I've been looking into this more and more. Here is a good blog post;

http://www.muscleoxygentraining.com/2021/01/dfa-a1-and-exercise-intensity-faq.html

thanks. It is an intriguing option for a training marker and way to estimate your AT.

I have tried it out for a few treadmill runs and steady indoor bike workouts using a polar h10 strap and the hrvlogger app.

So far I have found that for me training above the .75 threshold results in a lot higher HR and power than what I thought was Zn2 for riding. For running it results in a lot lower hr and much slower pace than my usual zn2 effort.

---

______________________________________________________________

Um, not to be Debbie Downer or anything, but do we really need a proxy for VT1 and VT2? Wouldn't the talk test be sufficient?

actually that is the point of this. He goes into the reason in the papers he has published that is is a surrogate for the lab based invasive testing for aerobic threshold. To get that number one has to do an invasive test and not everyone has easy access to the proper and skilled testing required for lactate testing and other forms of such lab testing.

So I think the Ursala upward answer is YES!

I’ve been using it on the indoor trainer only (no treadmill) and have found the same as you. I had thought my LT1 hr was around 140-142. Using HRV logger and a polar h10 with fresh batteries I doing get below .75 until the HR is about 148-150.

I didn’t do the incremental ramp test though. I just set the trainer at 200 watts for a ride (which tends to be around what I though was my LT1 HR of 142) and was at close to 1.0 for a couple readings (~10 minutes) then bumping it up slowly for a few 2 minute blocks. Eventually started getting close to .75 at 230ish watts and a HR of about 148. Probably need a bunch more rides to get a better handle on it.

Would love to find someone with a lactate meter to see if my values correlate.

---

USAT Level 1 Coach
Team Next Level
http://goteamnltri.com/

it is interesting that you have been using it. Have you used Altini's app? According to what he and Rogers say you should go one point past the break through point of 0.75 to assure that the change is consistent. Also they suggest durations of 5-8 minutes? to allow for sufficient data collection. The stepped approach still suggests going above the first point where 0.75 occurs to ensure consistency. Of course they have validated in their studies against normal lactate testing but of course that may not mean it is true for you or me. I am interested because if it works it may actually be a better metric than something like FTP because it should work under a variety of submaximal conditions?

Have used the app. It’s called HRV Logger. Only a couple bucks. I probably should do the prescribed ramp test which I now have a decent idea of where to start based on general erg trainer rides. I’d probably start around 210 after a decent warmup and then ramp up 10 watts every 6 or 8 minutes. The app gives you a reading for every two minutes so your steps absolutely need to be multiples of two. They say first two minutes isn’t “steady state” as you adjust to effort, so use the 4 and 6 minute reading and then step up again. Or do 8 minutes and get an extra data point. I have an endurance ride scheduled for tomorrow morning, maybe I’ll give the protocol a shot and see where I end up.
If the numbers I’ve been getting are good and accurate, then the top of my Z1 ( in a 3 zone model) is quite a bit higher than I thought and if it’s that way for a lot of people, then i think a lot of people need to readjust their idea of polarized and a z1 ride isn’t just tooling around nice and easy, it’s got a decent effort to it and especially if you stretch those rides out to a couple hours.

I’m a data geek and absolutely love this stuff

---

USAT Level 1 Coach
Team Next Level
http://goteamnltri.com/

I look forward to your insights. I am also a data geek and was really interested in how this might help with training.

Fascinating. I'd seen some of the earlier studies on using HRV as an indicator to determine VT1, but this looks like a cleaner way to do it. There might be enough information in those papers to re-implement the signal processing and analysis in GoldenCheetah, though I doubt I have the time for it.

As I read it, there doesn't seem to be any work done on doing a repeat or intervention study to track changes in VT1 using DFA a1 over time. It seems like a logical next step, but I can't find anyone with data yet. More than that, you'd have to see if there's really much of an accuracy gain rather than using a fixed percentage of HR or power versus an athlete's values at threshold/CP/FTP/MLSS, or maybe something else you could dig out of a mean maximal power curve. Maybe the ratio between ~1 hour power and ~2.5 or ~4 hour power? Whether or not knowing this value to that degree of precision means anything and if/how to build a training plan around it is another argument entirely, and one I don't want to touch.

