Saturday, March 20, 2021

DFA a1 behavior and HIT, a potential marker of fatigue

We have touched on the various external parameters that can affect DFA a1 (skin temp, food, exercise intensity) but have not formally looked at the effects of fatigue.  We are currently working on a study and draft manuscript on that subject.  However, I decided to do a N=1 look at what happens to the HRVT (a1=.75 on a ramp) before and after a combo HIT session.  The importance of this is twofold:

  • Does the HRVT change post HIT?  If it does, we need to make sure our warmup is mild, and not to do a ramp after a significant effort. In some tests (VO2 priming, fast start, FTP warmup) a brief HIT interval is actually recommended.
  • Can the DFA a1 be used as a marker of fatigue?  In other words, can we leverage the usual relationship between power and DFA a1 to gauge fatigue, and perhaps make sure it comes back to "correlated" values (about 1) before doing another HIT interval to get the best interval quality possible?

The protocol (total 2 hours):

  • Warmup 20 min
  • Ramp 20 min (5w/min from 130 to 230w)
  • Active recovery about 10 min at below VT1
  • 3 minutes at 110% VO2 max power (360w)
  • Active recovery about 20 min below VT1
  • Wingate 60s max (500w/min)
  • Active recovery below VT1
  • Repeat 20 min ramp as above
  • There were no temp elevations nor hydration effects - done indoors, AC on high, very good fan, no sweat on the floor and plenty to drink. 
  • Movesense ECG with sample rate 500Hz worn high
  • Kubios time varying output
  • The DFA a1 in black gradually drops at each ramp as expected, but plummets during the 3 min and 1 min HIT intervals.
  • There is a general decline in DFA a1 after the HIT even at power well below VT1 with values in the .75 realm.
  • Peak HR at the end of each ramp is the same, despite beginning HR much higher in the post HIT ramp.

Ramp comparisons:

Pre vs Post HRVT HR

Does the heart rate at which DFA a1 cross .75 change with fatigue?

  • The HR does not really change to a major degree!
  • Part of the reason could have been that I had eaten before doing the session (to avoid bonking) which may have elevated the Pre HR somewhat.
  • However, it is interesting to note that HR doesn't seem to help us much here.  Yes, there can be some cardiac drift, but that is usually caused by alteration of the HR x stroke volume relation with a stable cardiac output, not a definitive fatigue marker.

Pre vs Post Power:

  • This is the interesting outcome...
  • The power at HRVT Pre was 205w vs 170 for Post HIT.  The 205w is close to my usual ramp (a1=.75) value that I've repeated so many times.  The 170w agrees with what we see after the HIT intervals when doing active recovery - Despite staying 30w below VT1, DFA a1 stays at the threshold value.

Benefit of having the ECG

I was concerned about the possibility of pushing too hard in this session and inducing a cardiac arrhythmia so wore my Movesense ECG module.  Fortunately, there were no significant arrhythmia.  Here is an example of a single APC on the second ramp (after HIT):

 As a side note, the ECG waveform was excellent, here is a sample during the Wingate 60s:


Summary and Comments

  • HIT intervals will affect the behavior of DFA a1 during both a ramp and at constant power.  
  • One can look at this as a fatigue effect as well as a cautionary sign.
  • A post HIT ramp will not yield a valid HRVT (surrogate VT1).
  • Post HIT DFA a1 suppression seems to serve as a biomarker of a fatigued state.  
  • Real time observation of DFA a1 and recognition of when it resumes normal patterns for a given power may help in optimizing the quality of the next HIT interval. 
  • Next test to be done - a Pre Post double ramp as above, but just riding below the aerobic threshold for 1 hour without any HIT - will the HRVT remain stable?
  • Part 2 -
    DFA a1 stability over longer exercise times


  1. Hi Doctor, I tried to do an indoor bike workout by monitoring live a1 with HRV Logger and H10. I did a long and good warm-up to 1> 0.80, then 30 minutes of challenging ascent (Montirolo with positive peaks of 14.8%). later on a velodrome, to have a flat course, I did some HIIT exercises according to a1. I pushed up to a1 <0.40 and rested up to a1> = 1. My impression (however, this is the first time that I do it) is that a1 can be an excellent parameter to manage recoveries and restarts in interval exercises. Let me explain better, when I left I really had a feeling of fatigue (RPE) of recovery, but HR was still high due to the drift. If I had followed the beat I would have started much later, but I managed to do several repetitions even though I already started from a state of fatigue, due to the previous exercise, I believe that the management of recoveries made in this way is optimal. But I would like to have your opinion on what I have observed, in case a suggestion on what to investigate in the next exercise. Thank you

  2. Excellent observation and I agree. Although we don't have anything proven, following a1 post HIT, and waiting for it to "normalize" before another hard interval, may allow you do get a better quality of the high intensity interval. HR drift does not seem to be as helpful. In my opinion, getting good quality of the HIT interval is more important than just doing many low quality attempts.

  3. Try recording from a muscle oxygen sensor at the same time? Does dfa alpha 1 return as the muscle oxygen values return?