Detailed Topic Index

Monday, May 9, 2022

DFA a1 Review and Update Frontiers 2022

It's been an interesting 2+ years since my initial case report indicating that there was potential value of the nonlinear HRV index, DFA a1, in modulating training intensity, zone guidance and threshold delineation.  Given the rapid progress in the field, Thomas Gronwald and I put together a review and "synthesis" of exercise related DFA a1 topics which is now on the Frontiers website.  In many respects it also serves as a general FAQ with sections dealing with background, thresholds, limitations, usage with atypical exercise types, assessment of fatigue, durability, real-time use and of course, all the great apps to measure it.  There are some interesting comments and observations regarding ECG vs chest belt results, artifact effects, preprocessing/detrending (Kubios reverse engineering).

Below is the "essence" of the relationship of DFA a1 with exercise intensity - a combo figure showing the a1 to HR relation regarding training zones (A) and to defined threshold values (B) in a diverse group of users.

A few hopefully helpful comments:

  • Several requirements are need for optimal DFA a1 measurement:
    • Artifacts below 5% (3% even better)
    • Proper software that reverse engineers Kubios methodology such as Runalyze, AIendurance, Fatmaxxer but not HRV logger.  The alphaHRV app for Garmin units is in a separate class - although not using either Kubios preprocessing or measurement window type, it seems to correspond well.
    • Good signal to noise ratio of the cardiac electrical signal - this is generally not considered but is of major importance for a1 to be measured accurately.  
    • For thresholds make sure you are fresh, no illness, not over reached.  Try to use a fan indoors!
  • Does this method work in everyone?
    • No - there are always exceptions.  Even the gold standard approach of gas exchange will have uninterruptible results or even wrong determinations due to "user" error.
    • However, it is widely applicable for most individuals.
  • As I work with various participant data, I'm amazed at the frequency of rhythm disturbances - from atrial fib, very frequent APCs, intraventricular conduction delays (bundle branch blocks) and even ventricular aberrancy.  
  • I read in (fill in the name) blog or Twitter feed that DFA a1 insights regarding exercise intensity/threshold determination are based on "flimsy" evidence - what do you say to that!
    • To date, no published study has proven any of our conclusions false.  In fact another group, working with a large number of participants found essentially the same results regarding threshold values.  When reading someone's personal experience, several prerequisites should be present - the use of Kubios as the HRV interpreting software, publication in a decent peer reviewed journal and no conflicts of interest or emotional baggage.  Unfortunately, the later should not be underestimated as a motivation to discredit DFA a1.

Some final personal thoughts

Along the way, I've worked with some amazing teams on these articles and projects.  I would like to express my heartfelt thanks to all of them.  Without the data sets and expertise of people like Dave Giles and Laurent Mourot, we would not have been able to come up with these insights.  A special thanks to Thomas Gronwald for pioneering the concept of using DFA a1 for exercise modulation in general and being open for collaborating with a relative "noob" in physiology.  He is a brilliant researcher and has been an incredible friend. 


Heart rate variability during dynamic exercise