaker.
• It is something we discussed in other situations because I would not use HRV as a marker of fitness or the level of any athlete. (I have seen an elite athlete with low HRV) It could be due to genetics or other factors.
• However, there is a more substantial relation between HRV and physical activity when you are younger.
• Therefore, HR is a better marker of cardiorespiratory fitness. And that is why we use it to predict fitness at the population level.
• For each category, even if we divided it by age or BMI, the distribution of values of each group is too broad.
• So, these variables cannot explain the differences between people, and there are additional factors we need to consider to evaluate HRV better.
• People might want to start to analyse these parameters to change them and have a higher HRV. If you already have a healthy lifestyle, it might be impossible to change your HRV because of genetics, for example.
• It does not mean it is not handy to use it. It is not a marker you can optimise itself, but you use it to maintain your health or performance.
• HRV is a tool for tracking stress for people who are already healthy and active because all the other parameters do not change much.
• As always, context is essential when addressing this issue. We might want to improve the lifestyle, so we should see changes in HRV.

Stresses that affect HRV

42:14 -

• We looked at training intensities. First, we grouped them into low and high intensity and then into rest, low, average, and high intensity.
• The athletes were the ones who gave us these intensities.
• With increasing intensity, we saw increased changes in HRV. (HRV is more sensitive than HR with increased intensities)
• Then we had sickness, alcohol intake and the menstrual cycle.
• Looking at the training intensity response, the changes in HR are lower than HRV changes. Using HR is tricky because we cannot access those changes. (daily vs meaningful variations)
• If we look at HRV, we are more confident when we see a suppression in HRV because it means there was some stress involved in that change.
• The HRV change remains sensitive to the same stresses when comparing age groups. Whether you are 20 or 60 years old, you report the same HRV reduction in response to high-intensity training.
• Resting HR changes are less substantial when we look at the same response. And as you grow older, the changes in HR become smaller and smaller. Therefore, you cannot take any information out of it because of the daily variability.
• It was surprising because we know HRV reduces with age. So, with lower values, you would expect the daily variability to be less significant. (and we found the opposite)
• It means these parameters can capture training stresses at any age. This information is handy because we usually perform studies on a younger population.

Sickness and alcohol

45:45 -

• These factors are the most substantial stresses we found by far concerning HR and HRV.
• The changes in HRV and HR are four or five times larger with sickness and alcohol compared to training.
• It is crucial to understand this because we cannot rely on this technology if we do not improve our lifestyle first.
• If we drink three or four times per week, HRV will not present how we respond to training. It will only respond to how we are reacting to the alcohol intake.
• HRV is not a specific parameter to training but a general marker of stress. And the changes because of some stresses are significant.
• Hopefully, we do not get sick too often, but alcohol intake can be.

Menstrual cycle

47:08 -

• We replicated the results we have seen in previous studies. It means that during the second phase of the cycle, we have a suppression in HRV and an increase in HR.
• Those changes are relatively small compared with sickness and alcohol and more comparable to the training responses.
• It still matters because we will have these oscillations in athletes with a regular cycle to understand these changes in our daily lives.
• It is a valuable context indicator. There is no agreement on how athletes should proceed during these phases, but we can better understand the changes when they happen.
• Therefore, we have to think holistic to health and training.
• While HRV is more sensitive, it is also less specific. It means that any stress you have will affect the HRV.
• HRV may decrease, and we do not know the causes. If the stresses are intense (sickness and alcohol), you will also see a change in HR.
• Therefore, we can use HR as a more specific marker for addressing changes associated with health. (more alarming parameters)
• HRV allows us to capture more minute changes and evaluate our training response.
• All the responses will depend on your baseline. If you are well-trained, you should not see suppressions in HRV when training.
• If you are not training too much or too intense and HRV lowers, you negatively respond to training. (did not bounce back quickly after that training stress)
• If we have well-structured training with the proper training zones, we should have a positive trend in your HRV response.
• People think they should have a negative response in HRV to a hard session, but it is not the case.
• In an ideal world, you are always within your normal range.