At least the testing is simple, and could be run quite often at minimal cost to recovery. Especially after you have a baseline value, it'd be easy to replace a short easy ride or integrate into a longer endurance workout: short warmup, 6-8 minute ramp steps at 10 W increases starting from the previous VT1 power minus 20 W, then cool down or proceed with a longer workout.

I think the first application of this, as mentioned in one of the papers, is using a sub-maximal test for individuals who aren't advised to do high-intensity exercise: think injury rehab. I know someone recovering from a nasty case of post-concussion syndrome that would have liked something like this to track progress when hard work still triggered a migraine headache.

does HR at VT1/LT1 change over time with training within a semi-short period of time? I know as we get older HR @ LT1 probably drops. I actually think the real value is in just finding the HR where LT1 is, then using that to guide the majority of your low intensity training, and I'd guess, once a month or so during one of those lower intensity rides turning on HRV logger and seeing if power is changing. That being said, the ramp test is pretty simple and it's not a taxing workout, so you could do it often enough to get an idea of how aerobic power is improving. The more I think about this, it could be of big value to long distance athletes who do reverse periodization. Those folks get the FTP nice and high early on in a training plan and then work on extending the time they can hang at ironman or 70.3 power. FTP might not be changing too much, but I bet their power at LT1 improves a lot as their muscular endurance and just general aerobic endurance below FTP really gets stronger. That being said, if HR @ LT1 doesn't vary much week to week, month to month, then just keeping an eye on that power at a given HR on lower intensity rides is probably more than good enough (with the caveat that HR drifts around depending on lots of factors). I think knowing the number is important because then you can learn the "feels" and then adjust on days given rest/recovery/hydration/motivation/etc.

---

USAT Level 1 Coach
Team Next Level
http://goteamnltri.com/

Invasive testing is not required.
Clearly visible on a well performed Conconi test just as AT is.
Lactate testing is only required for confirmation in some individuals, usually the ones that struggle to maintain even power during a Conconi.

Good point, HR at VT1/LT1 probably doesn't change much. That's what I get for trying to cover up my limited physiology knowledge by just throwing more words out there. Perhaps use the relatively low change as a sanity check to make sure that a given test has good data?

Hmm, makes me curious about the opposite. I wonder how much my power at VT1 drops over the course of a cyclocross season where I'm basically just doing VO2/anaerobic work or recovery.

Yeah, another way to calibrate the RPE scale.

So, some other more fun rides than an aerobic ramp test got in the way the last few days, but this morning I did what I felt was a solid go at the protocol outlined to determine LT1. I warmed up for 10 minutes, then did 10 minutes at 185 watts (below where I was almost certain LT1 was), then I began the test proper @ 20 minutes into the ride. Did 6 minutes at each step recording in HRV logger (started at the start of the 10min@185w). Steps were 200,210,220,230,240,250. Based on nothing more than RPE and the maffettone calculation, I thought my LT1 was in the 210-215 neighborhood with a HR of around 137-142. The 210 step had an avg HR of 137, and the 220 step had an avg. HR of 142 (so wattage/HR was right where I thought they would be). DFA alpha 1 was nowhere near .75 for those. The last two readings of each step (throw out the initial reading @ 2 minutes because it's not "steady state") was .79 and .84 respectively. There was quite a bit of "noise" in the readings. I had the settings in HRV logger set correctly, and it wasn't picking up any artifacts, so data seemed clean.

Throughout the entire test, I never had more than two consecutive readings under .75 at any of the power levels. I'd get a low number here and there, but then the next one would be up above .75. The only place I got close to .75 was my final step @ 250 watts, avg HR of 148 (.68,.79,.59) and I couldn't pass the "talk test" where i try and recite the pledge of allegiance out loud. 250 watts is about 87% of FTP, so way above where LT1 is supposed to be. If I just straight up average the DFA readings for each step (don't throw out the first 2 minute block) then my 240 watt step puts me at .75 and an avg. HR of 145.

