The Ultimate Guide to Heart Rate Variability (HRV): Part 1

HRV measurement using HRV4Training, the only validated (and independently validated) camera-based app.

In this series of posts, I’ll provide an overview of best practices for your Heart Rate Variability (HRV) measurements (part 1), and tips on how to analyze and interpret your data over the short and long term (response to acute stressors, longer-term trends, etc. — in part 2). I will also include quite a few case studies so that you can clearly see how you can too make use of the data (in part 3). Finally, in the last post, I will cover a few misconceptions (utility with respect to subjective scores, non-training related use, strength training, etc. — part 4).

HRV is nothing new, and fairly simple to use effectively, but poor standardization and methodological inconsistencies make it difficult sometimes for people to make good use of the technology or understand what is reported in the scientific literature.

Hopefully, these posts will help, but please feel free to reach me (@altini_marco) on Twitter should you have any doubts.

You can find the other parts of this series at these links:

Part 1: measurement setup and first recordings

But first, Just a tiny bit of theory

HRV is a term that refers to ways to summarize in a number the variability between heartbeats.

Right, but why do we care?

For a simple reason: HRV is the most practical, non-invasive and cost-effective way we have to measure the activity of the autonomic nervous system (well, part of it, see later). Our body is continuously re-adjusting to maintain a state of balance, called homeostasis. Our heart rate, blood pressure, glucose level, hormones, etc. — react to the challenges we face (i.e. stress!) and the autonomic nervous system works to keep everything in balance so that we can function optimally (e.g. do not develop chronic conditions, or improve our performance).

Thus, heart rhythm (and therefore HRV) at rest is regulated by the parasympathetic branch of the autonomic nervous system, the one in charge of rest and recovery. Hence, measuring HRV is an effective way to capture how our body is doing while trying to maintain a state of balance in response to different stressors (training, lifestyle, etc.).

In particular, a reduction in certain HRV features (for example rMSSD, more on this later) typically means that parasympathetic activity is reduced, and therefore we have not fully recovered or in general, there is more stress on our body and we are unable to quickly jump back to our normal range.

This all assumes that data is collected using valid methods, and following best practices, which I’ll cover in the next sections.

Example of a few seconds of ECG data, including detected beats. The time differences between beats are called RR intervals and are the basic unit of information used to compute HRV. We need several RR intervals to be able to compute your HRV. This is why HRV needs to be computed over a certain amount of time, typically between 1 and 5 minutes.

How do you measure HRV?

You can measure your HRV with various sensors, either a full ECG, or a more practical device that measures the electrical activity of the heart (a chest strap like the Polar H7 or H10), as well as a small number of devices that use optical technology.

In terms of optical sensors, measurements can be taken either using the phone camera (HRV4Training is the only validated app that can do so) or a separate sensor, for example, the Scosche Rhythm24 or Oura ring.

Finally, the Apple Watch is also able to measure HRV reliably when using the Breathe app, and can also be integrated with HRV4Training. If you are an Apple Watch user, please read this other post as well. As I will explain below, data automatically collected by the watch is of almost no use, and you need to take your measurements first thing in the morning (please see also this blog for a deeper dive into the topic of night HRV data:

In this figure, we can see one minute of ECG data (similarly to the data we saw before, but a bit more compressed as we have 60 seconds instead of just 6, shown in lighter gray). We can also see one minute of PPG data collected with HRV4Training, in black. PPG data here was collected using the phone camera. We can also see detected ECG peaks (used to compute RR intervals), and detected PPG peaks, perfectly aligning over the minute of data. This plot should make it easy to understand that both methods are equivalent, and HRV data can be acquired accurately using optical devices. Note that not all optical sensors are able to do so, hence it is important to make sure the device you use has been validated.

To sum up, morning and night measurements are the only two well-established and reliable ways to measure your HRV.

Note that our application of interest here is determining chronic physiological stress level, which derives from combined strong acute stressors (e.g. a hard workout, intercontinental travel) and long-lasting chronic stressors (e.g. work-related worries, etc.). By measuring the impact of these stressors on our resting physiology, we can make meaningful adjustments that can lead to better health and performance (more on this later!).

We are not interested in sporadic spot checks during the day, as these would simply reflect transitory stressors (light physical activity, digestion, having coffee, talking to a friend, etc.), resulting in de-contextualized data with no practical use).

In the latest paper looking at the difference between morning and night measurement, titled Evaluation of nocturnal vs. morning measures of heart rate indices in young athletes, Christina Mishica and co-authors reported that “heart rate and RMSSD obtained during nocturnal sleep and in the morning did not differ”. This means that as long as we use the same method consistently over time, and follow best practices, both morning and night data are as good at measuring baseline physiological stress.

