Some notes on doing HRV readings with the Elite HRV app and a Scosche Rhythm24:
Elite HRV can’t find the Scosche Rhythm24 unless I briefly open the Rhythm Sync app first (without connecting to the device).
It is possible to get readings that are too high even when the app claims that no artifacts were detected and that the signal quality was good (the HRV4Training app has the same issue). The problem is obvious when looking at the RR-interval chart (which should look like rolling hills, not jagged peaks, see the screenshots below).
To get good readings, the strap should be tight on the lower arm, the arm should rest on a hard surface, out of direct sunlight, and covered with a dark piece of cloth (if not wearing long sleeves).
Here are a few exercise bike sessions, recorded over several weeks with a Scosche Rhythm24 and the myWorkouts app. The gray intervals are where I’m supposed to exert myself to the maximum. The outliers are bad data from before I figured out that I’m better off wearing the sensor on the upper arm rather than on the forearm. Looks like the sensor will rather report a too low heart rate than admit defeat…
Here are TRIMPexp scores for a few different activities (ELT=elliptical trainer, EXB=exercise bike, SKI=skiing, SWM=lap swimming, WLK=hiking), plotted by duration.
Elliptical trainer or exercise bike sessions would have been ideal for comparing the effects of different routines on recovery as they can be done regularly, and in a more or less controlled environment.
Unfortunately, activities with a score of less than 500 don’t seem to have a measurable impact on the next day’s HRV, which is how I was going to measure recovery. Activities with a score above 100 may have a cumulative effect if done daily (not happening), and activities with a score of 500+ are infrequent (once or twice a month), and not very controlled (i.e. I’m not repeating the exact same hike multiple times under similar conditions).
Nevertheless, I’ll keep tracking heart rate while hiking, as well as doing daily HRV measurements throughout this summer
Just got to do an activity with a score of almost 900 (which is about as far as I’ll go), and did not see an impact on HRV during the following days, despite being in obvious need of some recovery time…
Here are the HRV readings I have taken this year: The values are based on the RMSSD, normalized on a log scale to a 0-100 score. Outliers are marked red. The line is the 7-day moving average, and the background shading is the activity level for the day (based on step counts), with darker being more active.
High outliers are mostly bad readings (I suspect), and only one of the low outliers corresponded with anything obvious (sore throat).
If there is a signal in that data that reflects exercise “readiness”, it’s clearly being masked by day-to-day noise (caused by a not perfectly regular sleeping schedule and other factors).
Still holding some hope for long-term trends or seasonal patterns…
Here’s an alternative approach: Tracking how many days it takes for my body fat (or whatever it is the Withings scale measures…) to recover to its previous level.
Neither my diet nor my sleep nor my exercise during the “recovery” periods are consistent, so there is indeed too much noise to see any clear effects except in extreme cases (see chart), and those extreme cases are too few and too unique to support much experimentation.
I guess I’ll just post the code here… The code is very much custom and one-off, but some of it (like the TRIMPexp calculation, or the charts) may be copy-pasteable.
Have you looked into using the Polar H10 + HRV Logger to track your HRV? I have been using that stack for a few months and now want to start analyzing the data to see if I find any interesting findings.
As mentioned in the first post, chest straps didn’t work well for me (Polar H7). Used HRV4Training for a while (HRV Logger wasn’t available on Android) before switching to Elite HRV (paced breathing, multiple recordings per day).
Please share if you find anything in your data, or also if you don’t
Will def share! I just got 2 months of R-R intervals into a giant dataframe in Pandas now computing HRV.
Sorry if Im beating a dead horse but Im not smearing electrode gel on my chest strap and seems to be working fine…have you tried without and the only way you get readings is with gel? Just trying to make sure Im getting good data.
That was my conclusion. Even had the straps replaced, just to be sure. But this will depend on how dry/oily/sweaty/hairy your skin is. Should be obvious if the data is bad (gaps in the data, or noise like in one of the two Elite HRV screenshots above).
My yearly HRV analysis confirms better sensivity for standing measure. Also i’ve found that Oura ring nightly rmssd and polar h10 morning standing rmssd are both correlated with physical load. I’m using h10 with a little water instead of a gel and it provides enough accuracy to detect these effects.
My watch (Garmin Fenix 6S) automatically records an “overnight average” HRV, so I was wondering if that might be useful, and how those values compare to manually recorded measurements.
I used the HRV4Training app for 2½ months to take 1-minute readings shortly after getting up every morning, while standing, using my phone’s built-in camera (Pixel 6a). I repeated any readings that didn’t show as having “excellent” signal quality (washing my hands with hot water seemed to help).
I was slightly above average active during this time, but the two vertical lines were the only two days I “overreached” a bit (according to Garmin, and subjectively).
Results: There is no overall correlation between the two measurements (except maybe for a week or so). Garmin’s values appear to be a bit less random and are trending upwards, plus don’t require any effort to collect, so I think I’ll just go with those
easy hike, 
backpacking