HRV, sickness and training

53:20 -

• Every sickness is different, but in the initial phase of getting sick, you might have already a change in physiology before you realise you are not feeling well.
• You can notice it in your morning and exercise data, where your HR is higher for the same intensity.
• When you start to feel better after sickness, we do not have concrete evidence that your symptoms would be fading out before or after the data shows any changes.
• It is also complicated to track even with Oura and other companies doing this work during covid. They tried to see differences at the beginning of the infection, but it is too difficult to find these reference points because you can have ten to fifteen different types of symptoms.
• And you can also have these when you are healthy sometimes.
• You have to check when you do not start to feel well, and defining this makes it challenging to do a systematic analysis.
• If you have been sick for a couple of days and HRV has been low, you can use the data to understand when to return to regular training.
• You can notice an HRV change to previous pre-sickness values when you recover. (you are returning to normal and can start training)

Alcohol intake

57:35 -

• In the app, we have a field category with non-scientific choices. (no alcohol, a little or too much)
• In the study, we included only the extremes. People shared their data, and in terms of resting physiology, having a beer or a glass of wine does not affect the parameters.
• It does not mean that there are no other changes.
• The effect of alcohol is still up for debate, especially if you look at small quantities.
• We only capture tiny or no changes in resting physiology from the data.
• For example, last summer, I had some arrhythmia episodes, which triggered me not to drink anymore.
• Regardless of no changes in physiology, alcohol increases your risk factor for arrhythmias. So, if you have heart problems, alcohol is the first thing you should drop. And if you have arrhythmias, you cannot measure HRV anymore because there are too many disruptions.
• One good thing about the app is that it allows us to capture the frequency of some of our habits.

Conclusions and practical applications from the study

1:02:00 -

• Athletes or people that use these systems should not expect a change in HRV.
• The best way to use the data is for managing stress and capture the stress responses. It captures stresses better than HR.
• At the same time, it is non-specific. When you have a drop, it can be because of many factors, and we can see their relationship.
• If you have an increase in resting HR, you likely have more substantial stress, and you should be cautious about it.

HRV guided training

1:03:46 -

• HRV guided training means you adjust your training based on the data, which consists of making some changes in the planning, reducing the session's intensity, or even doing an extra session depending on the protocol.
• There was a study where athletes did more intense training than other groups. However, there are limitations to the number of intense sessions you can do in a week.
• The most commonly used protocols are the ones where we control the intensity if there is a suppression in HRV. When athletes present values below or above baseline, we reduce the intensity to control the stress on the body.
• Usually, we only control the intensity of the sessions because it is the intensity that primarily suppresses HRV. It causes several disruptions in the body.
• When we talk of ultra-endurance efforts, the duration can also be absurd, so we need to consider that.
• Sometimes, we only have to use common sense. If you have a suppression, you should adopt a lower intensity training and see how you feel.
• If you feel good at the low intensity and the HR is low, I think doing a more extended session is acceptable. I apply this methodology to myself.
• We always need to deal with real-life athletes that are not professionals. If you have a low score on Sunday, it does not mean you should not train because you have Monday to recover.
• It is on removing the intensity where we should be flexible. You think that workout might lead to some changes, but the research shows that if you are not in that state where you can assimilate the workout, it does not make much sense. You will not get better by doing it.
• There is no problem doing a short session if you are used to training every day concerning low-intensity training. The low-intensity sessions cause much fewer disruptions in the autonomic activity. And we can check this information with studies that look at HRV before and after exercise.
• Athletes that did a low-intensity exercise of one hour vs two hours have similar HRV responses, and it is always much better than a short high-intensity session, where HRV takes a long time to stabilise.
• If you only train three times per week, you might want to skip that workout.

State of evidence for using HRV guided training

1:09:53 -

• I think it is the same as it has been in previous years. We have more studies looking at the performance or other submaximal testing. (evaluation of HR at a specific power, for example)
• When using HRV guided training, the performance metrics are the same or better than the control group. The reason is the timing of the stress, which means we do not stress the body when it is already in a negative stress response state.
• It leads to similar or better outcomes. In my view, it is essential because there are a lot more stresses to consider, and these studies are short. (adaptations take a long time to occur)
• There are other factors to consider, for example, travelling and adding other stresses.
• So, if we can obtain the same or better results without the sessions that add stress that is not well assimilated by the body, we should explore further.