Here are the most important best practices in either case:

  • Morning measurement: you wake up, grab your phone, and measure either using the phone camera, a chest strap or a reliable optical sensor (such as the ones mentioned above). Your morning measurement will last between 1 and 2 minutes normally (longer is possible but not necessary) and return an HRV number.
  • Night measurement: you wear a device every night, and the device measures your HRV during the night, typically in chunks of a few minutes, then the average of these chunks is computed as your night HRV number. Note that if you are measuring just for a few minutes over the entire night, data will be affected by the circadian rhythm, sleep stages and misdetection of sleep stages, causing reliability issues that are described in detail in this post. Make sure you use a device that gives you at least 4–5 hours of data during the night, to avoid these issues.

What’s the difference then?

As mentioned earlier, both of these methods are valid and have been used many times in research and real life to quantify physiological stress (see an overview here), hence I believe you should really pick the one that suits you better.

There is no advantage in using one method or the other, however, there can be some differences based on stressors timing. Let’s try to clarify this aspect a bit.

Stressors (alcohol, but also training, food, etc.) impact your resting physiology acutely, and therefore late stressors will impact your night HRV data more. By the following morning, everything might be perfectly normalized (unless the stressor was particularly strong). In my opinion, having worked with both types of data for a while, morning measurements might be a better way to capture what you can do on a given day, while night data might better capture retrospectively your behavior. If you often eat late, drink, or exercise in the evening, morning data might be more helpful. If you have a daily routine with fewer late stressors, your night data will be very well aligned with morning measurements (see for example my data here).

This being said, the two measurements are effective in measuring baseline physiological stress, and therefore if you prefer to wear something over the night, by all means, get a device that does so. If you prefer not to wear something during the night and just to take a measurement in the morning, then go that way. If you are not sure this is for you, you can use your phone camera and invest as little as 10$ in measuring your physiology daily and accurately.

HRV4Training integrates with all standard sensors (chest straps, Scosche Rhythm24, Apple Watch) as well as the Oura ring, hence with either method you can get the same analytics and interpretations that I will show later on: measuring is only the first step.

Personally, I prefer the simplicity of a morning measurement using the phone camera, as I have used it daily for over 8 years since I developed the technology.

It is of course key that the sensor used to measure in the morning or during the night is reliable, and this is why we recommend only the ones above, since we have either directly validated them, or have seen validations with respect to ECGs. This cannot be said for most other systems on the market.

A recent independent validation looking at the accuracy of commercially available HRV apps and sensors reported the following median errors: 3% for HRV4Training + Polar H10, 4% for the Oura ring, and 5% for HRV4Training using the camera. This is confirmation of the quality of the work we have been doing for the past 8 years, starting with early analysis of the accuracy, then our validation paper, and finally with this independently run study confirming the accuracy of the methods we have developed for both the camera-based acquisition and artifact removal for RR intervals acquired from external sensors. Note how PPG (both HRV4Trainign and Oura) outperform even systems relying on chest straps, such as Firstbeat in this paper (10% error).

Remember, consistency is key

The list of confounding factors for HRV analysis is pretty much endless. Why? Because as I’ve explained at the beginning of this post in our short theory section, HRV is nothing less than a mechanism triggered by our body in response to stress (and guess what, pretty much anything is a stressor).

This explains also why it is absurd to think that HRV is irrelevant to a specific sport (more on this in part 4). Regardless of the type of sport you practice (or if you do not even practice any sport), stress is part of everyone’s life, and therefore we can use HRV to determine how we are responding to stress and implement changes if necessary. However, this also explains why it is key to either rely on the morning routine or on night data, and it is rather pointless to measure any other time if we are interested in actionable advice on our baseline physiological stress level.

Make sure to follow the following tips in order to acquire reliable data consistently over time:

  • Measurement time and position: try to measure always in the same body position, normally I recommend simply lying down in bed, as that’s the easiest for most people. If this is not possible (e.g. you have kids jumping on you), it’s perfectly fine to have a slightly different routine, for example going to the bathroom, sitting there, and measuring. The important bit is to try to do it, in the same way, each morning. Research using our app tested measurements taken 2 hours later at the facilities, which might be more practical in team settings, and found good correlations, hence you do have some margin, but the less you are exposed to confounding factors (activity, caffeine, or just getting anxious or upset after reading your email or social media), the better.
  • Measurement duration: clinical practice recommends 5 minutes of data to be used for feature extraction, however in recent years more and more studies were able to show that much shorter windows provide equivalent results, and more practical 60 seconds recordings are sufficient, especially when we look at time-domain features. We highly recommend using only 1 or 2 minutes of time for your measurements.
  • Measurement frequency: while 4–5 times/week will get you a good baseline, valuable information might be lost (e.g. weekly variability in measurements). Less than 3 measurements per week might be insufficient to get a reliable baseline hence we recommend measuring daily and trying to make it a habit. As HRV4Training uses the past 60 days of data to understand day-to-day variations in your physiology, the more data the more accurate will the app be at understanding when stress levels are consistently higher than your normal.

That’s all. See? Pretty easy, just relax, breathe naturally, and you’ll be collecting reliable data representative of your physiological stress level.

What does the HRV number mean?

HRV is determined by computing so-called features, starting from a series of RR intervals, or differences between heartbeats, as we have seen in the figures at the beginning of this post.

This means that on the contrary to heart rate, which can be thought of as an almost instantaneous value, HRV requires a certain amount of data to be accumulated, before it can be computed. In terms of features, the sports science community through the work of many in the past 10–15 years, settled on rMSSD as the most meaningful and practical HRV value to use in applied research and real life.

Why rMSSD?

Because of how our physiology works. In particular, the vagus nerve (representative of the parasympathetic system) acts on receptors signaling nodes to modulate pulse on a beat-to-beat basis while sympathetic activity has different pathways with slower signaling. Hence, beat-to-beat changes computed as rMSSD reflect parasympathetic activity (just math, based on how this feature is computed). The same can be said of HF (the high-frequency power), another feature often used, but with the shortcoming that HF is more tightly coupled to your breathing, and therefore rMSSD is preferable from this point of view as well.

Without going into another primer on HRV, parasympathetic activity represents our body’s rest and recovery system, and can be captured in terms of HRV: a stressor might, for example, induce a physiological response in terms of reduced parasympathetic activity, which translates into lower HRV as the nervous system modulates heart rhythm in response to such stressor.

The HRV number that you see in HRV4Training is a transformation of the rMSSD value, typically resulting in a score between 5 and 10. This score is entirely based on your HRV, to make sure you can assess how your body is responding to stress, physiologically. Your sleep, activity, and other questionnaire data is there for you to contextualize the data, but will not influence the score, as we believe the score should represent how your body has responded, nothing else.

Example of an acute reduction in HRV, representative of higher stress. In this case the reduction was associated to travel. We’ll learn more about how to use and interpret the HRV data you are collecting, in part 2 of this guide.

What’s next?

Now that you know how to measure HRV and understand what the data means, we can look at how to analyze the data in a meaningful way.

I will cover in the next section the concept of the normal range, but the important bit to remember, for now, is simply that HRV data is highly individual and has an inherently high day-to-day variability. This means that it is not meaningful to compare to others and that in your own data there can be large fluctuations between consecutive days.

​What are the implications? To make effective use of the data, we need to be able to determine what changes are trivial, or just part of normal day to day fluctuations (what we call your normal range), and what changes do matter and might require more attention or simply truly represent a positive (or negative) adaptation to training and other stressors (HRV4Training does the math for you).

In the next posts you’ll learn to analyze how training and lifestyle impact your physiology, and potentially make changes towards better health and performance.

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

He is the co-founder of HRV4Training and loves running.

Twitter: @altini_marco

New to HRV4Training?




Founder, Data Science @ouraring Lecturer @VUamsterdam. PhD in Machine Learning, 2x MSc: Sport Science, Computer Science Engineering. Runner

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

The Undeniable Power of Intermittent Fasting

A 10-Kilometer Epiphany

Christian Drug & Alcohol Rehab Centers Hockessin DE (855) 419–8366

The Different Branches of Yoga: How to Choose The Right One For You

Goodbye Apple Watch Exercise Ring Streak

What Is A Healthy Diet?

DIY Self Care: Do-It-Yourself Face Masks in the Kitchen

Managing Stress to Manage Burnout

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Marco Altini

Marco Altini

Founder, Data Science @ouraring Lecturer @VUamsterdam. PhD in Machine Learning, 2x MSc: Sport Science, Computer Science Engineering. Runner

More from Medium

3 Key Steps to Make Use of Heart Rate Variability (HRV) data

Why Multitasking Compromises Productivity

How to Look After Yourself When Treating SARS-CoV-2 at Home

Cold Swimming to Wake Up Fitness