Thoughts on Heart Rate Variability (HRV) measurement timing: morning or night?
Earlier today I was reviewing a journal paper looking at morning and night heart rate variability (HRV), a topic I have been thinking about a lot due to my work with HRV4Training and Oura. Reading Andrew Flatt’s tweet yesterday, I got thinking about this a bit more.
In this blog, I would like to highlight a few aspects that are often overlooked when it comes to night HRV data in particular.
First thing first: I do believe both measurements provide us with useful data to track long-term changes in physiological stress as it relates to training and health.
However, there are differences, in particular acutely (on a day-to-day basis), in response to stronger stressors, due to the different timing of these measurements. Similarly, morning and night data might be more or less useful in different contexts, e.g. despite heavy marketing to athletes, night data might be less relevant to them (more on this later).
Thank you for reading.
Please feel free to reach me (@altini_marco) on Twitter for any questions or follow-ups on this blog
Let’s talk about sleep
Sleep is the key restorative process of our body.
If we think about the theoretical benefits of measuring during or after sleep, arguably, an individual’s physiological response should be more relevant when measured after the restorative effect of sleep, not during.
Only measuring after sleep can inform us about our body’s state after the impact of 1) the previous days’ stressors 2) the restorative effect of sleep. This is the real actionable information.
I think it is also time to stop idealizing sleep as this process where nothing happens and is ideal for physiological measurements. During sleep we are unconscious but the autonomic nervous system is very active and variable on a minute-by-minute basis, with variations in HRV that are much larger than when we are awake.
As a matter of fact, wearables try to detect sleep stages based on autonomic nervous system activity, so if it was all stable as argued, we would not be able to estimate all these different states based on that data alone.
From an autonomic nervous system point of view, sleep is a very complex state where a circadian component (autonomic nervous system activity slowly changing throughout the night, for example heart rate reducing and HRV increasing) and sleep stages (autonomic nervous system activity varying greatly on a minute by minute basis, for example with lower and more stable HRV during deep sleep as opposed to REM sleep) make autonomic nervous system activity more variable than when awake in the morning (no sleep stage and no circadian effect).
This is indeed the reason why the only meaningful use of night data is averaging it all, something eventually most wearables figured out.
Measuring during the night is still valuable of course, but not for the reasons typically highlighted. The night is a good time for the same reason that the morning is a good time: it is easily reproducible (on a daily basis, you can measure under very similar conditions) and impacted less by other transitory stressors happening throughout the day (physical activity, diet, mental stressors, etc.) — even though probably impacted more than the morning, since it happens earlier.
Physiological measurements require a reproducible context, which is why it is typically of no use to try to assess underlying physiological stress outside of the night or morning routine.
What about stressors timing?
Another typical misconception I often see is the assumption that night data is more sensitive than morning data because it shows a stronger suppression after a stressor.
Let’s unpack that.
We should not confuse an HRV drop, showing sensitivity to a stressor, with the usefulness or meaningfulness of that drop.
If a larger reduction was the ideal case that makes a measurement protocol not only more sensitive but also more useful, we would measure right after the workout: our HRV would be much lower, that’s as sensitive to the stressor as you can get.
The larger drop is simply due to measurement timing: the night comes first, and therefore your HRV is suppressed even if you might have bounced back by the morning.
Is the night better then?
Or is the night simply entirely missing the point and only reflecting the acute drop, instead of the actual response? Because if it is the latter, the night is the worst time for measurement.
I am not arguing that this is the case but merely highlighting the fallacy in this reasoning.
To reiterate, is an acute suppression captured in night data, but not in morning data, valuable information? Or is it of no use, because you have recovered by the following morning?
Analyzing a response to a workout at different time points (night, and then morning) simply shows how physiology gradually returns to normal (unless the workout is more stressful and impacts also morning data, which is often the case, as shown in our research), but it is insufficient to determine the usefulness of one measurement or the other in capturing individual responses to training load.
Below you can see an example. A typical difference in acute responses for morning and night data following large or late stressors (e.g. hard, long workout followed by a late dinner). HRV is suppressed in the first part of the night, causing a low average (left pic, Oura ring night data), but is back within normal range by the morning (right pic, measured while sitting using HRV4Training, phone camera).
Looking at the data above in more detail, for night data, in the first half of the night we have: HRV = 61 ms, heart rate = 49 bpm. In the second part of the night, we have: HRV = 69 ms, heart rate = 44 bpm. This shows how in the second part of the night physiology is back within normal range, just like in the morning.
Unfortunately, we are stuck with no reference when it comes to recovery, and therefore it is very difficult to establish the effectiveness of one method or the other outside of very simple acute responses.
Is lying down the ideal position to measure HRV?
I have recently changed my mind about this after several conversations with a trusted expert in the field (Andrew Flatt), self-experimentation in different positions over an entire year (so that I could look at the relationship between the data and various acute and chronic stressors) as well as reading the available literature.
