Tuesday, March 23, 2021

DFA a1 stability over longer exercise times

In the last post we saw how DFA a1 can be a marker of "fatigue" after intense short bouts of effort.  Since it represents a marker of "organismic demand" and autonomic balance, typical values for a given level of power will be altered post HIT.  On the other hand, what about behavior over the course of longer durations of mild to moderate intensity?  With this question in mind, we are going to examine two examples in this post looking at DFA a1 behavior over the course of a 2 hour easy pace cycling ride and one just below the AeT.  The questions are:

  • Does DFA a1 "drift" during a constant power effort below the AeT (aerobic threshold) over 1 to 2 hours?
  • Does the HRVT (AeT related DFA a1 threshold) change after 1 to 2 hours of "just below" AeT cycling?


The first example is from a friend who is a former (very) elite athlete who is still active in cycling and running sports.  His usual AeT by HRVT was about 210w.  He did a double ramp test, each separated by 2 hours of easy riding at 140 watts (2w/kg). The ramp consisted of 5 x 6min at 100w, 130w, 160w, 190w and 220w. He used a Polar H10 with nearly no artifacts. 

Here is how the entire session looked up to the second ramp:

Ramp comparison:

Because the ramps were not the usual incremental type I use, I decided to simply superimpose the ramps on top of each other.  In addition, this athlete had improved their fitness and instead of the usual HRVT being about 210-220w, it is now higher, and not totally captured on the tracing.

DFA a1 on Pre and Post ramp:

Heart rate on each ramp:


  • There was no significant alteration in HRVT dynamics from Pre to Post 2 hour cycling ramps.  
  • There was noticeable HR "drift".
  • DFA a1 during the 2 hours was in the highly correlated range (above 1), consistent with very easy effort.
  • From this example it appears that the 2 hours of easy cycling did not impact either the HRVT nor the ability of DFA a1 to gauge intensity status.
  • DFA a1 stability was in the face of some cardiac drift.


Example 2

The next set of data comes from me.  I'm about 25 years older and recreational level - so a good cross section comparison.  What I did was a 20 min warmup, then first ramp (incremental, 5w/min, good cooling, Polar H10 with <2% artifact, Kubios time varying analysis). Immediately after the first ramp, I cycled at just under my AeT (5-15w) for one hour.  This was about 30 watts above my typical "easy" days, but I was curious to see if pushing the limits would be an issue on the second ramp.  

After the hour, a second ramp was done, identical method to the first.


  • DFA a1 was both above and below the .75 for the hour of "just below" AeT riding.
  • There did not appear to be major cardiac drift (red is HR) despite stable power.
  • Histogram of DFA a1 during that 1 hour at just under AeT is below
  • Although some of the time was spent below .75, the bulk was not.
  • This is important to note as some users of this method have worried about variation in the DFA a1 during a relatively stable power session - the fact is variation is normal, but the average was well above .75.  The official mean was .89 in this example.


Ramp comparisons (Pre and Post the 1 hour):

Ramp overlays (as I did above):

DFA a1:

  • The early parts of the ramps were a bit different since the second was preceded by the "just under" AeT 1 hour, but the DFA a1 later on looks remarkably similar.

Was there any HR drift?

  • Except for the first few minutes (as noted), there was no noticeable cardiac drift.


HRVT in each ramp:

Heart rate at HRVT:

Power at HRVT (calculated from the time:power relationship)


  • Both the HR and power at HRVT were about the same during each ramp before and after 1 hour of cycling just under the AeT. 
  • This was even in the face of some of that hour session having DFA a1 dip below .75 on occasion.
  • There was no cardiac drift (good fan and cool room).


  • It appear that 2 hours of continuous cycling well under the AeT does not lead to a change in DFA a1 behavior and does not alter the HRVT.
  • It also seems that 1 hour of continuous cycling just under the AeT does not alter the HRVT or the DFA a1 relationship to power.
  • DFA a1 behavior appears to be independent of HR drift (first example had it, second did not).
  • On a practical basis, one can be reassured that DFA a1 values over longer time spans still accurately reflect the exercise intensity as long as HIT was not performed. Even prolonged cycling at an intensity a few watts below the AeT did not seem to affect the HRVT after 1 hour.
  • Of course, longer time frames (>2 hours) may entail a very different set of results.  
  • Finally, for enforcement of a low intensity session, real time monitoring of DFA a1 appears to be a promising aide to endurance exercise training.


  1. Fascinating, this is an excellent analysis. I do wonder though do the DFA a1 values vary significantly on a day to day basis due to differing external factors, emotional stress, sleep deprivation, caffeine intake etc?

  2. Great question. There's nothing currently published on reproducibility of a1 threshold data. Yes, many things theoretically can affect it, the issue will be by how much.

  3. Enjoyed that Bruce, interesting to see you making the point ref occasional dips below 0.75 not really affecting the purpose/validity of the ride.

  4. Great to see this as I have wondered about stability of DFA1 overtime. One comment though on the cardiac drift in ex1, do you think that DFA1 relationship to AeT remains the same during longer endurance events? Put another way, if you measured lactate across this entire test would the lactate still remain low during the 2nd test when the cardiac drift is observed? It is often said that one should be able to maintain an effort below LT1 for hours w/o cardiac drift. This example, albeit with an n=1, might suggest that seeing cardiac drift doesn't necessarily mean one has entered zone 2 (3 Z model).

    1. As I mentioned in my recent podcast, cardiac drift can occur independently from any a1 change.
      Causative factors overlap but not totally. HR drift is usually from heat/skin temp. If those are well controlled, you may not see it, but still see a1 decline post HIT.
      As to the question - do you think that DFA1 relationship to AeT remains the same during longer endurance events - depends on how long. I doubt it would remain the same after a 6 hour run at slow pace.