Detailed Topic Index

Saturday, May 12, 2018

XC Skiing from the NIRS perspective

One of the most metabolically intense, cardiovascular demanding activities is the sport of cross country skiing.  The athlete is using upper body, trunk, abs for propulsion while poling with continuing leg motion with the skis.  To some extent it is the perfect storm for cardiac output redistribution since there really is no safe muscular site to withhold blood flow to.  Extreme VO2 maxs, high respiratory flows have historically been recorded which is not surprising.  Although there is some NIRS data published during a XC ski session, there has not been very much.  I recently was fortunate enough to view an athletes data and although there are perhaps more questions than answers, the results are fascinating. I would like to thank "XCSkier" for sharing his data and especially educating me about the sport.

First off, for those who are not familiar with the sport, there are several permutations to training.  There is a classic, skating and poling variety.  The poling involves using the arms, trunk to "push" along especially uphill.  The double poling up a hill leads to intense muscular demands since upper and lower body are working hard.  

Let's look at the literature first and see what has been published with respect to muscle O2 monitoring.  I think the most appropriate paper was done on British bi-athletes.
The 2 best members of the team were monitored over a course with sensors on the VL and triceps (Portamon).
And this is the NIRS data:
Despite the fact that each subject finished with virtually identical times, the desat/resat patterns have some differences.  But lets look at the 4 segments one by one.

Double poling/flat-arms only: Both subjects desat the triceps to a similar degree. Note the saw tooth pattern, is this related to each individual contraction?  No major changes to the legs.
Slight incline-arms and legs: Subject B has deeper desats to the legs than A, and looks like the triceps are more affected as well.  Less of a saw tooth pattern in B.  If this is related to single contractions, perhaps A is able to reperfuse better between strokes.
Downhill- no muscular effort/in tucked position:  Both subjects reperfuse the triceps well but for some reason A does not show a rebound in the legs.  The authors made some comments about the lack of resaturation in the downhill part of subject A:

There may be an alternate explanation.  If we look at the exertion immediately before the downhill, subject B had a desat of about -8% but subject A did not desat as much, about -4%.  Therefore not as much compensatory vaso dilatation would have been expected in subject A and he did not resaturate as much.  As has been shown elsewhere, the deeper the ischemia, the more pronounced the overshoot (subject B).

Steep incline- both arms and legs presumably more intense than the slight incline:
Subject B desaturates quicker, and deeper than A, also with similar elapsed times.  Whether this is related to stroke volume difference, efficiency, alternate muscle usage is unclear.
Regardless, it is interesting data and shows that we should be careful in how to interpret data and expect between subject variation in desaturation.

Let's now look at tracing of our friend XCSkier.  
A Moxy sensor was on the RF(1st sensor) and chest(2nd sensor), (same placement as recommended a few posts ago for respiratory muscle monitoring).
There are 4 intervals, 1,2,4 were easy, #3 intense(high heart rate):

The red is heart rate, green RF and background gray is chest.  There certainly is significant RF desat, bottoming out to zero on the hard section.  The chest results are definitely less prone to lack of low end dynamic range, they do get low but still have some range left.  When I originally looked at this, my guess was that the sensor was not reading accurately.  In sports such as running, cycling RF desat to zero is seldom seen.  However on further reading, XC skiing leads to the highest leg desats recorded.  Upper body desats are significant as well.  Nonetheless, if these figures are correct, this indicates that our friend probably would have difficulty keeping this pace for very long.  
Looking at segment #2, the heart rate is moderately high and the RF has a consistent downward drift, chest does desaturate but remains stable.

XCSkier then went on to do a lactate test on a treadmill.  This was his description and values:

It was a a continuous effort. Every 4 minutes or so, I jumped off
the treadmill to get my lactate values taken. The very last stage I only lasted about
a minute. So, I do get about 15s-20s break after every stage when the lactate sample was taken.

Lactate values were:
rest  1.39
1       1.25  (1:00)
2       1.21  (5:00)
3       1.40  (9:00)
4       1.90  (13:00)
5       2.61  (17:00)
6       4.25  (21:00)
7       7.41  (25:00)
8       8.40  (26:30)

So at about 21 minutes he is in the 4+ lactate range, which corresponds to the inflection points of the O2 desat curves in both RF and chest(ballpark estimate, no math involved):

RF O2 sat does reach zero at the end of the test (without the competition from arm use in poling).  But with simultaneous arm and trunk competition, would the point have been different in terms of heart rate?
Lactate >4 occurs well before max heart rate.
RF rebound greater than chest O2 rebound 
Dips in RF at each pause for lactate sampling noted(?)

Next set up with a Moxy on the VL:
Five intervals:
1. Easy skating
2. Fast skating
3. Easy double poling
4. Fast double poling
5. Easy skating

And the Tot Hb, O2 sats:
I drew in the green lines showing the more rapid desats in the 2nd and 4th interval of Fasts.
The difference between the 2 Fast bouts was the rise in total Hb on the 4th, presumably from muscular compression of venous (not arterial) outflow.  This could mean more intense VL contractile force was present.

The competition for available blood flow and O2 supply between multiple active muscles and the respiratory system is complex.  Instead of the previous example of cycling legs vs costal readings, we are faced with leg vs arms/trunk/abs vs costal (respiratory) in a constantly changing mix of use case scenarios.  When reviewing this data and the subject of skiing, I was overwhelmed with the multitude of permutations that the athlete could employ to achieve the same end goal.  Some may use more upper vs lower body or perhaps be more efficient in either upper or lower body (and even not realize it).  
What we can say though is that if muscle O2 is continuing to drop to non sustainable levels, it is time to back off (or hope for a downhill).  The site chosen to best monitor these changes is unclear.  Do we look at the triceps, RF, VL or chest/costal.  I know for myself, in the case of cycling, I have had the "best" guidance with the costal location.  Unfortunately because of technical reasons the Moxy has trouble at that site and we used the lateral chest wall instead.  This area should have been minimally affected by poling but that is not proven.  Further testing, especially of longer times (>10-20 min continuous effort), which would best leverage the ability of the sensor to inform about exceeding lactate tolerance may be helpful.  For example, in my case, sensor usage would help me pace above my lactate steady state just long enough to get a distance gap, but not get into decompensating threshold difficulty.  Could that be useful here as well?  Conversely, in a 1 hour time trial (no power meters) of cross country skiing, could sensor readouts (pick your site- leg vs arm vs chest) better guide you to your best possible time.  This would include uphill sections where you would be in the red (but avoiding severe decompensation), downhill recovery (maybe with some muscle assist if O2 sats are rising) and long flat sections where you would want a steady curve without downward trends.

On a practical note, the use of an android device with Ant+, Ipbike software(it has text to speech capability) and bluetooth headphones could give you an audible sensor readout on a real time basis while training and racing.

So does this help our friend in training and racing?
If we go back to the published literature there is still as of this date, no answer to the question.
A recent review article on muscle oximetry in sports made the following conclusions and recommendations:

However even with the above qualifiers, we can clearly see well defined and reproducible patterns in the above tracings.  Cross country ski related muscle oximetry characteristics will understandably not resemble those of cycling, running or weight training.  I commend and encourage further use of this modality even knowing it may just be of academic use. 
Certainly much more data is needed.

On a theoretical basis, cross country skiing could be a prime use case for costal O2 monitoring, given the variable muscles used, intense overall efforts, high muscular O2 extraction making active muscle site observation problematic for interpretation.  

But here may be an example of cardiac output redistribution nonetheless:

1. Easy skating
2. Fast skating
3. Easy doublepoling
4. Fast doublepoling
5. Easy classic (just striding)
6. Fast classic (just striding)
7. Easy biking
8. Fast biking
9. Easy running
10. Fast running
With Moxy on the RF
And associated leg EMG data:

If we ignore the first easy skate(lower O2 possibly just a first set phenomena), one could say that the easy DP leads to more blood flow diversion from the legs(because of high arm usage), hence the lower sats.  The easy DP also does not have higher EMG signals from leg activity making this plausible.
The fast bike does lead to a near zero O2, another example of the amazing O2 extraction capability XCSkier has.  
The fast run has the highest HR, highest EMG but not the deepest O2 desat!  This certainly underscores the pitfalls to simplistic interpretation on NIRS data.

Given my original criticism of the use of leg sensors for race pacing, I wondered if in this case, he could use the leg in this context.  The next tracing was a 30 min time trial, associated with a lactate of 7 at the end, Moxy on the VL.

The O2 is really pretty steady, with each O2 dip associated to a Tot Hb rise - orange markers (outflow compression?).
Given that we don't have a power meter to use in cross country skiing, using this for pacing would be interesting (Ipbike text to speech).
Note: that the HR is always rising, which is less than helpful in how hard to go.  Given that the course was completed with a highish lactate, he could probably have gone an hour with an O2 sat just a tad higher.  I hate to say it (given how critical I have been of sensor usage) but this is compelling as a pace modality in this sport.

Bottom line:
  • Cross country skiing is an extremely demanding sport from multiple standpoints, without any established power measuring devices.
  • O2 extraction rates are among the highest seen, and very low numbers are probably not artifacts.
  • Muscle O2 monitoring (especially through text to speech) may have potential usefulness while racing/training.



  1. I would classify myself as an experienced and well trained endurance athlete, but I'm a newbie when it comes to muscle oxygen. I do SkiErg ( on a regular base.

    Where would you place the sensor (I have recently bought a Humon Hex)?

  2. That is a very neat device!
    Placement depends on what you are trying to monitor. Many of my posts focus on costal O2 to give feedback for race pacing, which reflects cardiac output redistribution and probably respiratory compensation point/work of breathing. However, if you are trying to optimize the amount and duration of hypoxia to the working muscle, I would aim for the active segment you are training. So triceps, lats, maybe some deltoid. It's easy enough to do some short intervals on each and observe the curves. The idea would be similar to weight training with light loads - find the best combo of effort vs time to cause the optimal hypoxic stimulus.
    Of course, short intense bouts have their own advantages, but you wouldn't need a sensor for that.

    1. I think the costal thing could be more something for me, as I'm more a endurance (I do SmartBike, SkiErg and RowErg in different combinations) guy; not so much a sprinter or big strong guy. In the range of 55-95% FTP I'm pretty fine/good. Above threshold (105-110% is ok'ish, >110% I suck) my breathing becomes my limiter, I think.

      I've read most of your posts, which I did like very much. Your doing really great work with you blog. I like you approach of not talking marketing BS, instead you explain/show what's possible (or not ;)). #kudos

    2. Thanks for the kind words.
      One thing that you may want to consider is using a Powerbreathe RMT. The literature definitely is favorable, it's easy and if you are getting some breathlessness at high intensity it should help. My last post discusses the RMT process in detail.

    3. About kind words: You deserve it. Really. :)

      About post discusses the RMT: I'll give it a read.