I think the biggest obstacle in using DFA a1 is getting good feed without artifacts (or at artifact level which would not skew the metric) with casual HR straps like H10. This or better artifact handling in a software.
Thanks also for poining indirectly to HRVAnalysis 1.2, will try that out. I never had luck with Kubios, it either had graphic issues (on the MS Surface) or runs so sluggish that it is not usable. 2Gb runtime is a bad idea anyway.
That is a concern, especially with rates above 3%, which is easy to do when in motion. We are evaluating that issue now and it is an important one to be aware of.
Each has it's advantages. Kubios premium (latest version) is a super tool, especially the time varying analysis. We could not have generated the data seen in Figure 3 of the article otherwise. The time varying option allows one to look at a recalculation of HRV data every X seconds (we usually do every 5 or 10 sec) and window length (we use 2 min) over the course of a session. But - yes, it's costly and RAM intensive. HRVAnalysis is an accurate, free tool, best used for looking at a few fixed width time intervals.
Yes an interesting study but on a different topic. :-) @bjrmd I found my Gas exchange lab test results so I can finally compare it with DFAa1 method. If I won't get lost in Kubios.
If you would like, send me the data (?PM) and I'll do the VT1 and DFA a1 plots for you. Gas exchange by hand can be tricky and the automated analysis is frequently wrong.
Hi, I looked at your graphs that you sent to my blog. Let's get into a better discussion about this. First, I need the raw gas exchange data since I don't trust automated readings off the metabolic carts. Unless you had a really good physiologist looking things over, we should document that the "gold standard" is indeed gold. Second, I would also like the raw RR series. If artifacts are high and/or there are APCs, the DFA a1 can be markedly shifted. The auto correction mode in Kubios does a reasonable job of this but the standard "threshold" correction method in the free version can be way off. I have the premium version so I have both methods. In addition, only the premium version allows "time varying" output - a 2 minute measurement window with recalculation every 5-10 seconds (user defined). Yes, you can use the free version for a very rough guide or for low artifact constant power intervals. But for incremental ramps, it's tricky to do that. Send me a PM
These are 2 activities where I did the step test. 2nd one got much more artifacts, 1st one is quite clean. I tried to measure at the end of intervals where I got 0 or max. 1 artifact per section and skipped the sections with artifacts. I'm using Kubios free so I have only Threshold correction but I tried to avoid artifact sections for calculation of the parameters.
