Aerobic Decoupling and aerobic threshold


I’ve found few webpages describing method of finding aerobic threshold by using aerobic decoupling of 5% as a marker of aerobic threshold. But overall there is very little about this method, Joe Friel uses it as an estimation of aerobic fitness but as far as i’m aware he does not link it with aerobic threshold.

I suppose i could provide bit of meat over the bones here:

Last february i rode 2hours at 220watts. HR was on average 144BPM, and Pw:HR was 0,5% after warming up.

This week i held 230 watts for 2 hours, average HR 159BPM (quite hot conditions). My Pw:HR was 2% after warming up (it took 30 minutes for heart rate to stabilize). So basically i wasn’t even “close” to my aerobic threshold.

My FTP is around 280-290watts, there is no major difference between two sessions when it comes to fitness. Both tests were done indoor on same trainer using Zwift and ERG mode. I used training peaks to give me the averages. So for what it’s worth equipment and software are similar.

And on regular day i consider my february heart rate of 144BPM to be normal. Heart rate of 230watts session is way above normal values. It correlates to something like 250-260watts on normal day. Heart rate at FTP is around 174BPM.

So despite huge difference in average heart rate, decoupling should be within aerobic threshold’s bounds (infact way below). Question is: Is this a reliable method?

And another question: If my heart rate would have been similar 144-146BPM in both session with decoupling being similar, would that change anything if my high heart rate would ruin the comparison?

I don’t know if this is a stupid question. On one hand it seems valid testing method, and on other hand maybe this apples and oranges comparison (=lactate vs heart rate as a marker).

1 Like

Hi @antti.hankonen

Actually that’s a really good question and not one that I personally have a full answer for. I do agree with Joe Friel that cardiac drift is a great indicator of fitness. But I’m still not personally willing to use it as a way to exactly estimate aerobic threshold. There’s just too many variables. One of which you’re pointing out - many factors affect both heart rate and power. So you can have two rides at similar wattage with very different heart rate responses.

Here’s two examples from my own riding. This is a graph I developed based on Dr Seiler’s formulars - the green line indicates heart rate to power and a rise shows cardiac drift. Ideally the green line is very close to the 1:1 horizontal black line you see below.

First, here’s a ride I did last weekend. As you can see there’s a little cardiac drift but mostly the green line stayed flat and very close to 1:1:

By contrast, here’s a ride from this time last year. Fitness was a little better last year and last year’s ride is actually also a little easier (power-wise) than the ride above. Notice how much higher the green line gets:

The big difference in last year’s ride was that it was an extremely hot day and I wasn’t hydrating nearly enough.

My guess is that something similar was going on with you between your two rides. The 159BPM ride might have been in the heat, or you might have had some heart rate depression in the 144BPM ride.

I check the cardiac drift graph on every ride for myself and my athletes and doing that, I think you can see trends over time that allow you to draw some conclusions about stamina and aerobic threshold. So, yes, if I understand your question correctly, if your heart rates had been similar and you had seem similar cardiac drifts on both of those rides, then you can start to say “all else was likely equal” and draw some conclusions.

Here’s an example, this is another ride from me where heat and other factors had been accounted for. You’ll also notice a consistent green line that kicks up in the last 45 minutes of the ride:

I feel pretty confident based on this graph that I was riding very close to my aerobic threshold.

Hope that helps!

Thanks for a great answer!

So it should be done often, couple times a year just doesn’t cut it. And there is no certainty of it giving correct threshold.