Heart Rate Variability and its Applications in Sport Science

A major aim of athlete monitoring in sport science to understand the body’s responses to psychophysiological stress, overall adaptations, and recovery needs (Thornton et al., 2019; Coutts et al., 2018). One proxy measure collected to achieve this aim is heart rate variability (HRV) (Altini, 2024). At a very high level, HRV represents the variation between successive heart beats. Consider an example with two individuals who have a resting heart rate, or number of heart beats over a period of time, of 60 beats per minute. While the number beats over those 60 seconds is the same for the two individuals, the heart beats are not necessarily occuring every second. For a visual representation of this, see the figure by Stephenson et al. (2021). Looking at ECG traces and the R-R interval durations shows that there can be differences in the time between the heart beats while heart rate is the same. The varying time between successive heart beats is a result of the influence of the autonomic nervous system and other physiological processes. Notably, there are different ways to calculate HRV, or HRV indices/features, that reflect different physiological underpinnings. However, this post will focus on the HRV index/feature that is most commonly used in applied sport science and reported by commercial wearable devices.

Heart rate and HRV are modulated by the autonomic nervous system via the sinoatrial node, or the pacemaker of the heart. The primary autonomic nervous system itself can be broken into two branches: the sympathetic and parasympathetic nervous system. The sympathetic nervous system is associated with stress and arousal and is characterized by the “fight or flight” response. On the other hand, the parasympathetic nervous system is associated with relaxation and recovery and is known for redirecting the body’s resources to prioritize “rest and digest” processes. While the autonomic nervous system affects heart rate and HRV through the sinoatrial node, the autonomic nervous system itself is regulated by a complex network of the body’s systems and organs such as the limbic system, hypothalamus, medulla, brainstem, and spinal cord. The connection of the autonomic nervous system to various other physiological and psychological systems makes it a useful mode to understand human capacity, recovery, and psychophysiological state.

The sympathetic and parasympathetic branches of the autonomic nervous system have distinct effects on heart rate and HRV. The changes in the heart’s rhythm via the effects of the autonomic nervous system are a result of different neurotransmitters depending on the branch. The primary neurotransmitter of the sympathetic nervous system is norepinephrine, or adrenaline. During the “fight or flight” response, norepinephrine elicits the cascade of physiological processes associated with a state of stress and arousal such as pupil dilation, increased salivation, the release of stress hormones, and increased muscle tension and blood pressure. Along with these changes are an increase in heart rate and decrease in HRV. Importantly for monitoring athletes, a chronically elevated sympathetic nervous system response can have negative, unintended consequences such as overtraining syndrome, symptoms of burnout, and elevated risk of injury and illness (Stephenson et al., 2021). The main neurotransmitter of the parasympathetic nervous system is acetylcholine. When acetylcholine is released, it very instantaneously binds to receptors near the sinoatrial node and can slow down even the very next heart beat (Altini et al., 2024; Russo et al. 2017; Katona et al. 1982). Within milliseconds of acetylcholine’s release, the effect on HRV can be observed and occurs much faster than effects on overall heart rate. In addition to increasing HRV, increase in parasympathic nervous system acitivty also causes pupil constriction, vasoconstriction, decreased heart rate, and increased stomach and gastrointestinal functions.

Since the body’s response to stress can be observed through the autonomic nervous system’s impact on heart rate and HRV, these measures are used in athlete monitoring to understand to the specific acute stressor of training and competition. Training shifts the autonomic nervous system activity towards a sympathetic dominance which manifests as elevated heart rate and decreased HRV (Jeukendrup et al., 1992; Plews et al., 2012; Stanley et al., 2013; Altini & Plews, 2021). For instance, our paper exploring nighttime sleep and recovery across as National Championship women’s ice hockey season revealed that, across the entire season, there was a significant decrease in HRV coupled with a significant increase in heart rate (Merrigan et al., 2023). Anyone who has participated in or been around a collegiate sports season can empathize with the draining feeling that accumulates across a season, and this cumulative fatigue was elucidated by HRV and heart rate. It is important to note, however, that other confounding acute factors (e.g., not including long-term considerations such as age or fitness level) such as alcohol intake, illness, menstrual cycle phase, and other stressors also influence autonomic nervous system activity and, subsequently, heart rate and HRV. In my opinion, that is why heart rate and HRV are useful for identifying athletes who need a specific recovery intervention to prevent sustained elevation in sympathetic nervous system (e.g., floatation-restricted environmental stimulation therapy) and not an absolute determination of an athlete’s condition.

As mentioned earlier, HRV can be calculated in a number of different ways and each represents a different specific physiological effect (Altini et al., 2024). These HRV processing methods can be categorized as frequency, time-domain, or non-linear methods. Frequency and non-linear methods are not commonly reported by commercial off-the shelf wearable devices and require more extensive analysis to compute and longer (5-minute) epoch or window sizes. Therefore, this post will focus on the time-domain features and specifically the root mean square of successive differences in the adjacent N-N intervals, or the time between normal R-R peaks on an ECG trace. Again, and HRV feature is a way to compute beat-to-beat/R-R intervals in a certain amount of time into a single value. The certain amount of time is the window size and also called an “epoch” (I genuinely still don’t know if it’s pronounced “epic” or “eee-pock”). The length of the epoch depends on the length of the total duration of data that were collected and the processing methods or features being calculated. The most common time-domain features are the average of beat-to-beat intervals (AVNN), standard deviation of the beat-to-beat intervals (SDNN), proportion of consecutive beat-to-beat intervals that vary by more than 50 ms (pNN50), and finally, the root mean square of successive beat-to-beat differences (rMSSD). rMSSD is associated with short changes in heart rhythm and is commonly reported on commercial wearable devices such as Oura ring, Whoop, and Apple Watches and also most commonly used in research. This feature best reflects parasympathetic nervous system activity and is regarded as an accurate measure of vagal tone (Altini et al., 2024). rMSSD is calculated with the following formula:

RMSSD = sqrt( (1 / (N - 1)) * Σ (RR_{i+1} - RR_i)^2 )

Where RR_i represents the inter-beat intervals (in milliseconds) and N is the total number of intervals.

rMSSD is calculated for each given epoch over the duration that is chosen (it is time-invariant so generally the duration of windows does not significantly alter the outcome, but most wearables use 60- or 300-second epochs) and then averaged across all epochs (Altini et al., 2021).

Generally, it is not productive to compare HRV between individuals or across groups. While a low (rMSSD < ~25 ms) is associated with poor fitness, metabolic syndrome, and other cardiovascular or neurodivergent (e.g., ADHD) disorders, HRV has very high between-individual variaibilty. For example, a recent audit of over 30,000 HRV records collected during sleep from collegiate athletes found the standard deviation of HRV rMSSD was ~30 ms (with an average around 90 ms; publication in review). HRV is most useful when it is analyzed longitudinally within a single individual (Altini et al., 2024; Stephenson et al., 2021). For example, understanding how a given day or week’s (weekly coefficient of variation) value compare to what is normal for that athlete. It is important to establish a baseline HRV for subsequent comparisons. This generally involves consistent, daily measures for a minimum of 1-2 weeks but, in my opinion, closer to 4 weeks. Once this is established, within-individual statistical comparisons can be used to understand recovery states. Fortunately, statistics for individual analysis in applied sport science are becoming well-described and easier to understand (REF). One commonly utilized calculation is a standard score such as a z-score. This involves calculating the difference between that days value from the individual’s average divided by the individual’s standard deviation. With this, HRV rMSSD z-scores greater than 0 may reflect completed recovery while values below 0 may reflect incomplete recovery. In my opinion, based on the HRV values I’ve seen, I would consider using a band of -1 to 1 for normal day-to-day variation and values above and below these thresholds as reflective of relevant recovery statuses.

This turned into a way longer post than I was anticipating, but I think it very accurately encapsulates the complexity, but also utility and value, of HRV in athlete monitoring. The most important takeaways and practical applications are 1) heart rate and HRV are most valuable when within-individual comparisons are made and, since HRV specifically, has so much between-individual variation, it is not productive to compare your value to your friend’s or family member’s; 2) heart rate and HRV reflect autonomic nervous system influence and the body’s reaction to acute stressors and acute stressors can be many different things. It is important to remember this when interpreting the values in athletes, especially collegiate athletes, because there can be so many different things that an athlete is experiencing. With that being said, heart rate and HRV can be used to inform recovery interventions and identify athletes who may be experiencing sustained elevations in sympathetic nervous system activity, but not an abolute determination what the isolated responses to training and competition are. I do trust the usefulness of floatation-restricted environmental stimulation therapy for athletes experiencing extended periods of sympathetic activity, or failing to experience complete recovery. If this fails to modulate autonomic nervous system activity and behviors in other hours of the day are being taken care of responsibly, then it would be worthwhile to consider changes to training volume/intensity in my opinion. Again, the important thing to remember is that training and competition are not the only stressors athletes are being exposed to (or exposing themselves to).

In my next post, I will show and example of retrieving R-R interval data from Polar’s API after measuring it with an H10 and processing it into HRV features.

A caveat that I personally find important: not all stress is bad. When Hans Selye coined the term “stress” in the 1930s, he was using it as a term that represented a general stimulus that elicited the general adaptation syndrome in rat (REF). When we talk about acute stressors/or acutely being in a stressed state, this is not inherently negative or detrimental to an athlete. In order to adapt and develop improvements, there need to be stressors and overreaching. However, it is whwen that stressor is not intentionally withdrawn or reduced and there are chronic elevations in sympathetic nervous system activity that negative consequences are possible. I wanted to include this disclaimer because stress has taken on a significant negative connotation when that was not necessarily the original intention and was not my intention in this post.

…sorry, one more caveat:

HRV is affected by menstrual cycle phase in eumhennoric women. During the luteal phase, HRV is decreased while heart rate is elevated. (physiology) This is particularly important for those who are using HRV to monitor “readiness” among female athletes as understanding menstrual cycle phase can provide context for what is eliciting the change in HRV. A practical example of this interpretation would be if an athlete is in an intentional overreaching phase of training and then the volume and/or intensity of training is reduced but the athlete still presents decreased HRV, knowing menstrual cycle phase (and contraceptive use, which may also affect this interpretation but I am unsure of this especially depending on the different types of contraception) helps the practitioner understand what is going on with the athlete’s response to training. If the athlete presents a significantly supressed HRV compared to her norm and is in her luteal phase, the HRV may not be an indication of poor adaptation from training but instead a result of the menstrual cycle. In contrast, if the athlete is not in her luteal phase and presenting a supressed HRV following the training block, she may need additional recovery time or interventions to help augment her adaptation.

REFERENCES

Altini, M. (2024). How Data Can Capture Recovery: The Case for Heart Rate Variability. In P. Düking & B. Sperlich (Eds.), Individualizing Training Procedures with Wearable Technology (pp. 7–19). Springer International Publishing. https://doi.org/10.1007/978-3-031-45113-3_2 Bailey, C. (2019). Longitudinal Monitoring of Athletes: Statistical Issues and Best Practices. Journal of Science in Sport and Exercise, 1(3), 217–227. https://doi.org/10.1007/s42978-019-00042-4 Buchheit, M. (2014). Monitoring training status with HR measures: Do all roads lead to Rome? Frontiers in Physiology, 5. https://doi.org/10.3389/fphys.2014.00073 Flatt, A. A., Allen, J. R., Keith, C. M., Martinez, M. W., & Esco, M. R. (2021). Season-Long Heart-Rate Variability Tracking Reveals Autonomic Imbalance in American College Football Players. International Journal of Sports Physiology and Performance, 16(12), 1834–1843. https://doi.org/10.1123/ijspp.2020-0801 Heart Rate Variability in Professional and Semi-Professional Soccer: A Scoping Review—Google Search. (n.d.). Retrieved January 12, 2025, from https://www.google.com/search?q=Heart+Rate+Variability+in+Professional+and+Semi-Professional+Soccer%3A+A+Scoping+Review&rlz=1C5GCEM_enUS1104US1104&sourceid=chrome&ie=UTF-8 Heathers, J., Quintana, D., Angus, D., Krygier, J., Kemp, A. H., & Rosnay, M. de. (2018). Water consumption as a source of error in the measurement of heart rate variability. OSF. https://doi.org/10.31219/osf.io/83exy Merrigan, J. J., Stone, J. D., Kraemer, W. J., Friend, C., Lennon, K., & Hagen, J. A. (n.d.). ANALYSIS OF SLEEP, NOCTURNAL PHYSIOLOGY, AND PHYSICAL DEMANDS OF NCAA WOMEN’S ICE HOCKEY ACROSS A CHAMPIONSHIP SEASON. Nuuttila, O.-P., Kyröläinen, H., Kokkonen, V.-P., & Uusitalo, A. (2024). Morning versus Nocturnal Heart Rate and Heart Rate Variability Responses to Intensified Training in Recreational Runners. Sports Medicine - Open, 10(1), 120. https://doi.org/10.1186/s40798-024-00779-5 (PDF) What Is Behind Changes in Resting Heart Rate and Heart Rate Variability? A Large-Scale Analysis of Longitudinal Measurements Acquired in Free-Living. (2024). ResearchGate. https://doi.org/10.3390/s21237932 Plews, D. J., Laursen, P. B., Kilding, A. E., & Buchheit, M. (2012). Heart rate variability in elite triathletes, is variation in variability the key to effective training? A case comparison. European Journal of Applied Physiology, 112(11), 3729–3741. https://doi.org/10.1007/s00421-012-2354-4 Stephenson, M. D., Thompson, A. G., Merrigan, J. J., Stone, J. D., & Hagen, J. A. (2021). Applying Heart Rate Variability to Monitor Health and Performance in Tactical Personnel: A Narrative Review. International Journal of Environmental Research and Public Health, 18(15), 8143. https://doi.org/10.3390/ijerph18158143 Thornton, H. R., Delaney, J. A., Duthie, G. M., & Dascombe, B. J. (2019). Developing Athlete Monitoring Systems in Team Sports: Data Analysis and Visualization. International Journal of Sports Physiology and Performance, 14(6), 698–705. https://doi.org/10.1123/ijspp.2018-0169