Thursday, April 14, 2022

The DFA a1 App matrix

In keeping with our upcoming review of "whats new" in DFA a1 related exercise topics, I wanted to consolidate a cross comparison listing of current apps that will display this metric either on the Web or in real time on smart devices.  I will continue to add/subtract based on features, issues and programming changes made.  Most of the apps have separate posts except for the new iOS beta app.  Below the matrix listing I will make some personal comments about each app as well.  It was interesting to go back to my first impressions of many of the app's initial agreement with Kubios (which is what we use in published studies) - they were uniformly not good.  However with cooperative developers, excellent agreement to Kubios was achievable.  Many personal thanks to those individuals (you know who you are) for being open to constructive criticism. 

 


Runalyze - An excellent web based app that does it all - will track power, muscle O2, Core body temp, HRV etc.  Great development team willing to adapt and adjust the product per user feedback.  It allows one to determine where to set the stop, start timing points for ramp analysis.  Csv download contains DFA a1, power, HR for easy Excel analysis.  Artifact percentages are displayed for each 2 minute measurement window. 

AIEndurance - Another excellent web based app.  The focus here is perhaps more DFA a1 vs power/HR related than Runalyze.  There is automatic calculation of threshold values if a progressive HR/power rise is noted during the first 30 minutes.  In addition, the app will keep track of historical trends of the a1/power relationship to determine degrees of baseline fatigue (inappropriate low a1 at warmup power) as a measure of readiness.  The app will also track the a1:power relationship during a longer session as a measure of exercise fatigue/durability.  These are handy observations to have and our group is in the process of further exploration/validation of this concept.  The development team is very willing to take suggestions to improve the product.

Fatmaxxer - Although android only, as near perfect a monitoring app that I could hope for.  Why - excellent Kubios agreement, nice graphical display of pertinent data (including artifacts), signal strength of the ECG waveform for optimal belt positioning and perhaps most important of all - ECG tracing snips pre/port each artifact detected.  That one feature alone may potentially save a life someday for detecting a serious arrhythmia.  Designed by an enthusiastic, talented developer who is interested in making a quality product that is free of charge.  Finally, the RR output file is "unfiltered" meaning that it contains the raw RR data for potential Kubios analysis. This a good tool for researchers looking to collect HRV for later Kubios import, while also getting real-time feedback on a1 and signal strength for optimal belt position.

 

Alpha One Endurance - This is an app designed by Ash Burrows and more info can be found here.  Although I do not have detailed specs at this time, the app uses Kubios preprocessing methodology and testing so far has shown excellent agreement to Kubios.  Further review and comments pending.  It is very welcomed since there is a current lack of an appropriate iOS based DFA a1 app.

alphaHRV - The app that I thought would be impossible to create (I'm happy to be wrong) - an accurate, realtime a1 display for Garmin watches and head units (without a smartphone being needed).  Although it is using an alternate method of preprocessing and a measurement window of 200 beats (instead of 2 min), the data is close to that of Kubios and updates each second (user defined).  A masterwork of programming and imagination indeed!  Artifact % is displayed and there is a 2 min running graphical display of a1 on device.  Garmin fit files will contain this data so you can extract for further analysis.  Garmin Connect treats this like any data field (with artifacts showing as well).  The development team is helpful, motivated and willing to adjust the product as needed.  Please see my review for why we can't expect precise Kubios agreement, but IMO, this is close enough for our purposes.

Fitness Dashboard - A PC based app, still in alpha but stable for every day usage.  Uses Kubios features including the proper detrending and the huge bonus of "automatic" artifact correction, not seen in any other app except Kubios Premium.  Also tracks other Ant+ devices such as Moxy, core temp, cadence etc.  Will eventually have realtime ECG for the Polar H10.

Watchletic - An app for Wear OS and Apple watch. Uses Kubios detrending and threshold correction. Watch display shows RMSSD, DFA a1 and artifacts (along with other sports metrics). Does not require the phone to be present for record and watch display!!

 

HRV logger - The first app for realtime a1 tracking.  Unfortunately, it suffers from inherent limitations in recalculation ability (stuck at 120 sec) and non optimal preprocessing that may lead to poor a1 agreement to Kubios.  The app was designed as an add on for an already mature product as opposed to Fatmaxxer, alphaHRV and Alpha One Endurance which were created solely as an a1 tracking app.  An analogy for this app is that of the first cellphones.  I remember my first "bag" phone, a real game changer for a busy doctor - I didn't have to stop at a pay phone (yes really) to answer calls after my pager went off.  So yes, the bag phone was great in 1988.  But in 2022, we can (and should) do better.  This app has not been updated and the developer has lost interest in DFA a1, making improvements in programming, interface, and Kubios agreement unlikely.  As they say, "the right tool for the right job".

To the FAQ....



Saturday, March 26, 2022

More pitfalls in HRM belt position

I was recently asked to evaluate a blog reader who had high artifact RR recordings.  This is the way it looked:

We are seeing near 5-10% artifact rates with occasionally higher values.  Potential causes include poor contact, low battery, arrhythmia, module failure or non optimal belt position.  To track down the cause, the athlete was asked to record a ramp session with software that records an ECG from the Polar H10.  

Here are the results:

What we are seeing here is a very low R peak amplitude, so low that some beats are not even counted properly (yellow marks).

Since the R peak has a low signal to noise ratio, the DFA a1 can't be properly determined.  Recall that this index is based on correlation patterns of beat to beat timing - if the timing is not precise, the correlations will be lost.  Kind of like looking through a dirty camera lens or set of eyeglasses and not recognizing a persons face.

Here is the DFA a1 result-

  • The bottom panel (black) is the DFA a1, which is all over the map, no relation to heart rate (blue). The top panel shows the ECG and artifacts (tiny vertical lines) which are above 5%.


How do we fix this?

  • Just invert the signal polarity in the preference settings of Kubios.  

This is what the result yields:

 
 

  • Now DFA a1 is gradually dropping as HR rises.
  • Notice also that all the artifacts are gone - not a single one is present! 
  • The ECG strip on the top panel shows good R peaks. 
  • Excellent regression plot of HRVT.

Unfortunately, the H10 does not auto invert the signal.

To compensate, this user should flip the module (Polar logo upside down) to reproduce the software polarity inversion I did.

More info from Polar (even they recommend these mods):


 

Bottom line:

  • Some users of chest belt HR monitors will need to either invert the module or move it to the left to optimize the signal.
  • Usage of Fatmaxxer (Android) or one of the Polar H10 ECG logging apps (ios) can help one perform this type of assessment.
  • Although I can't say how common this type of issue is, it probably occurs more often than we think.
  • Going forward in DFA a1 monitoring, I recommend an initial signal optimization protocol to help produce the most precise a1 values. Besides the issue of proper software detrending (as Runalyze, Fatmaxxer and AI endurance are doing), signal optimization may help you reproduce published findings.





Friday, February 18, 2022

HRVT validation in Elite Triathletes

Up until now, there have been 3 studies examining the relationship of the aerobic threshold exercise intensity with that of where DFA a1 crosses 0.75 - 

  1. Our initial report in Frontiers (runners), 
  2. one in the Journal of Clinical Medicine (cardiac patients) 
  3. and another in runners (however they used different methodology).  

Therefore, although many of us are using the index during cycling activities, there are no published studies validating use DFA a1 with this particular exercise type (yet).  Today, the journal Sports has published our article - An Index of Non-Linear HRV as a Proxy of the Aerobic Threshold Based on Blood Lactate Concentration in Elite Triathletes. This report did use a cycling model and will be briefly discussed below.


 

The article is a relatively straightforward read and should be a review for most users of the index.  However, instead of using gas exchange to obtain the AeT (VT1, GET), a lactate based protocol was used.  This was the "normal operating procedure" during the training camp, for the team we followed.  As I discussed previously, lactate threshold determination is not universally agreed upon - therefore, results could have been different with variations of threshold concepts.  

As usual, many thanks to co authors Sander Berk and Thomas Gronwald and the athletes participating in the training camp.

Study Results

The bottom line of the study was that yes, the LT1 (via log-log method) agreed closely with that of the HRVT in a group of elite triathletes (2 of them women).  Although the match in each individual was not perfect, it was in line with other studies that looked at the agreement between the VT1 with the LT1.

As part of my habit of adding a little bonus material to the published work, the following is a set of plots detailing the HRVT and lactate trajectories of one participant.

This was a female triathlete, testing protocol as per the above article.

LT1 from log-log plotting (we actually used the formula from Newell et al, but the results were identical to this graph):


  •  The HR at the log log threshold is 168 bpm.

The HRVT plot (DFA vs HR):

  • Excellent agreement in this individual for LT1 and HRVT.

Analysis of power at LT1 and HRVT power were done with different methodology (see article if you are interested).  

Summary:

  • This study represents the fourth "validation" of the HRVT concept but in elite triathletes performing cycling exercise.
  • The regression metric and limits of agreement were in line with other published data comparing various threshold methodologies.
  • No test method is perfect therefore YMMV, but the majority of users should be able to reproduce results similar to this.
  • Shortly after the above was published, the following was released - Validity of DFA a1 to determine intensity thresholds in professional cyclists with similar results


Heart rate variability during dynamic exercise