Friday, January 4, 2019

Sensor Measurement Variability - Location and Device

It is well known that muscle O2 sensor use in different locations, contexts and sports applications, will yield day to day numerical variation in results.  This fact makes it difficult to have rigid cutoffs for zone training or threshold monitoring.  For instance, on one day the absolute value for a lactate max steady state in the vastus lateralis could be 45%, but during another session may be 5% different.  This post will explore a major source of this variation, that of sensor position.

A study done over 10 years ago was one of the first to look at this issue.  They measured the O2 saturation changes in multiple quadriceps muscles during various 6 minute cycling interval sessions.  There were several important observations.

The speed of desaturation was different for various sites and there was significant heterogeneity in various subjects:
In subject B for example, the various sites had large variation in degree of desaturation and rate of desaturation which was not seen in subject A.  In fact, in subject A, except for one site (dotted line), the curves look remarkably similar.  Presumably, placement would have little impact in that person.

Observations in one subject:
Another example of the difficulty in finding "benchmark" values for training is the following tracing.  They compared all sites as well as site 3 (vastus medialis?) and site 7 (rectus femoris):

As seen above, the degree of desaturation and speed of resaturation varied at each site.
Interestingly they did remark that the time constants of desaturation were similar at proximal and distal VL sites:
Regarding the comparison of the distal- vs. proximal site
(i.e., S1 vs. S10 and average of S1–S3 vs. S8–S10), there were
no significant differences of TDp, p, or MRTp for either
moderate or heavy exercise.

Proximal vs Distal locations:

In a given muscle (VL) does a location lower (distal) or higher (proximal) affect the desaturation pattern or absolute values?  
A look into this issue was done several years ago.  The study was designed to look at proximal, distal VL as well as RF in relation to speed/extent of desaturation as well as any reserve in desaturation potential compared to a tourniquet test.
Some interesting data follows:
MOD represents a moderate pace and SVR a more intense (severe) pace.
This was calculated from the gas exchange threshold (GET) with the MOD being 80% of the GET (GET=resp compensation point) and the SVR 120% of the same GET (closer to VO2 max).
There was some variation in the baseline saturations of the proximal and distal VL with VLP lower in MOD but higher in SVR, perhaps illustrating the issue of placement, sensor calibration or day to day variation.  There were some small changes in amplitudes but kinetic timing seemed very close.

The closeness between distal and proximal was evident here:
Although there may be some statistical difference between tracings, they are very similar in desat degree and shape.  Also of interest is the rapid drop in O2 saturation, followed by a slow rise.  This pattern is very prevalent in my personal tracings.

However, the O2 "reserve" between VLP and VLD was different.  This reserve potential refers to how much (if any), desaturation ability (compared to a max occlusion value) is present at a particular site under heavy dynamic load (max cycling).
The VLD (and RF) during dynamic exercise was unable to reach the desaturation peak of an occlusive cuff.  Therefore the VLP is able to desaturate to the max occlusive value (while cycling) but the VLD is not.

 With the conclusion:
It would not be unreasonable to conclude that although the saturation changes are similar in the VLD and VLP, some fundamental physiologic difference may exist between the two.

Another important point is that the depth of sensor targeting into the muscle is also a factor in saturation as well as O2 reserves (but won't be covered here).

My data:
Armed with the knowledge of the studies above, I decided to measure the VLP/VLD saturations on 2 different days using 2 different BSX sensors.  The sensor positions were switched on the second day in case one sensor estimated differently from the other.

Day One

360 watts x 3 min (max aerobic power or VO2 peak)

 Wingate 60 (all out 60sec)

Although the baseline and amount of desaturation is different, the curve shapes and recovery are remarkably similar between VLP and VLD.

                                           VLD           VLP

Baseline                             55%          68%     baseline VLD<VLP
3 min nadir                         30%          47%
Wingate 60 nadir                31%          48%      

Max difference                    25%          21%


Day 2 - sensor position swap

360 watts x 3 min (max aerobic power or VO2 peak)

 Wingate 60 (all out 60sec)

Although the baseline and amount of desaturation is different, the curve shapes and recovery are remarkably similar (again).

                                           VLD           VLP

Baseline                             66%          60%            baseline VLD>VLP
3 min nadir                         45%          39%
Wingate 60 nadir                42%          40%      

Max difference                    24%          21%

Points of interest:
  • There was a change in comparative baseline saturation after the sensors were switched.  Initially, I thought that the distal site had a lower baseline than the proximal, but that seems to be related to the sensor unit itself (despite the same brand).  Ambient temps, skin thickness and external pressure were about the same.  One of the sensors seems to have a higher baseline than the other, day 1 VLP sensor unit was used in day 2 VLD location (68 vs 66%).
  • Absolute saturation change from baseline was similar at each site.
  • Pattern of saturation drop, then slow rise was also similar.
  • Despite the baseline numbers being quite different between sensor units, the "Max difference" (baseline minus nadir) was virtually identical on each day.
  • Both Wingate 60 and a 3 minute VO2 max interval yielded similar nadir values at each individual location on a given day.  So the nadir at the VLP will be about the same during a Win 60 or 3 min VO2 max interval.

Final thoughts
  • The literature regarding monitoring muscle O2 for sprint and endurance performance is large and increasing.
  • There are studies indicating that "breakpoints" in the desaturation curve may be indicative of physiologic thresholds of various kinds.
  • However, there will be measurable differences in saturation depending not only on which muscle is chosen, but what part of it, the sensor is upon.
  • Baseline values may differ at nearby locations of the same muscle group.
  • Different sensor units may also show different baseline saturations.
  • It seems however, that the absolute change in saturation (normalized from the baseline and nadir) as well as the pattern of change is similar within the muscle group.  Other nearby muscles may recruit differently, so we can't assume their pattern will look similar.  Either distal or proximal VL locations should show similar patterns and absolute degrees of desaturation.
  • Since real time monitoring of a locomotor muscle can not normalize at a given site, it follows that using O2 saturation info as an "on the bike" threshold, training zone or race guide will be problematic.  One may still be able to retrospectively glean valuable information however.
  • Lastly, day to day, or longer term comparisons may also be of questionable value given the variation in sensor placement that is bound to occur.  Multiple sensor units should also be directly compared and not assumed to be equal in data output. 
  • Additional data : Sensor comparison - BSX, Hex and Moxy

1 comment: