Assessing fitness and detecting over-training: the running version of the sub-maximal test

The art of effective training is maintaining a training load sufficient to produce steady improvement while avoiding the damaging effects of over-training.   In my preceding post, I described a test designed to help achieve these goals by assessing improvement in aerobic fitness while providing a sensitive indication of the early signs of over-training.  The first key principle in the design of the test was measurement of heart rate (HR) and heart rate variability (HRV) under various circumstances, thereby assessing multiple aspects the way that the autonomic nervous system regulates the heart, and avoiding dependency on a single potentially erratic measurement.  The second design principle was that the test should fit within the framework of a warm-up preceding training, making it practical to perform it frequently within only minor interference with a training program.

At the time I was doing much of my training on the elliptical cross trainer so I designed the first version of the test for use on the elliptical.  The protocol started with measurement of HR and HRV while standing in a relaxed state.  Then after very brief warm-up to allow initial stabilization of the cardio-vascular system, HR was measured at three different levels of work intensity: easy, mild and moderate; followed by an assessment of heart rate recovery (HRR) during a minute of standing and finally HR was assessed again in the final minute of a short easy-paced recovery.  The evidence accumulated during a 10 week period of aerobic base-building indicated that the first version of the test provide a sensitive measure of improved aerobic fitness, while also quantifying stress levels during a period when life-events were challenging, and in addtion it proved also to be a sensitive indicator of impending over-training.

The consistency of beats/Km

More recently I have developed a version of the test for use while running.  The structure of the test is similar, but the biggest practical challenge to overcome was that fact that it is less easy to maintain a constant work rate over a period of several minutes when running compared with exercising on an elliptical cross-trainer with a power meter that allows regular monitoring of work-rate.  The key thing that provides the solution to this problem is the observation that the number of heart beats per Km changes only slowly with changing effort across the aerobic range.  Therefore, absolute stability of pace is not essential if the measurement of interest is heart beats per Km at each work-level.  Furthermore, it is not essential to achieve the same pace at each of the three work levels  on different occasions.  Small discrepancies in pace from session to session have a minimal effect on the measured beats per Km at each level, and larger discrepancies can easily be dealt with by simple correction.

The course

I do not have access to a track, so I developed the test on a 230m  loop of woodland path – slightly undulating and laced with treacherous tree roots, but on the other hand, an inspiring environment in which to run.  At each of the three work-rates, I do three laps at approximately constant pace.  During the first two laps, HR stabilises.   Heart rate in the final lap at each level is used to estimate beats/Km at that level.    The final lap should cover the same terrain on each occasion, and preferably should be a closed loop so that variations in wind speed from day to day have little effect.  If you have access to a track, 2×400 m would probably work as well.  The first 400 m lap allows stabilization.  Beats/Km in the second lap is the measurement of interest.

Easy, mild and moderate levels

The easy, mild and moderate work levels correspond to 75%, 82% and 88% of maximum heart rate.   The highest level (88%) is near the top of the aerobic range.  The subjective effort required to maintain this pace increases quite markedly when over-training occurs, so for the purpose of early identification of over-training, it is desirable to have a level as high as this, though 3-4 minutes at this pace is perhaps a little more demanding than an ideal warm-up.  Perhaps this level could be omitted on days when you want to perform the test before a demanding training session, though I find the full version of the test works well as a warm up for moderately intense interval sessions.   The HR measured during the final lap of the recovery  (performed at a subjective effort similar to that of the easy first phase of the test) reveals a substantial increase in beats/Km at that effort, demonstrating substantial mobilization of adrenaline during the test.

Target HR or pace

At least on irregular course, it is more practical to aim for a particular HR than a particular pace, though it is necessary to allow for the fact that it takes a few minutes to reach a near equilibrium state.  If you have a foot pod that provides a continuos display of pace, using it to maintain a constant pace might be preferable.   Nonetheless, I find that at least at mild and moderate effort levels I can maintain pace such that time per lap is constant to within less than 2% without using the foot pod.