How Marco uses HRV

1:12:15 -

• I am busier with working than with training, and I have to judge my goals concerning where I want to perform.
• I live from my work, not my training.
• For example, the relationship with work stress is often more robust than with training.
• If I have periods in which mentally I am not in a good place, or maybe it is spring, and I have allergies, the data always looks pretty bad.
• The way I came to use it is not frustrating anymore, which could be the first thought people have when people see the data telling you to go slower for weeks. (because maybe your health is not significant)
• I have to work with what I have. We have to be aware of the current state of the body, and instead of always trying to perform in a way we would like to perform, we do what we can on a given day based on everything that is happening with us.
• It means I go out and do a lot of easy training and use the better days for some intensity if everything looks good.
• I try to achieve a place where the consistency in the day to day scores is better. For me, stability is more critical than achieving higher scores.
• You will train intensely more often if you can get more stable values. This regime is more complicated to achieve if you do not have that stability. (lower day to day jumps in HRV)
• I do not go hard when HRV is lower in terms of intensity. This aspect is easier to control because my workouts are not frequent.
• In conclusion, I try to manage the values in terms of training, stress and health concerning all the mentioned above.
• We have a lot of sensors and data to analyse. But the analysis we do from data is crucial to apply it correctly. And this is a learning process.

General questions

1:16:57 -

Can you give three pieces of advice to athletes to take from this interview?

If you decide to measure your resting physiology, use a proper tool for measuring in the morning or during the night. You should do that consistently over time to get good data and try to relate that to the different stresses you face. Pay attention to not making changes for several weeks or months. For example, you can analyse how your body reacts to different amounts of alcohol, sickness or training intensities. If you see that HRV is helping you, you can try to be more flexible with the workouts and the high-intensity efforts to maintain your HRV levels in the normal range. (which should most of the time) A lower HRV score on a given day does not mean your HRV is getting suppressed. We have daily variability in the scores, and we should remember that when we analyse the results.

What is one thing within coaching or training you are now learning/curious about or fascinated by and why?

Staying on the HRV topic, I have been experimenting a lot more with measurements before and after training based on some research Stephen Seiler did 15 years ago. There has been no study following up ever since. There seems to be no impact in HRV scores with one hour vs two hours easy sessions.

Typically, we want to look at HRV during the morning or the night to look at the big picture. But if you want to look specifically at training, we should measure before and after. After means at different time points. (5, 10, 15, 30 minutes after training) We can analyse how HRV bounces back to pre-training value. (if they do) Your HRV could be higher after training if the session were low-intensity. And that is a surprise to me. I always thought that you train, and it will take time to get back to normal. I did some testing on myself, and I found those results. So it will be interesting to understand what kind of intensities allow this to happen. Moreover, we can better study how different intensities affect suppression or longer recoveries. We can use the information to adapt the intensity of the session.

Rapid fire questions

What is your favourite book, blog or resource?

Training Essentials for Ultrarunning: How to Train Smarter, Race Faster, and Maximise Your Ultramarathon Performance by Jason Koop

What is an important habit that benefited athletically, professionally or personally?

I think it is about putting in the work. Personally and athletically, I did not get very far. Professionally, I have never been very clever, so it was always about working, studying and keeping motivated to learn.

Who is someone you have looked up to or who has inspired you?

I do not have anyone in particular, but that does not mean there is no one I admire. I mean that I admire anyone doing good work with solid ethics. Anyone who puts ethics and values before other things in their work and life is someone I look up to and help me guide my work.

LINKS AND RESOURCES:

• Marco's website, Medium blog, Instagram, Twitter and Research Gate
• HRV4Training website and Instagram
• Heart Rate Variability (HRV) – Applications and Insights in 2018 and Beyond with Marco Altini | EP#144
• What Is behind Changes in Resting Heart Rate and Heart Rate Variability? A Large-Scale Analysis of Longitudinal Measurements Acquired in Free-Living - Altini and Plews 2021