First thing first: what’s the difference between positions? As Andrew explains, when lying down, HRV represents cardiac autonomic nervous system activity during undisturbed rest. On the other hand, sitting or standing HRV represents cardiac autonomic nervous system activity in response to physiological stress (orthostasis) with a strong influence from the baroreflex. The physiological challenge exacerbates your response so that if something is off (there is more stress, sickness, or anything else), there will be a much larger change in HRV (and resting heart rate) with respect to the change you would see if you are lying down (sleeping or in the morning).
Thus, an orthostatic stressor, which here means measuring while sitting or standing, makes it so that HRV becomes more sensitive to the stressors you are facing, and therefore more useful. This is why I now recommend measuring while sitting. For more details on these mechanisms, please check out Andrew’s great blog here.
Measuring while sitting or standing can be particularly beneficial if your heart rate is low, e.g. low 50s or lower, since parasympathetic activity is already high. Measuring while sitting or standing, you might also be less prone to another rare issue of measurements taken lying down, parasympathetic saturation (a situation in which your HRV measurement does not reflect parasympathetic activity, basically your HRV would be suppressed despite actual parasympathetic activity being elevated). This is particularly true if your training volume is high.
Personally, I now measure while sitting, and therefore can only do this first thing in the morning, as opposed to during the night.
Below is an example showing acute sickness, and how the orthostatic stressor (i.e. sitting in the morning for a measurement) captured very well the acute phase, possibly with an early warning, as well as the recovery phase, while night data failed to capture any actionable information apart from the acute phase (one day).
What are you interested in?
The whole point of using HRV is to try to capture the body’s response to training or other stressors.
Quite possibly, measuring after the restorative effect of sleep is a better time to assess your state before a new day.
Measuring closer to the stimulus, we capture the stimulus better (how hard it was), but we don’t really capture the response (how we recovered after e.g. 10–15 hours, before the next workout).
Night HRV can be very useful to better understand aspects of your behavior: eating habits, alcohol intake, weight loss, and exercise timing, for example. Similarly to morning measurements, night data can also track very well anything health-related (acute sickness, chronic disease).
However, in the context of daily training monitoring or recovery, I currently think that night data might in fact be suboptimal at times, as too closely coupled to the training stressor, especially if you train in the afternoon or evening. Additionally, the high parasympathetic state during sleep, makes the data much less responsive to daily stressors, have we have seen in the example above in response to sickness (captured well acutely, but not in the recovery phase).
If your goal is to make daily adjustments, morning data might be more informative. I’d recommend checking out Andrew Flatt’s recent experience with sickness for some more insights on the relationship between morning and night HRV.
How is your lifestyle?
Given all of the above, it follows that morning and night data might tell us a different story based on our lifestyle.
However, it is also important to highlight that here we are really going into the nuances of morning and night measurements, which are otherwise often capturing very similar trends, especially under certain circumstances. Here I’d like to show such an example.
In personal experience, morning and night data can align very well over long periods of time, see for an example below:
In the graphs above we have 1) an acute drop (the lowest data point in both graphs, towards the beginning), which was due to sickness (food poisoning). Then, we have 2) a negative response to the heat, which HRV4Training’s trends analysis detects as maladaptation, color coded in orange, followed by 3) increasing HRV: due to good training and a cooler climate. Finally 4) we have a small drop on the last day, for both measurement times.
This level of consistency between morning and night data is likely lifestyle-dependent: my evenings are very boring: eat early, go for an hour walk, no alcohol, go to bed early, same story Monday to Sunday. If your lifestyle is similar to mine, then you should see high consistency between morning and night measurements in the long term.
There is a caveat though: I am not a professional athlete, and I mostly experience negative stress responses associated with work and non-training-related stressors. The relationship between morning and night data might differ for athletes and people with overall higher parasympathetic activity or depending on how well we handle stress.
Wearables today are heavily marketed to athletes. However, when we combine a fairly stable lifestyle, potentially higher parasympathetic activity and a superior ability to handle stress, night data might show less day-to-day variability, basically the same values would be reported on most days. This means that we might get less information out of night measurements.
Part of the issue here can be linked to body position as I have discussed above, as supine measurement can be less sensitive and confounded by parasympathetic saturation. Adding a little stress (e.g. just by being awake during your measurement, or sitting instead of lying down), might better capture your capacity to deal with stress for the day. Challenging the system (by changing body position), might better reflect your readiness, as opposed to a measurement taken while asleep.
is HRV adding anything?
Another way to look at the potential lack of sensitivity to daily stressors of night HRV is to analyze the correlation between your resting heart rate and HRV. As shown in our research, HRV is more sensitive to stressors, with respect to resting heart rate.
