Theoretical Model for Using Muscle Oxygenation as the New Training Target

Moxy Muscle Oxygen Sensor in the light guard.

Moxy Muscle Oxygen Sensor in the light guard.

Technology capable of measuring muscle oxygenation has now been available to athletes and coaches for a few years.  When it first came out, there was hype that live readouts of SmO2(Muscle Oxygenation Saturation) would revolutionize the way we train.  So far, the revolution is a bit delayed. Training with direct force power measurement is still the gold standard for consumer-level athletes.  But that doesn't mean that muscle oxygenation is dead. I've now been using the MOXY Monitor muscle oxygen sensor for about 6 months and have a theory about the way we might get serious benefit out of this little blinky-light box.  This post is (hopefully) the beginning of a series, and a conversation, where I will outline, test, and discuss a potential new method of using muscle oxygenation measures for daily training.


As a trained scientist, I feel obliged to disclaimer this upfront.  This is, at its heart, inductive science. I have a basic understanding of the systems involved but am not sure if there will be utility and thus wish to explore this by gathering data and exploring what the data suggest.  Feel free to follow along, or spin off your own research from this. The intent of publishing this is to possibly accelerate the development beyond my own limited capacity as a full-time grad student and coach.

First, What is Watts?

Cat 5 Cyclocross Racer Power Profile Curve

Cat 5 Cyclocross Racer Power Profile Curve

Before we can dive into possible future directions, I think it is valuable to understand where we are right now as an industry that uses power output over time(Wattage or Watts).  Watts are a direct force measurement of work. That is, how much energy is being transferred over a period of time. 1 Watt is also equal to 1 Joule of energy transferred per second.  For our purposes in sport, it is a standard measure of output that, when combined with the knowledge of how heavy the object is that is being moved, provides an apples to apples comparison of athletic potential.

An athlete's specific wattage output ability for a specific duration divided by their mass (Watts per Kilogram) can give you a really accurate estimation of if they are competitive in specific races.  For example, an athlete that can output 25 W/Kg over a 5-second sprint is considered world class.  Translating that, assuming average US male weight of ~89Kg, that would be someone averaging 2,226 Watts for 5 seconds.  For reference, most world tour level bike sprinters top out around 1600-1800 Watts(but they also weigh about as much as a damp kleenex too).

Power meters were first adapted for cycling in the late '80s, but didn't see wide usage in the pros for another decade, and consumer usage really only started to boom in the last five.  it is still prohibitively expensive for most age group or hobby level athletes, so why use it? Well, it works better than anything else we have had, it is well supported, and it provides a useful yardstick to benchmark athlete potential.  All my clients use it. I train on it. If you want to learn more about it, I really recommend  Joe Friel's "The Powermeter Handbook."  I want to be clear, I'm not here to take down power training.  I believe in its utility.

But training with power has limits.

1) It doesn't translate well to other sports.

Despite the hard work of a lot of sports tech companies and athletes, running, swimming, lifting etc still don't have a convenient power metric that provides the same level of utility as it does in cycling.

2) It is a relatively indirect measure of the changes that actually matter.

In the process of training, the athlete wants their body to become the most effective at the collection, transportation, and utilization of oxygen and glycogen as well as the removal of CO2 and metabolization of lactic acid.  All of those systems working together produce output, but at best, wattage is an indirect measure of any one of them and thus has limited utility for prescribing training conditions that will result in improvements of each of those systems.  This is why athletes do blood tests for lactate threshold and carbohydrate/fat metabolism, and respiration tests for volumetric efficiency.

3) It is a delayed measure of physiological change.

For the same reasons it’s indirect, it also means that the wattage number you see today is the result of stress/strain adaptation to efforts that were weeks prior.  Your output is the combinatorial result of your different physiological systems changing in different ways over time. If you get better at transporting oxygen, but worse at metabolizing lactic acid, your output might go down or up or nowhere depending on the relative changes.  As a coach, when I see strange wattage numbers from an athlete, I have to go back to past work done and ask "how does this output relate to this past work?" Coaches get better at identifying and planning for these delayed and combinatorial effects, but no client enjoys being told that their unexpected poor performance today was because three weeks ago I told them to go too hard for too many days or because I didn't properly ramp their workload over the past three months.  Ain't nobody got time for that, but that is the training world we live in right now. So there is still room for improvement.


And this is where we get into Muscle Oxygenation (SmO2).

What is SmO2?

SmO2 is short for Saturation of Muscle tissue with Oxygen.  It is represented as a percentage of your blood flow that is carrying fresh oxygen to the muscle tissue as well as the oxygen that is currently stored within the myoglobin in the muscle tissue. (e.g. %O2)  A measure of 100% would theoretically be claiming that every blood cell passing by the sensor is carrying fresh oxygen to the tissue and every myoglobin within the tissue is bonded to an oxygen molecule. But we know there are other types of cells in the blood stream, not every cell gets a fistful of oxygen everytime it goes through the lungs, and not every myoglobin holds an oxygen molecule all the time. So the reality is the average healthy individual at rest will see a reading between 80-98% depending on the location of the sensor, altitude, and volumetric efficiency(amongst other factors).  If you want to deep dive on the what and the why of SMo2, I suggest MOXY's Muscle Oxygen Physiology course.

Why does it matter to Athletes?

At the most fundamental level, a muscle that gets no new oxygen has a finite amount of work that it can do before it just stops.  Thus an athlete that can deliver and utilize more of that incoming oxygen, can do more work over any given period of time. As discussed above, the athlete that has the highest work capacity to mass ratio for a given duration has a high probability of winning a race of that duration.

1) It's a more direct measurement.

SmO2 more directly measures the amount of oxygen being delivered, stored, and used in the process of doing work.  Yes, there are still questions as to what is being measured specifically, but the research is being done and the early data support this claim.

2) The measurement translates across sports

The sensor is portable.  A runner, soccer player, gymnast or swimmer can use it in the same manner. Generally, I don't have to buy a new sensor for every bike or every sport.  One and done. More only if you want to get way into the details of how your body prioritizes oxygen delivery under specific strain. (Sidenote: Moxy reports that they test all the units for water ingress while under pressure but still suggest limiting the depth of usage because high pressure can push the power button.  Also some swimmers keep the sensor in a ziplock baggie for additional protection without any interference issues with the optical sensor. Anecdotally, the sensor is durable. A sliding bike crash right on the sensor did more damage to my skin and ego. The sensor was barely scratched and never missed a datapoint.)

3) There is less delay between work done and adaptation.

I'll be honest, this is the most tenuous of my claims and needs a lot of exploration, but here is my theory.  There has been a lot of research done to identify what changes in SmO2 data reflect in the actual physiology of the athlete.  With this, we have a good idea that if we see certain indicators, we have changed the collection, transportation, or utilization systems.  Thus, if delivery improves, we don't have to wait until that improvement produces an increase in power output before we see change. We see it sooner and can adapt training relative to observable changes.  Theoretically, we can even get precision down to daily workouts. More on that below.

Why hasn't this happened yet?

Good question.  These sensors have been available for a while now and yet, outside of the niche community, no one has been able to sell SmO2 as a significant alternative to power.  One of the two sensor sellers closed up shop already. There are some limited use cases, but there is still research to be done. So why should you listen to some grad student? Two reasons:

  1. Calculus

  2. Nike

Yup.  Maths.  And a shoe retailer? Hear me out.

First the maths.

SmO2 as a direct measurement for training, during training, is not available to the consumer. Yes, we can do post workout analysis and know more about the physiological changes,  We can also see what out saturation percentage and flow volume is moment by moment, but those aren’t useful during the exercise.  Here is an analogy to help illustrate.  If I told you, you are driving a car at 70 miles per hour, what do you do?  How do you use that measure in the moment? Do you speed up? Slow down? Turn?  You're smart so likely you'll ask me for more information. Where are you? Why are you driving the car? Are you speeding up or slowing down?  All these questions give you context to make a decision at the moment.

Measurements are only useful with context. Calculus gives context.

WKO4 SmO2 Time in Saturation Zones Workout Analysis Chart

Calculus transforms SmO2 into a measure with more context.  If you take that same speed measurement in the above analogy and transform it to the first derivative(that's the calculus part), you get acceleration.  Now you know your current speed but also if you are speeding up or slowing down. If you take the first derivative of SmO2, you get the rate of change of oxygen saturation.  Now you know if your muscles are utilizing oxygen from your blood faster than you can put it back in(desaturating), delivering oxygen faster than you can use it(saturating), or at an equilibrium where your muscles are using the amount that your body can deliver(balance).

This additional information about the oxygen balance could, theoretically, be used to define training zones similar to power training but with more accuracy day to day.  The underlying theory of power training zones is the overcompensation theory of physiological response. In brief, our body will overcompensate when repeatedly exposed to a measured amount of stress for a specific duration.  If we exceed a certain threshold of stress and duration the workout is destructive. If we don't get close enough, the effect is limited(but there is a positive effect).  Currently, Functional Threshold Power(FTP) and our individualized Power Profile, provide a wattage-based guide to roughly model that threshold.  From that threshold, we set wattage zones(ranges of output) as a guide for how to accomplish athletic goals.

In theory, Saturation/Desaturation zones could be used in lieu of wattage zones.  There is no reason why we couldn't adopt the same concepts from power profiling and thresholds to this derived measure.  And the live measurement during exercise could be naturally adaptive to the current capacity of the muscle tissue and oxygenation system that day.  That is, theoretically the equilibrium saturate/desaturate point will float respective to your ability to utilize and transport oxygen on a daily basis. (As I think about this more, this may be negatively correlated with fatigue. Needs further investigation.)  In other words, we don't have to wait three weeks to see a change and adjust our targets. We may be able to improve targeting resolution and mark physiological adaptation on a daily, or even workout to workout, basis. There will still be cumulative effects that won't show up right away, but the possibility of a more precise and timely measure of a significant part of the subsystems affecting capacity to do work is exciting.

What does Nike have to do with this?

This application for a patent.  Buried in that application is a discussion of how their sports performance lab is already investigating exactly what I'm talking about. They deem it worthwhile enough to invest in research and patenting a wearable that would do this exact transformation of the SmO2 data.  They go further into regression modeling, expert systems, and translation to a simplified points system, but the concept is there. When I started exploring this transformation and asked around, finding this application was validation enough that we need to dive deeper.

So what's next?

I've done preliminary work on this and have a few thoughts on where to go from here and some of the foreseeable hurdles.

  1. Get the number transformed.  I'm not a programmer, but if someone is, Garmin's Connect IQ platform can be used to create a data display and capture channel that transforms the raw data from the MOXY sensor into this derived rate measure.  I have built out a WKO4 Chart for post-workout analysis for now, but until we can get the live number, we are limited on what experimentation can occur. If you are interested in pursuing the programming side or using the chart, contact me.  I'll share my chart and I have an in so we don't have to start from scratch on the Connect IQ stuff.

  2. Figure out limitations of the metric as a tool.  As I already mentioned, the way that equilibrium probably floats may be really useful, or something that we have to adapt to.  Additionally, when I first started charting this, I realized that if we are measuring only one leg, there probably periods of both saturation and desaturation in every pedal stroke/step/kick.  As an athlete's cadence changes, this may produce strange artifacts in the measure relative to the sampling rate of the device. I think this can be programmatically filtered out, but definitely needs to be explored.

  3. Analyze current data available for correlations between time in desaturation ranges(this should indicate a muscle and/or oxygenation system under strain) and long-term changes in FTP and Watts per Kilogram.  Ultimately, the athlete wants to achieve a certain capacity to do work. Utility in this measure will only prove out if we can show that it is somehow related to improvements in that capacity.

  4. Start gathering data across sports and compare it to changes in functional measures.  Runners and swimmers, does time in any particular desaturation range correlate to getting faster?  How might we use this in weightlifting? Gymnastics? Alpine sports? Does this have the portable utility that theory suggests?

  5. Share alike.  Please. I learned much of what I know about SmO2 from the users who are already sharing on the MOXY forums and from those who were willing to share their expertise and hard earned research with a random guy who emailed them questions.  Cooperation has gotten us this far. As a coach, this has the potential for very high utility, but as a full-time grad student working on a Ph.D., I can't devote enough time, money, or effort on this alone to make serious, timely progress.  So I've put it out to you and all I ask is that you pay it forward and share what comes of this back to the coaching and athletic community.

  6. Ask questions and spread the word.  I am not Joe Friel or Ray Maker. I can't just send a tweet or email and have a lot of sports scientists or coaches read my little tiny blog.  But if you made it this far(congrats and thank you!), you are someone, or know someone, who finds value in what I've written. Pass this along and let me know.  If you are confused, agree with me, or disagree, leave a comment or email and I'll try to provide a valuable and functional response.