Thursday, April 16, 2020

The future of consumer muscle O2 monitoring - trainer derived MLSS?

The recent demise of the Humon Hex on top of the prior exit of BSX Athletics from the field of muscle O2 saturation monitoring makes this an appropriate time to talk about several issues.
  • Why can't these companies turn a promising idea into a successful ongoing business?
  • What choices do we have left in monitoring muscle O2?
  • Is there still a role of this technology in the consumer/amateur exercise field?

I can only conjecture on the first question, but it may have something to do with my whole point in starting this blog many years ago.  After reading the forums on the Moxy website, I was appalled by the lack of hard fact and abundance of pure anecdote.  Wild claims were made for all of these devices (read my review of the Hex for example) that although they may have a shred of theoretical logic, most are just plain nonsense.  Therefore consumers buying these units will ultimately be disappointed by the promise of unrealistic expectations.  No - they don't help in recovery, effort modulation (well, without a proper ramp and data analysis), optimize your workout routine!

Believe me, I am not "trying to kick a dog while they are down" or an "I told you so".  But, I did have an honest conversation with the Humon folks before launch and respectfully asked them to be honest about the claims.  So if you were a buyer and hoped for the above features, you would have been very disappointed.

What is still available now?
Unless you can get a "new old stock" or used BSX/Hex on Ebay, the only option is the Moxy.  The price is just over $800 for 1 sensor.  They have been around the longest and have a good reputation for standing behind their product.  But, the device has not seen a major hardware revision (ever?) and is still a clunky box.  Given that they are probably struggling to stay in business, I understand the lack of funding for a redesign.

Now the key question - Is there any (major) role for the sensor in current athletic training.
The answer is yes - there are now enough studies showing that identifying the "breakpoint" in VL or RF desaturation that occurs with increasing exercise effort, can be used as a surrogate for a MLSS, RCP, VT2, LT2 intensity.  The question is - is it worth the cost?  Alternatives include buying a lactate meter and strips (about $450), then performing a series of intervals to generate a lactate curve.  Advantages of a lactate test is that you may be able to also get a LT1 intensity to help with zone 1 demarcation.  However, the issues with lactate testing are many and the result may not be precise.  Other modalities to get a VT2/LT2 include gas exchange testing (about $150 for my local center), and doing an FTP test which we have covered.  
The choice is yours and they all have their merits.  But the Moxy price just does not seem to make sense for what you are buying (forget the hype).

If someone came up with a unit in the $150 price range to calculate your MLSS, that would be a good cost:benefit ratio.  In fact, since you really need a smart trainer with power monitoring (or a treadmill) to do this accurately, it could be offered as an inexpensive add on for the trainer purchase.

In the past, I have struggled to make ramps "work".  It is very difficult to both cycle hard, keep an eye on power as well as ramp time.  The appearance of affordable smart trainer has made doing a ramp a much different experience.  With a feedback controlled trainer using Zwift (or equivilant) all you need to do is pedal and the rest is done for you.  That makes doing a desaturation ramp easy.  Below is another example with both the RF and VL:

The following data is an example of how trainers ans O2 sensor tech can work together.  
The test subject is myself (VO2 max 55 ml/kg/min and a RCP/LT2 power of 275+- watts).  
The ramp was done on an Elite Suito with my Assomia Duo pedals providing the cycling power data.  The Humon Hex was on my L RF and the BSX Insight on my R VL.  Thanks to Zwift, it was easy to create an incremental ramp starting at 100w, with an increase of 10 watts per minute up to 310 watts (almost my maximum which is near 330w).

The first plot is the Cyclinganalytics.com power curve:


  • The cadence is the same throughout and the power rises quite smoothly.  Even though it's set for 10 watts per minute, Zwift is able to raise resistance continuously, so no abrupt transition occurs.

The Data:
We will need to plot this out officially in order to interpolate the MLSS/RCP from the SmO2 data.
For example, let's say the breakpoint occurs at 1080 seconds into the test, what is the equivilant average wattage (intensity) at that point?

Here is an Excel plot (with regression formula):
 According to the formula, at a time of 1080 seconds, wattage was 278.


Now we look at the Vastus desaturation curve with calculated HHb.  Remember the VL shows a plateau starting at the MLSS as opposed to the abrupt rise seen in the RF behavior.


And the Rectus Femoris:
Here we are looking for the departure of a linear desaturation (increase in HHb), with an upward change in slope:

The lines are drawn in and show the following:
  • VL plateau begins at about 1080 seconds.
  • RF curve slope steepens at about the same time, 1080 seconds.
  • Power vs time formula gives us a derived wattage of 278 with a time (x) of 1080 seconds.
  • The curve shift needed given the ramp speed (10w/min) is about 3 watts.
  • Final MLSS/RCP is about 275 watts.  Very close to what my previous gas exchange and lactate testing showed.

Although this seems easy to do, it is not as simple as one may think.
Some potential limitations and source of error (yes, I've seem them all):
  • Wrong sensor position - such as between the RF and VL area.
  • Poor control of incremental power rise (common).  In the days before a smart trainer, this would be very difficult to pull off and I rarely got a good looking curve.
  • Difficulty in plots and graphs.  Although not at the rocket science level, one does need some basic math and spreadsheet skills.

Final thoughts and wish list:
  • Is this the end of the consumer muscle O2 story?  I hope not.  It still is a viable alternative to measure RCP/MLSS breakpoints, which is helpful in setting up your zone 3 training intensity.  Yes, lactate testing can be done, if you are willing to jump those hurdles.  Yes, FTP can be done but is exhausting and can vary in accuracy.
  • Given the interdependence of both accurate ramp intensity structure with muscle desaturation curves, it makes sense that either trainer manufacturers and/or new startups in sensor tech build an integrated framework around this typical usage scenario.  Take the data plotting out of the picture - simply put the sensor on, start a ramp and get your result.  BSX started with this but it was before smart trainer availability.
  • The sensor vendors need to abandon the hype and hypothetical claims.  They should rely on the solid foundation of the VL or RF breakpoint as a means of RCP/MLSS identification.  Having one solid usage scenario is fine, we don't need five.
  • Price points needs to be reasonable.  The hardware costs can't be much higher than a Polar OH1.  Bundling this with many smart trainers should enter into the product volume calculations.
  • And here you have the published evidence: 
  • Combining NIRS and DFA a1 for critical intensity estimation  

 See also: