Who, what & why?
The measurement of the amount of training someone has done is called their Training Load (TL), and is important as a means of controlling training so that sessions can be created which ensure an athlete does the right amount at the right time.
However, there are a variety of measures in common use and they all have limitations – see our earlier review here.
All ithlete users will (hopefully) have noticed that a hard training session brings down their HRV score the next morning, and important review papers such as Cardiac Parasympathetic Reactivation Following Exercise: Implications for Training Prescription found HRV consistently lower after high intensity training (HIT). Now a group of French researchers, led by Damien Saboul, have set out to create a more formal method to see whether HRV recording before and after a session could provide a better insight as to the TL that session had produced.
What did they do?
Eleven well-trained distance runners participated in a study where they had to perform four different training sessions with full recovery in between. They had their HRV (RMSSD) measured immediately before, 5 minutes and then 30 minutes after each session. The runners had to sit quietly and were allowed 0.5L of water but no food between the post exercise measures.
The composition of the four sessions was as follows, the intensities refer to the individual’s maximum aerobic speed (MAS).
| Session | Warm up | Middle | Cool down |
| 1. Aerobic | 10 mins @ 70% | 34 mins @ 70% | 10 mins @ 50% |
| 2. Long intervals | 20 mins @ 70% | 3x 10 mins @ 85% | 10 mins @ 50% |
| 3. Short intervals | 20 mins @ 70% | 8x 2 mins @ 95% | 10 mins @ 50% |
| 4. All out max | 25 mins @ 70% | 6 mins @ 100% | 10 mins @ 50% |
The Authors defined a TLHRV index as follows:
TLHRV = ln (T x (Pre5 – Post5)/(Post30 – Post5) )
In other words this new formula calculates the ratio of the change in HRV before to 5 minutes after the session to the change during the first 30 minutes recovery. This is then multiplied by the duration of the session in minutes.
Bannister’s TRIMP (Training Impulse) and Foster’s session RPE were used to compare the HRV calculated training loads for each runner in each of the four sessions to commonly used measures in the field.
What did they find?
The table below from the paper contains all the key data – the HRV values, effect sizes of the changes in HRV, and data to enable calculation of the three different TL measures.
All three techniques rated the long intervals as having the largest TL, followed by the short intervals, but the Bannister method rated the aerobic session as more demanding than the all out effort whereas the Foster & HRV method both rated the all out as more demanding than the aerobic session.
As expected, the pre to post session HRV difference correlated to the session intensity whereas the post 5 to post 30 difference was not correlated either to intensity or volume for all participants. The authors point out that it did provide normalization with respect to inter-individual differences (always an issue in HRV studies).
What does it mean?
This is an interesting study that suggest that HRV has the potential to deliver a sport independent comparable training load number, a goal long sought after by sports scientists and coaches.
Best of all this is a technique which can be quite easily deployed by ithlete users – simply perform three one minute measures, just before your workout, 5 minutes after and then 30 minutes after. All three will be shown in the app dashboard and you can get to work in Excel to calculate the TL values (remember to transform the ithlete HRV numbers back to raw RMSSD = EXP(ithlete HRV/20)).
It does require further studies to validate and refine this technique and we would be pleased to hear from coaches and students who would like to do this in their sports settings.
You can find the original paper here.
I find HRV very useful.
However I think the issue to overcome in the article is “full recovery in between”. This is not going to happen in the real world. Surely it would skew the interpretation if no full recovery
Thanks for the comment. That raises a couple of interesting issues:
1. Clarification – I believe the full recovery in between was to ensure the runners were able to do these 4 demanding sessions without any cumulative impact.
2. Would the same session produce the same TLhrv number when performed fatigued as opposed to fresh? (and should it?). I think this would need to be tested in a new study.
I would be worried about the number produced if the athlete was already overreached. Plews et al showed data that indicated the day to day variation in HRV being reduced when a triathlete was in this state, and Andrew Flatt (hrvtraining.com) is also looking into the CV% of daily HRV over time. So you might think that the Pre & Post readings would be lower but also closer together, and generating a lower TL score, unless the Post5 – Post30 were proportionally reduced.
Do let us know if you test this out!
This approach has appeal since it implies that the training load is a function of your body’s ability to respond to it. It would be a nice complement to “static” measures like TSS used in cycling. For example if you see a divergence (HRVtl going lower relative to TSS) then you may be getting more fit and sand-bagging your FTP (or CP) numbers. Conversely, if the HRVtl is going up relative to TSS then over-training would be indicated.
With all these nice features, I gave this approach a go. While the calculations are straight forward, I have observed that it only works well when the body is sufficiently awake. For example I got good HRVtl results for workouts in the afternoon because of sufficient separation of the Pre, Post-5 and Post 30 measurements. However for early morning workouts, my Post-5 were less than my Pre number = NO HRVtl. The ones that did work seemed to give more insight into the quality and intensity of the work than just the TSS number, which a also scaled using the natural log.
Thanks for trying this out. I have an idea for the early morning HRV suppression (probably cortisol awakening) effect. Do a 5-10 min gentle aerobic warmup before measuring the Pre workout HRV. I suspect this will be higher than the morning wake measure and has also been proposed (Buchheit et al) as a way of ‘washing out’ morning stresses for team sports players testing HRV at their training ground.
It would be great if you can share some of the numbers you found for different kinds of workout here.
I came to this blog looking for ways to interpret hrv values to work out an individual’s ability to absorb high levels of acute training load and was thinking that an option on Ithlete to record the previous day’s TL with either ATL or possibly cumulative training load could lead to personalisation of training peaks performance manager. My thoughts were that careful monitoring of hrv would allow riders to personalise their more intense training blocks based on heir track record of coping with rapidly aquired CTL or notional ceiling for ATL accepting that this would be constantly evolving and subject to all sorts of external factors I think this is still something I’d like to explore so if there are other studies you could point me too then thanks. However this artical also raises for me the additional possibility of knowing how the combination of TL and intensity factor could produce a more predictable point at which a rest day would be required.