However, when measured during the night, the correlation between resting HR and HRV tends to be near perfect (close to 1). If this is the case, it means that HRV is really not adding much to the picture. Below is an example where we can see a very strong correlation between HR and HRV for night data (top plot), but not as strong for morning data (collected while sitting, using HRV4Training’s camera-based measurement):
If changes in HRV are almost perfectly explained by changes in resting heart rate, then HRV is not particularly helpful, when measured in this context.
Below is an example of how morning data can capture all sorts of stressors, both planned (e.g. travel, altitude, training camps) and unplanned (e.g. sickness, or other psychological stressors), even for athletes that tend to take better care of the basics (sleep, diet, etc.). The data comes from a recent case study of a Paralympic Medal-Winning Cyclist.
Here we can see how HRV is in fact more sensitive than resting heart rate to most stressors, when captured in the morning:
If our lifestyle is far from stable (we eat late, drink a fair amount, etc.), then night data might capture well our behavior and provide us with useful feedback on such behavior, but this does not mean that it would be as informative in the context of readiness to perform the next day (more on this in the next section).
Similarly, night HRV might track well with our health over very long periods of time, but be less useful for day-to-day guidance.
The best tool for the job might depend on the question you want to answer, and what you want to track, in relation to your lifestyle, baseline parasympathetic activity, and ability to handle stressors.
What does the research say?
Early research comparing morning and night data is in line with what I have discussed in this blog. In particular, Christina Mishica and co-authors looked at the correlation between night and morning data and found that “heart rate and RMSSD obtained during nocturnal sleep and in the morning did not differ” (see “Evaluation of nocturnal vs. morning measures of heart rate indices in young athletes”).
This is expected when we look at the data at the group level, meaning that if a person has a relatively low HRV, it tends to be low both in the morning and during the night. However, this analysis does not look at how data changes in response to stressors, for example how morning and night data respond differently to a workout, something I’ve discussed above.
A new article by Olli-Pekka Nuuttila and co-authors tries to address this issue, looking at how HRV collected in the morning and during the night relates to prior stressors. The authors report:
“It could be argued that the morning .. being further away from previous stressors and closer to the following, would provide more relevant information on the current state of homeostasis”
“Furthermore, it can be speculated that nocturnal recordings would rather reflect the physiological and psychological load of the previous day than the actual state of recovery and readiness to perform on the following day”
Stressor timing matters, and in the context of training, morning data might better represent your recovery and ability to assimilate additional stress on a given day.
What about data quality, movement, and arrhythmias?
Night data is very prone to artifacts. We can easily take a morning measurement avoiding any movement, but that is not the case when we sleep. Any movement, either tossing in bed or going to the bathroom will cause issues in data quality. Ironically, the more you are having trouble sleeping, the more inaccurate will your data be.
For the few wearables that show you the full night of HRV, you can easily spot peaks (“high HRV”) when moving more, which has nothing to do with your parasympathetic activity at that time, but simply highlights artifacts in the data (poor quality data leading to erroneous measurements). Remember that poor quality HRV can only result in higher values, hence be skeptical of those.
Finally, an important issue rarely discussed is actual cardiac abnormalities (arrhythmias). In the context of measuring HRV, arrhythmias are artifacts. As I have described elsewhere, a single beat out of place will cause a disruption and artificially increase HRV. Normally, when we have such isolated events, we can deal with them and provide accurate estimates of HRV. However, if the issues are more frequent, it can be difficult or not possible to measure HRV. Unfortunately, if your arrhythmia is frequent during the night, there is no point using a device that measures as you sleep.
In this case, the only way to measure HRV is to take a morning measurement during a period in which you have no or fewer ectopic beats. This is not to say that devices using night measurements are inferior in terms of artifact detection or removal. However, in the morning you have control, you can wait a bit, you can assess if the measurement was artifacted, etc. — in the night your data will be impacted by ectopic beats and there is really nothing we can do.
Note that harmless arrhythmia such as premature ventricular contractions have a prevalence between 40 and 75% in the population (i.e. pretty much everyone has them). These issues should be carefully evaluated in sports settings as athletes tend to have a higher prevalence of ectopic beats, especially in the context of endurance sports.
Personally, I had some episodes of arrhythmia that caused clear discrepancies between morning and night data (where night data was noisy, see here), to the point that I tend to be skeptical of HRV captured in the night, unless I can see the same in morning data. Is it really looking normal or trending positively, or is it just picking up some ectopic beats? Can you ever be sure? When I measure in the morning I can feel any potential issue and see the PPG waveform, hence I can trust the data 100%.
Wrap up
The goal of this blog post is to get you to think a bit more critically about issues that are often overlooked when measuring HRV during the night.
Night HRV data is promoted as optimal due to the fact that we are unconscious during the measurement, and have more data available.
However, looking at stressors timing, the restorative effect of sleep, movement artifacts, parasympathetic saturation, lifestyle and potential arrhythmia, night data might in fact be less than optimal in many circumstances, especially if we are interested in using it to guide training.
For any questions or doubts you might have, feel free to reach me on Twitter
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
