For an updated discussion see this
As discussed in the previous posts, observation of non locomotor muscle O2 saturation may be quite valuable in assessing both training and racing intensity zones. In addition, looking at a surrogate, namely the respiratory motor group (costal) may provide a better snapshot of current physiological status. When I began using an O2 sensor, the real time dynamic usefulness of the data did not seem worth the trouble. Yes, there is supporting literature in using muscle O2 drop in looking at VO2 max/lactate thresholds but that is not of interest to most athletes.
But what I have been trying to work on is a more "real time" marker of exercise intensity sustainability. In other words, knowing how hard you can push yourself, without decompensation. Obviously, knowing your muscle O2 sat will be of only academic interest in a 1 min all out interval. However, getting a handle on the potential zone of non sustainability would be quite helpful.
To that end, from this post on, I will be showing simultaneous tracings of costal O2 with locomotor areas (like the vastus). Using 2 sensors with Ipbike (the android app I use while riding) is a bit problematic. There is one channel for muscle O2, but for a second sensor, we need to hijack the speed/cadence channel (the BSX will emulate cadence equaling the O2). So some of the graphs will have a bit of a scaling difference. Garmin Connect will display up to 3 sensors and will also be used occasionally.
So back to the the real time observation of a "sustainability", decompensation curve in both a short maximal effort as well as a longer interval of ten minutes. This type of data has been presented before, but let's look from the vantage point of VL vs Costal data measured at the same time.
First, a one minute all out (about 525 watt avg), followed by about 1 min at 150 watts, then 1 min at 250 watts.
The costal O2 rapidly drops (baseline 67%) during the intense 1 min (nadir about 5%), stays down flat almost 30 secs into the recovery 150 watt zone, then slowly rises to about 34% until the 250 watt zone, where it drops slightly to 30%, and does not recover until I stop pedaling. It finally comes back to baseline about 2 min later, but at no point does it overshoot.
The continued costal desaturation even post interval may also be related to the continued high respiratory rate and work of breathing. Although the legs have reduced their O2 consumption, the oxygen debt is still high and the respiratory system is still working hard.
The vastus lateralis O2, also drops rapidly from the baseline of 65% to 44% at about 13 sec into the interval, but then slowly rises throughout the remainder of the maximal effort(costal O2 is still dropping). It continues to rise with the 150 watt zone (costal still bottomed out at 5 %) but does drop again with the 250 watt zone. There is a rapid rebound thereafter during coasting to 72%.
This does make sense, from the cardiac output redistribution perspective. The initial VL drop is a result of the combination of external muscle compression reducing inflow (like weight training) and high demand. But since the muscle power is dropping with time and vasodilitation is occurring, the O2 sat gradually rises. There will be other individuals, muscle groups, fiber depth variations that will show an alternate pattern, but the bottom line is that the O2 drop is multifactorial and highly variable. Once the power backs off, the O2 rise continues in the VL, almost reaching baseline. Certainly, this cannot be construed that you are ready to hit it again. The next minute at 250 watts does show a significant drop in the VL (52%) but considering the narrow dynamic range, perhaps not very obvious.
The following is a closer inspection of the VL only, done on a different day:
The pattern is quite similar, but the scaling limits are different.
On the other hand, the costal O2 falls throughout the max interval, and stays at nadir for quite some time after (when the VL is getting refueled). During the low intensity zone, it does recover slowly, but abruptly plateaus at a markedly reduced level before further dropping again with the higher effort. Here is a major clue that the cardiovascular system is not back to baseline and further intense efforts will be potentially problematic. There is no O2 rebound after, possibly because the locomotor muscles still have priority in blood supply.
Next is a 10 min interval using the 2 sensors, also costal and VL. The VL baseline is about 65% and rapidly drops to about 48%, and stays between 45 and 48% throughout, with a rebound during coasting. In regards to the lowest the VL O2 sat can reach (in my case) that would be the 44% during the maximal 1 minute effort. But that was transient, and a slow rise was seen after, even under effort. Therefore, the VL "cruising" O2 sat of 45 to 48% seen is probably near the realistic minimum and does not add any clue to how best to optimize race pacing.
Costal O2 starts at 64%, drops more slowly to about 35% and very gradually drifts down to 30%. It certainly has a wider dynamic range than the VL, making on the spot, during the ride decision making easier since the curve slope is much more obvious. Also of note, the 34% is very close to the 35% zone seen after the maximal effort in a post recovery moderate intensity situation. So the costal O2 "cruising" O2 sat is well above the nadir seen during a maximum 1 minute effort, and it has room to spare. Watching this residual buffer zone is something that can be done while racing (or training) in real time. Certainly, a much larger sample size than 1 rider is needed to see how this holds up.
Additionally, it may be possible to extrapolate the downward costal O2 curve to predict a time vs power relationship.
From my perspective, keeping the costal O2 in the 30% area (for me) is a way to avoid rapid anerobic decompensation. The VL data does not appear helpful in this regards.
Is there any outside literature supporting this method of training/racing/pacing?
A new sensor vendor (Humon Hex) clearly advertises the following on their website:
They boldly state that O2 sensor monitoring (of locomotor muscles) can guide pacing "without worrying about burning out" as well as provide recovery knowledge. I was very curious as to how they came up with these claims, since no research paper I was aware of supports this.
The reference in their "white paper" is to "Regulation of Increased Blood Flow (Hyperemia) to Muscles During Exercise: A Hierarchy of Competing Physiological Needs". I would strongly urge anyone interested to actually read this. It is an excellent review of the physiologic interactions spoken about on the last few posts. But there is not a single reference, figure or text on any of the "unique insights" promised above. In fact a search of the document fails to even find "NIRS" present and it does not address monitoring muscle O2 sat with real time sensors at all.
In closing, O2 sensor use has some intriguing potential use in real time cycling and weight training. Unfortunately, the manufactures continue to make unsubstantiated claims that in the short term may sell more units, but will ultimately harm themselves (BSX) as well as mislead the end user.
Instead, they should be at the forefront of true peer reviewed academic research to expand their markets and be more profitable long term.