Figure 1 shows the time course of heart rate during the full test.

Figure 1: The tine course of beat by beat variation in heart rate during the sub-maximal test.   The black event markers mark the end of each lap.  The horizontal bars beneath the R-R trace for the final lap at each intensity level indicate the time period over which average HR is calculated to provide an estimate of beats/Km at each of the three intensity levels.

Figure 1: The tine course of beat by beat variation in heart rate during the sub-maximal test. The black event markers mark the end of each lap. The horizontal bars beneath the R-R trace for the final lap at each intensity level indicate the time period over which average HR is calculated to provide an estimate of beats/Km at each of the three intensity levels.

Figure 2 shows the beats/Km in the final lap at each of the three effort levels on tests performed prior to an interval session in three consecutive weeks of my current phase of training, in which I have introduced interval training after a phase of low intensity, moderately high volume base-building.   On all three occasions, subjective effort during the final lap at the third level was either 11 or 12/20.  During this time I have not experienced over-training, though past experience indicates that effort level would rise to 14 or 15/20 at this pace during the early phase of over-training.

Figure 2: beats/Km at each of the three test levels and during the final lap of the easy recovery run, recorded on three consecutive weeks during which a marked improvement in fitness was observed.

Figure 2: beats/Km at each of the three test levels and during the final lap of the easy recovery run, recorded on three consecutive weeks during which a marked improvement in fitness was observed.

The amount of improvement over the three weeks is surprising, though it did correspond to a marked improvement in my fitness observed during other sessions in the same weeks.   I will describe my current training program in greater detail in a subsequent post.  Note that beats/Km decreases gradually as effort level increases.  From the slope of the line a correction factor can be estimated if the intensity at any of the levels on a particular occasion is appreciable different from the intended level.

Standing HR, HRV and HRR

The other measurements of interest (standing HR and high frequency HRV during standing at the beginning , and heart rate recovery while standing for one minute after the final lap at the highest intensity level) are performed as in the elliptical version of the test described previously.  Interestingly, in the three weeks depicted in figure 2, initial standing HR was 52, 57 and 52 beats/min while the three values of RMSSD (a measure of high frequency HRV that reflects ‘restful’ parasympathetic activity) were 51, 30 and 57 milliseconds.  Thus both standing HR and high frequency HRV indicated reduced parasympathetic activity in the second week, possibly indicative of a somewhat increased stress level.  In general, high standing HR and low high frequency HRV provide a sensitive indication of general level of stress.  But neither standing HR nor HRV provide a consistent and sensitive measurement of improved fitness because they are too sensitive to various sources of stress.

The interpretation of HRR is complex, though the earliest phase (in the first 30-40 sec) is largely determined by the ability to re-engage the parasympathetic system, which is a good indicator of healthy cardiac regulation.  One feature I attend to closely is the development of strong beat to beat variation within 45 sec of stopping running (illustrated as the fuzziness of the trace in figure 1) confirming strong parasympathetic re-engagement.  Although I have not experienced an instance of over-training since I began the development of the running version of the test, the preliminary evidence acquired with the elliptical version revealed low HRR during over-training.  Furthermore, during the over-trained state, HRR was low despite moderately large high frequency HRV during the initial period of relaxed standing, suggesting that high initial parasympathetic activity but reduced ability to re-engage the parasympathetic system might be a marker for over-training, but this is only a tentative interpretation.

About these ads

8 Responses to “Assessing fitness and detecting over-training: the running version of the sub-maximal test”

  1. Ewen Says:

    Canute, thanks. That looks like a good test to predict the onset of overtraining. Coincidentally, I’ve just read the chapter on overtraining in Phil Maffetone’s book. He mentions 3 stages, where stages 1 & 2 can be recovered from fairly quickly if stress (from training or other sources) can be reduced immediately. Stage 3 can take months or even years to recover from!