So overall, I feel like my data is suspect, there are too many peaks and valleys, almost always after a low reading, the next one jumps up quite a bit. Overall the trend is downwards, but there certainly wasn't a clear spot where it goes below .75 and stays there. Maybe it's time to find someone with a pro version of Kubios to run my file for me :)

---

USAT Level 1 Coach
Team Next Level
http://goteamnltri.com/

We have uploaded a Python chart for activity analysis in GoldenCheetah, based on Marco public implementation in Python (https://github.com/marcoalt/Colab).

The easiest way to use it is installing GC v3.6 (Last development or Snapshot builds from http://www.goldencheetah.org/#section-download), Check included Python is enable in Settings and use View > Download Chart from Activities View.

Hope it helps, Ale.

---

Ale Martinez
www.amtriathlon.com

AI Endurance is now tracking DFA alpha 1 as well:

https://aiendurance.com/blog/dfa-alpha-1-app-to-track-your-heart-rate-variability-thresholds

Bruce Rogers, one of the original authors of the research papers, has given us his seal of approval in terms of accuracy:

http://www.muscleoxygentraining.com/...ccuracy-initial.html

The obvious benefit is correctly setting your training zones. Especially for the aerobic threshold this is a big step forward since the 'conversation test' is the only current alternative without having access to a physiology lab.

The DFA alpha 1 ramp test is generally not very taxing as it only requires you to go past your (an)aerobic threshold for a few seconds to determine power/pace thresholds.

I think that this is going to be huge for automatic threshold tracking. We regularly push past the (an)aerobic thresholds in our activities when we just dial up the intensity enough even without doing a ramp test, for example when we encounter a steep enough hill. However, the dialling up of intensity has to be steady and consistent enough, resembling the ramp test protocol sufficiently to detect thresholds.

As a counter example, if your FTP is 300 W and you would go from coasting to holding exactly 400 W instantaneously then your alpha 1 would cross 0.75 and successively 0.5 within a couple of minutes or so which obviously doesn't mean your thresholds are 400 W. So threshold crossings like this are not useful for actually detecting pace/power thresholds from HRV data.

If however there is a portion of an activity that sufficiently resembles the slow ramping up of the ramp protocol, AI Endurance is already automatically detecting your power/pace and heart rate HRV (an)aerobic thresholds.

There are also obvious benefits for our AI as the more often it gets information on the fitness state, the better it can correlate an athlete's training input with the latter.

---

Founder of AI Endurance
https://aiendurance.com

I was hoping that GC had some form of this enabled but I am not sure I have the depth of knowledge to follow your instruction Ale. How do you get the data into GC in a way it can work? What data does it use? And then how does it spit out metrics like DFA Alpha 1?

I have tried the HRV logger but it only sends out DFA a1 data every 2 minutes which is rather slow to capture crossing points. I feel like it is likely that this info will become more useful with applications that send out DFA a1 data on a faster time interval so that you can see the change in the value with better resolution?

Has anyone found a good app for the live tracking that gives more data points than HRV Logger?
This is an interesting subject and I am still trying to find an option that works.

Usually importing a FIT file from your activity, although other formats are supported.


R-R data, so you need a Heart Rate Strap capable of recording them and a device or app to collect the data and generate an importable file.


This is how the chart looks like:

---

Ale Martinez
www.amtriathlon.com

Here's a better chart:

---

Eric Reid AeroFit | Instagram Portfolio
Aerodynamic Retul Bike Fitting

“You are experiencing the criminal coverup of a foreign backed fascist hostile takeover of a mafia shakedown of an authoritarian religious slow motion coup. Persuade people to vote for Democracy.”

Hi all,

Has anyone who's done this gotten wildly different HR thresholds from this test? I did both the Bike (this morning) and the Run (just now) test today, and I'm getting much different readings

For Bike: 215w, 164 bpm
For Run: 5:20/km, 144 bpm (treadmill)

I'm 40 years old, so the usual HR - age formulae would suggest that ~140bpm is ballpark, so 144bpm isn't that far off. The only thing I can think of is that I used a high quality strap for the run (TICKR), but I used a lower quality ($40 Amazon special) for the bike, since that's what was lying around this morning.

Though based on 288w FTP, so 215w is also reasonable as a Zone 2/3 boundary. I'm a stronger runner, though I've spent much more time below these figures on the bike than on the run.

Any thoughts? Should I retest the bike with the better HR strap? or was fatigue maybe an issue with the run test since I did it 2nd workout of the day?

ETA: I'm using HRV4Training as the app

for one thing the heart rate strap could definitely be the source of problems try using the best available but the same both times. Just a note it works best if using blue tooth as well. Of course the heart rate vs DFA a1 for cycling and running should likely be different since the type of muscles used etc are very different. Also of note it has been said that running can lead to issues with clean heart rate data so again could lead to some variation because DFA a1 likes god clean data. So there are numerous reason why they would be different but most important is the strap to start with.

Are you just eye balling ti based upon the HRV logger live data output? The 2 minute lapse between reporting data is not a good thing. You can though export the HRV RR data then import and merge to a run or bike file in Runalyze, this will used the raw RR data and you can see the DFA a1 at 0.75 or at any point for that matter.

What sort of ramp did you use for the two workouts? Again the ramp method could impact the result. Heat is also known to cause results to drift so cooling is important. There are a lot of things that can impact results. Start with just trying to get similar results for the same activity then if that is working for you the next step is across activity types.

There are many things that could cause the difference. The above is my uneducated list of guesses ;-) !

Thanks for the assistance.

The ramp was indeed different for the two tests. For the bike I did 5 mins @160w to warm up and then started at 170w, increasing by 5w every 5 mins, so it took the better part of an hour to ramp up to where it registered.

For the run, I assumed that I had a large buffer, given the higher than expected HR reported on the bike (I assumed HR would be similar for both), so I ended up almost at 0.75 right off the bat when I started the treadmill. I then adjusted on the fly (in 2 min blocks) to see how I could hold 0.75 as I went.

Looks like the plan is to go back with similar HR monitors, and to redo them both using the gentle ramp.

I'll definately check out Runalyze.

Thanks!

I am not certain the gentle ramp is best. I have played a bit with ramp and imho as long as it is consistent then the result is relatively consistent, but that means nothing for accurate. Bruce Rogers is the guru of this and does a lot of posting on his website to explain how this may work. check out: http://www.muscleoxygentraining.com/...e-intensity-faq.html
he answers virtually all of your questions/ issues. He has stated a recommended ramp rate and range. on that page: Can I do a short, fast, rapid rise ramp?

• Possibly, although most ramps are done with a 5 to 10w/min rise, you may be able to use the 30w/min protocol - see DFA a1 HRVT and Ramp slope

So this is far from a set in stone protocol. They are still learning and refining what they know about it. Runalyze is a great tool for doing the analysis but does not have the ability to export the file after merging. Also there are some setting in Runalyze that can help with the analysis, such as limiting the range of values to between x and y time rather than the whole file. It is a very powerful tool once you figure it out. What I do is add the x.FIT file from the ride without RR data, then merge the RR data from the HRV logger. This works well from what I can see. Runalyze allows you to set the frequency of results, but 1 second or 5 second rolling results both work well for me. I prefer the higher frequency daa. You can see how DFA a1 varies with time and that should look like your power data or our HR data, for running it can even correlate to pace data.

This is really a rabbit hole once you enter there may be no way out just ask Alice! ;-)

Can you comment on LT2 @ DFA alpha 1 = 0.5? Is this reliably correlated?

I’ve just found this topic and as others have said, I’ve gone down the rabbit hole, I find it very fascinating, especially given recent LD race results from the Norwegians - who seem to be at the cutting edge with this type of data.

I’m sure many will judge it but I’ve recently “dumbed” down my zones into 3 zones, with zone 1 56-75% ftp, zone 2 low end 76-90% ftp, zone 2 high end 91-105%, and Z3 +110% ftp. So I’m approximating my Aerobic Threshold at that 75% mark and trying to stay lower as often as not.

I did find one article from Rogers regarding Anaerobic Threshold correlating to DFA alpha 1 = 0.5 but I need to get my head around the details still. Would appreciate your thoughts.

I’m planning to work through a gentle ramp test that I put together, 6 min warmup 55-65%, then increasing 5w every 8 minutes so I have 6 steps within a 60 min workout. Obviously that’s my own “protocol” and I don’t have a Polar H10 which seems to matter? Interested to see what the results are anyway.