    He says that “one key to avoiding the overtraining syndrome is biofeedback, including measuring the resting heart rate, training at your maximum aerobic heart rate, and the MAF test.” (He also mentions HRV as a way of monitoring overtraining). I agree with you that h/beats per km in narrow effort levels is an easier way to measure results than the classic MAF test – 5 miles on the track at an exact HR & measuring changes in pace for each mile. Maffetone also says that “The interesting phenomenon of a sudden increased performance being associated with the early stage of overtraining is due to the fact that overactivity of the sympathetic system temporarily improves muscle strength and raises blood sugar, but it’s done at the expense of overall fitness and health.” Also that “chronic overtraining in the 3rd stage is associated with a lower resting heart rate due to the parasympathetic effects accompanied by the loss of normal sympathetic function.”

    In Figure 2, regarding the improvement of h/beats per km at the three intensity levels, do you put that down to the ‘efficiency’ of faster running? I’m presuming it’s because at higher speeds, more force is applied to the ground, generating more elastic recoil for the next ‘flight’ stage of the stride.

    • canute1 Says:

      Ewen
      Thanks for providing those details from Phil Maffetone. What Maffetone describes as stage 1 of over-training corresponds to the state of increased adrenalin that produces the pattern of increased HR shown in fig 4 of my post on 25th June, while Maffetone’s stage 2 corresponds to the pattern shown in fig 3 of that post. As I indicated, that state resolved after 2 days rest. If one ignored the warning sign and continued demanding training, there would be a high risk of suffering the enduring damage that corresponds to Maffetones’s stage 3. The various measurements that Maffetone recommends potentially provide similar information to that provided by my test . However, unlike Maffetone’s tests, my test was designed to that it could be incorporated into a warm-up prior to training with minimal disruption making it easy to accommodate within a training program. At times when you are pushing the limits of training volume or intensity, I would recommend performing the test weekly. It does require some calculations which appear a little daunting at first, though with the aid of a computer spread sheet such as excel, the calculations are quite easy once you are familiar with them, and they are actually quite interesting for anyone who likes playing with training data.

      You raise an interesting point regarding the improvement of beats/Km at higher intensity. In part this is due to increased mechanical efficiency as you suggest, but it should be noted that the contribution from cardiac output required to serve baseline metabolism makes an increasing relative contribution at low speeds. If you moved at a snail’s pace, HR would be scarcely greater than baseline but it would take hours to cover a Km and beats/Km would be large. In addition I believe that the increased metabolic efficiency of force generation in muscle fibres also contributes to the decrease in beats/Km up to the moderate intensity range depicted. Beyond that point metabolic efficiency of force generation decreases, stress level rises and beats/Km begins to rise.

  2. Robert Osfield Says:

    Your last two post are fascinating exploration of how we might derive objective measures of training response. I dabbled a bit last year with by recording the calories reported by my HR monitor for each run and then normalizing these for HR drift and hills to give an a normalized efficiency (calories per mile) for each run.

    The results did show training effects with my efficiency improvements as training progressed, but also picked out times of stress, either mental or physicall stress (such as catching colds) where my efficiency dropped significantly. The results I found were broadly similar to your own with calories per mile being similar to HR beats per km. I also observed that my efficiency for runs at moderate to tempo pace were typically my most efficiency, with recovery runs and faster that LT runs being less efficient.

    These approaches do all require some work on part to process all the data we record which takes it out of what the average runner will be prepared to do. To take this type of approach out to the masses I feel that mobile phone apps may well be the critical element to make it easier to monitor.

    One element I haven’t been able to test is HRV as my HR monitor doesn’t report this. What devices do you use to record it? I do wonder if bluetooth enabled HR monitor linked to a phone with an appropriate app might be able to record HR and the HRV, I really don’t know the details of recording HRV so don’t know yet whether this might be possible. Do you have any thoughts on this possibility?

    • canute1 Says:

      Robert,

      Thanks. It is interesting to hear that your own observations are largely consistent with what I observed.

      I obtain the beat by beat (R-R record) required to estimate HRV from my Polar RS800CX heart rate monitor. It is possible to produce an iPhone app that processes the signal transmitted from a Polar HR chest strap. Simon Wegerif has produced such an app for his Ithlete device, though this requires a custom ECG receiver. Alternatively, a Cardiosport Blue Chest Strap Heart Rate Monitor can transmit heart beat information directly to an iPhone using Bluetooth Smart Technology without the need for the ithlete ECG receiver. The Ithlete is programmed to provide an assessment of high frequency HRV at rest. My own view is that the multivariate test I have developed is potentially more informative because it is not dependent of a single variable that is sensitive to minor fluctuations in stress level. In principle, it would be possible to produce an app which processed the multivariate data from my test, though there needs to be a way of inputting the event markers required to estimate beats/Km. One might use a foot pod to produce a signal indicating distance travelled. Alternatively, it might be feasible to employ GPS data. I am not familiar with the consistency of GPS data. It would need to be consistent to within a few metres.

  3. EternalFury Says:

    This protocol needs a little calibration maybe.

    Here is how I went about it:

    1) I calculated my 75%, 82% and 88% of HRmax.
    2) I looked back over the past few weeks and recorded my typical pace at 75%, 82% and 88% of HRmax, P1, P2 and P3.
    3) I defined 3 pace ranges: P1 +/- 5 seconds, P2 +/- 5 seconds and P3 +/- 5 seconds.
    4) I made sure there was at least 30 seconds between P1 and P2, and P2 and P3.

    Now, HR being a dependent variable of pace, I would rather not use HR to set my work rate corresponding to the easy, mild and moderate levels.
    HR can be finicky and what the pace I can hold at 88% of HRmax can vary quite a bit, even though the amount of mechanical work produced remains the same.

    As for measuring HRR after the moderate level at pace P3 +/- 5 seconds, I settled on recording standing HR for 2 consecutive minutes. (http://www.cardiology.org/recentpapers/AJCHRR.pdf)

    I will test this approach tomorrow at the track.

    • canute1 Says:

      EF
      That is interesting.

      Your protocol sounds reasonable though I would allow at least 60 sec (and perhaps even a bit longer) for HR to stabilise at a new pace; even longer if there has been a large step from the preceding HR. As for HRR. I think that will measure something different from what the first 60 secs measures. The change during the first minute is largely dominated by re-establishment of parasympathetic influence. At present I mainly use visual inspection of the re-establishment of large amplitude HRV within the first minute to interpret what is going on. I am still exploring the most useful interpretation of HRR.

      The question of whether or not it is better to aim for a target pace or a target HR is one that I am still pondering, though on balance, I think that a target pace (or power output) is preferable because when racing, it is pace that matters. For a distance runner, the goal of aerobic conditioning is reducing beats/Km at race pace.

      On the elliptical, it is easy to monitor power output continuously, so I opted for target power output in the elliptical version, whereas I had developed the running version on an irregular woodland path and it was more feasible to aim for target HR. When beats/Km (or beats per watt-hour on the elliptical) is the measure of interest, it is reasonably easy (though a little tedious) to make the small correction to allow for failing to achieve the correct target pace using the plot of beats/Km v pace. When one is monitoring increase in fitness over a period of several months, if you aim for a particular target HR, pace at that HR will increase systematically, so each level will be at a faster pace.

      The question of whether beats/Km at a particular pace or at a particular HR the variable of greatest value is assessing fitness is made even more debateable because HRmax decreases as you get fitter (and also as you get older).

  4. EternalFury Says:

    Back from the track, after my first attempt to implement the test.

    First, my watch set-up was incomplete, so there was no HRR collection this time.

    Second, using GPS around the track for pacing won’t work. I kind of knew this already, but I had to make sure it was the case with largely submaximal paces.
    So, I will need to rely on my foot pod for pacing around the track. I calibrated the foot pod on the track today.

    Third, paces I had estimated, P1, P2, P3 was not right. P1 was about right to reach 75% of HRmax. P2 was not fast enough to reach 82% of HRmax. P3 was not fast enough to reach 88% of HRmax.
    I will adjust P1, P2, P3 for next time.

    Now, to clarify a few things…

    1) Each level is 2 laps (2 x 400m) around the track. The first lap is aimed at reaching the target pace while stabilizing HR.
    2) As you indicated, I only intend to use the second lap of each level for reference.
    3) When I said I would measure HRR for 2 consecutives minutes, I meant exactly that. I will have average HR over the first minute and I will have average HR over the second minute as well. I can use the first average as a measure of the “re-establishment of parasympathetic influence”, as you indicated. And I can use the second average as a further measure of HRR. (Most of the HR recovery should happen in the first minute. I can compute the % of HR recovery that happens during the first minute and also use this as an index. Example: If HR at the end of the moderate level is 155, if average HR over the first minute is 120, if average HR over the second minute is 115, then 87.5% of the HR recovery happening over the first 2 minutes happens during the first minute. That percentage 87.5% should be a good index of HRR.)

    I will update you when I am done with this initial calibration of the test.

    • canute1 Says:

      EF,
      Thanks for the update. I have little experience with GPS – because I often run in woodland I suspect it would have limited accuracy for me, and it sounds as if it is not a great deal more precise on the track either. I find that a Polar S3 stride sensor measures distance with an accuracy of about 1% which should be fine for the test.

      It might be that you had mis-estimated the pace required to reach desired HR because you had used data recorded during a longer session to estimate the relationship between pace and HR. Within the 10-15 minute duration of the test, HR probably only reaches approximate stability, at least at the higher effort levels where slow upward drift is anticpated. If one trained hard for 30 minute before the test, I suspect HR values at a given pace would be higher, so the test should always be done during warm up. Nonetheless, when developing the test I did demonstrate that beats per Km at an easy pace (corresponding to 75% HRmax) was very similar after a total of 8Km run at that easy pace, as it had been after 4 minutes of very slow jogging and 3 minutes at the easy pace, so there is little there is virtually slow upward drift of HR at that pace after the initial stabilization period.

      With regard to the clarifications you listed, my comments are in capitals:
      1) Each level is 2 laps (2 x 400m) around the track. The first lap is aimed at reaching the target pace while stabilizing HR. YES
      2) As you indicated, I only intend to use the second lap of each level for reference. YES
      3) When I said I would measure HRR for 2 consecutives minutes, I meant exactly that. I will have average HR over the first minute and I will have average HR over the second minute as well. I can use the first average as a measure of the “re-establishment of parasympathetic influence”, as you indicated. And I can use the second average as a further measure of HRR. (Most of the HR recovery should happen in the first minute. I can compute the % of HR recovery that happens during the first minute and also use this as an index. Example: If HR at the end of the moderate level is 155, if average HR over the first minute is 120, if average HR over the second minute is 115, then 87.5% of the HR recovery happening over the first 2 minutes happens during the first minute. That percentage 87.5% should be a good index of HRR.) THAT IS INTERESTING. IT IS A BIT DIFFERENT FROM MY MEASURMENT OF CHANGE FROM BEGINNING TO END OF THE FIRST MINUTE. IT SHOULD BE MORE RELIBALE THAN MY MEASURMENT, AND IT IS POTNETILLY A GOOD MEASURE OF THE RE-ESTABLISMENT OF PARASYMPATHETIC INFLUENCE.

      I look forward to the next update.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


Follow

Get every new post delivered to your Inbox.

Join 72 other followers

%d bloggers like this: