Sunday, July 29, 2018

Trainer vs road cycling, psychological and physiologic comparisons

We know all too well the hesitation to use indoor sports equipment such as treadmills and exercise bikes.  Let's face it, they just are not pleasant to use.  I'll take a ride in the rain any day rather than get on a good turbo trainer.  However, putting preference aside, do they provide the same degree of physiologic challenges to the cardiac, respiratory and local muscular systems?  While using my own bike on the turbo trainer, I noted the perceived effort was way higher than my physiologic metrics indicated.  However, the metrics were similar to the road riding data.  I decided to look into this further and report on some literature as well as my personal measurements that help reconcile this.

First off, is my perception of a higher relative effort valid?  
A study was done comparing an outdoor 40 km time trial with an equivalent indoor simulated time trial (subjects personal bike on a computrainer).  The instructions were:
Therefore, same bike, same perceived effort but no feedback such as power, heart rate were allowed.
Results were quite interesting:

The outdoor trial was associated with a much higher average power (205 watts) than the indoor trainer (163 watts).  With the higher power came a higher average heart rate in the outdoor trial (152 vs 143).  The RPE and core body temp were statistically the same.  Therefore, despite the same bike/riding position, an indoor ride will have a drastic reduction in power at the same perceived effort.  Conversely, an outdoor ride, at the same perceived intensity, will yield greater physiologic stress and presumed benefits from the work out.  Other studies have not shown dramatic difference in physiologic metrics, but were done differently.  Either instructions were given to do a max effort, or heart rate/power was available for the subject to see.  
For example:
 The above study was done at maximum time trial effort:
"After a short rest the subject was asked to
cover a distance of 40km in the shortest time possible achiev-
ing the highest average power output. Feedback in the form of
heart rate response (b ´ min±1), elapsed time (min: sec), and
percentage distance covered were the only visual cues given
during the trials"

Bottom line:
At equivalent power and heart rate, the relative perceived effort will be higher on an indoor trainerTo reach the same physiologic targets, a much higher discomfort level must be endured.  No wonder most people don't like riding indoors!

It appears that the physical benefits can be equivalent on a trainer but only with sufficient effort.

However, what about issues such as muscle specificity and overall cycling efficiency while riding a turbo trainer?
This study looked at exactly that using EMG as well breath by breath analysis for calculations of efficiency.  The test condition was cycling on a treadmill vs using an electronically controlled turbo trainer.
From the paper:
"Pedalling technique was quantified by a novel parameter that
described the distribution of power during the pedal revolution.
The minimum power outputs during the pedal revolution (i. e.,
the dead centres) were expressed relative to the overall power
output and described as ‘dead centre size’ (DC) [ 21 ] . Leirdal and
Ettema [ 22 ] showed a positive correlation between efficiency
and DC and thus pedaling technique."

The EMG data did show some differences:
The RF was the main muscle of difference and theoretically could be an issue especially for Powercrank users (such as I), since studies have shown that the RF is more heavily utilized with that pedaling modification. What the significance of these differences would be in practical terms is not clear.
Here is a further look at mean muscle EMG activity as power is increased.  There is again differences in the RF and BF (hamstrings):


Dead center data:
And finally their conclusion:
Their bottom line:
Turbo trainer EMG and pedaling technique are subtly different than road cycling but do not appear to impact efficiency.
One condition they did not test for was the effect of a road ride up a slight incline.  The trainer ride does feels more like a mild steady incline than a flat road (the flywheel helps of course but is not the same as a road feel).  Perhaps that condition would have yielded different EMG patterns closer to the trainer?  Arguing against this simple explanation are studies showing that pedaling mechanics and muscle EMG patterns are not altered with uphill riding.   This is a figure from the above paper showing that seated uphill and level cycling have very similar muscular patterns (as opposed to standing):
On close inspection, there are tiny subtle changes of unclear statistical and physiologic significance between incline and level seated riding.

On to my data.
My original purpose in getting a trainer was to recreate the interval efforts I have been doing and then determine finger-stick lactate values over this time.  Outside blood measurements in the heat/humidity of Florida are just not possible with the Lactate Scout (error 1 always seen).
As I was using the trainer, I was struck by how hard I was seemingly pedaling but the "numbers" were not showing costal O2 drop, heart rate elevation as they should.  This now makes sense in view of the first discussion above.
Lets now look at my physiologic metric comparison, specifically Hexoskin cardiac, respiratory data, costal and leg muscle RF SmO2 vs power.

The following interval tracings pairs were done in the same session, first a road ride of about 2 hours (1 hr warm up, interval, 1 hour back), then immediately putting the bike on the turbo trainer.  Another 10 minute or so warm up is done then the same interval was repeated.

Even start on the road at 330 watts:


Measurements include 
Costal O2 dropping from 58 to 23%
L RF SmO2 dropping from 47 to 8%
Heart rate max 168
Minute ventilation at 30 seconds - 105 L/min
Minute ventilation at end 210 L/min, with transient rise to 226 afterward
Respiratory rate at 30 seconds - 46/min
Respiratory rate at end 59/min

Even start on the trainer:




Measurements include 
Costal O2 dropping from 56 to 13%
L RF SmO2 dropping from 46 to 6%
Heart rate max 167
Minute ventilation at 30 seconds - 118 L/min
Minute ventilation at end 214 L/min, with transient rise to 233 afterward
Respiratory rate at 30 seconds - 39/min
Respiratory rate at end 55/min

  • Allowing for some random fluctuation and noise, remarkably close.
  • There is some additional O2 drop on the costal area, but the hypoxic shaped curve is very similar.  
  • The RF data was pretty close as well, but since O2 extraction was near complete in both cases it is difficult to judge in my opinion.  
  • Certainly, heart rate (single lead EKG) and respiratory rate/volume were close

Next case, Fast start at 310 watts:

Fast start Road interval with bilateral costal O2:


Same graph interval but looking at R RF muscle O2:
Measurements include 
R Costal O2 dropping from 73 to 36%
R RF SmO2 dropping from 75 to 54 rising to 61%
Heart rate max 163 early in the interval
Minute ventilation at 30 seconds - 100 L/min
Minute ventilation at end 209 L/min, with transient rise to 219 afterward
Respiratory rate at 30 seconds - 40/min
Respiratory rate at end 56/min


Now for the trainer data
Bilateral costal O2 and power:


 Hexoskin:

R RF muscle O2:

Measurements include 
R Costal O2 dropping from 72 to 36%
R RF SmO2 dropping from 69 to 55 rising to 58%
Heart rate max 162 later in the interval
Minute ventilation at 30 seconds - 100 L/min
Minute ventilation at end 208 L/min, with transient rise to 219 afterward
Respiratory rate at 30 seconds - 44/min
Respiratory rate at end 50/min

Comments:
  • The R RF deoxygenation curves are very similar.
  • The costal deoxygenation curves are also extremely close to each other.
  • Both respiratory rate and minute ventilation are on par with each other, although rate seemed a bit less on the trainer toward the end.
  • Heart rate curves are slightly different but max values about the same. The heart rate rapid rise is present in both cases but there is a slow rise on the indoor trainer as opposed to a slight decrease riding on the road.
One of my continuing interests is in the Fast start pacing strategy of Dr Stephen Bailey.  A difference in technique (amount and duration of fast start) may not be the only variable when comparing my observations to that published.  They use a trainer indoors (understandably considering what they are measuring):
Would the results be different on an outside road condition?  Probably not, but I need to keep testing conditions in mind as further trials are done.  One of the major differences that I am seeing with the Fast vs Even start pacing is that of heart rate trajectory.  At least on this limited look at indoor trainer vs outdoor comparisons that differential may become blunted.

Road/Outdoor vs Trainer/Indoor cycling generalizations:

  • Perceived effort is higher on an indoor trainer.  Whatever the reasons (distraction riding outside reducing RPE), it is something to consider in attempting to simulate an outside road ride.  It also helps to explain why many of us consider indoor riding so unpalatable.
  • Overall cardio-pulmonary impact at equivalent power outputs is very similar in both conditions.  Heart rate, respiratory rate and minute ventilation were all similar between test conditions.  However they may be some subtle differences in heart rate patterns in the fast start road vs trainer.  This may also translate to lower costal and RF SmO2 at interval end.  I need to explore this further.
  • There were subtle alterations in pedaling mechanics and muscle usage with a turbo trainer.  Whether this is going to impact road riding specificity is not clear.  It is also possible that these changes resemble outside riding up an incline however literature does not support the EMG changes that are seen with this.  The RF is the muscle group that seems to most affected by trainer use.  However, NIRS measurements of this muscle did not show marked difference while using a trainer.
  • Lastly, the Hexoskin, BSX sensor combo is a helpful tool for home physiologic testing and metrics.





This post is dedicated to my son Kipp who uses a trainer exclusively (performing multiple intense intervals) and has become a gifted cyclist as a result. 




Sunday, July 15, 2018

Hexoskin shirt, review and observations with Interval exercise



The past several posts have revolved around the interaction between locomotor and respiratory muscles in relation to their respective muscle oxygen saturation during exercise.  As the exercise intensity rises, acidosis ensues and a higher rate/volume of breathing is needed to help remove the CO2 that results from acid buffering.  However even before that, at moderate work loads, more O2 is needed and respiratory rate and volume are boosted.  To better quantify this process it may be interesting to follow the respiratory rate and minute ventilation alongside the costal O2, power and heart rate.  The Hexoskin shirt has been available for several years to do just that with study data to back up it's accuracy.  It provides the following:
The placement of the costal O2 sensor is such that a conventional belt monitor like a Polar H10 will get in the way of the device.  As a result, I have been using a Polar OH1 which is ballpark at best, and worthless in a sprint.  The Hexoskin may also be the answer to better heart rate monitoring for me since I can't wear a belt and O2 sensor simultaneously.  The following post is meant as a commentary of my initial experience using it for both heart rate and respiration monitoring.

The shirt basically has sensors for an EKG, 3 axis accelerometer and of course sensors for monitoring chest/abdominal expansion.
The fit was reasonable (size large for me at 178 lbs), comfort good (even in hot weather), and there really is no software to set up except the PC app to transfer the data from the bluetooth module (which is small).  If you are going to use the android/iphone app, you may not be able to pick up the BTLE heart rate on another device (watch, bike computer) or even on the same device using a different app (Ipbike).  Since this can be a limitation of bluetooth pairing, I decided to just use it as a heart rate sensor for live data, and analyze the respiratory particulars after the ride.  Eventually I would like to get both Ipbike to record heart rate and run the android app in real time concurrently on the same device.

On my initial use I had issues with the O2 sensor moving around under the Hexoskin.  My solution was to layer a Skins shirt over the Hexoskin. A coban wrap light sensor on top of the Hexskin but under the Skins shirt was also used to block light and keep things tight.  There are 2 "belts on the Hexoskin, the sensor is placed between them, as noted in red.
 

The results:
The following tracings are from 2 rides, both in very hot, humid weather (90+).  The minute ventilation was probably calculated on the weight/height I put in (which were not accurate, I was initially in a rush and didn't bother converting into exact metrics).  However, the breathing rate would not be affected and the ventilation is comparable within my personal data (on a relative basis).

The first interval is the typical 1 min max on the same hill as previously shown in older posts.

Several comments:
Heart rate (blue) seems spot on, consistent with no drop outs, recorded with a Sony Z5 premium with Ant+, BTLE using Ipbike
Excellent costal/deltoid O2 drop.  


Now the Hexoskin data
(A potential shortcoming, at least with cycling is the lack of power metrics on the tracing.  We can get around that by looking at the Activity data (green arrows).  During a significant interval or sprint, the cadence and activity are higher as shown.)

The heart rate track is identical of course(same data).

The respiratory rate starts at 32 (was as low as 20 before that) and rapidly rises.  The spike of 63 may be real.  At that point I was getting very winded and probably taking shallower breaths at too high a rate for maximum effectiveness.  Interestingly, the minute ventilation rises relatively smoothly and is highest after the interval stops.  No wonder the costal O2 stays down even with low power (150 watts) and immediately after ending, the respiratory muscles are working quite hard.  From previous observations, the leg muscles are rapidly resaturated post interval, but they are now resting whereas the respiratory muscles are still at maximum load.  There appears to be no "coasting" for the respiratory system!
Take home point: 
This may be a good illustration of what happens when respiratory muscle O2 drops below a certain point leading to less than optimal muscular efficiency as well as the prolonged time to recovery after such an event.

Several days prior I did a similar 1 min max interval but the O2 sensor moved.  Here is the Hexoskin data:
Unfortunately I had a difficult flat and tire change (60 min walk to get air-a long story) well before this point and was probably dehydrated leading to the high heart rate.  
Again we see a spike or 2 in respiratory rate without corresponding bump in ventilation.  Although artifact is possible, I think it is from shallow, fast breathing.  Minute ventilation smoothly rises despite that.
(This was also my first time using the shirt and band fit slightly off.)

Also on my first outing, a 5 min Fast start interval was done (very similar to previous posted data).
The costal O2 drop was a bit higher than usual, either related to the lower baseline, or from the dehydration.  Regardless, it was relatively stable as was the chest/pectoral O2.  Notice the last 30 sec or so, power was boosted and costal O2 took a nosedive.

Hexoskin data:
The activity (in yellow) gives a good guide to the start, stop of the interval.
Heart rate rapidly and smoothly rises, falls a bit after the power backs off then rises at the end as expected (higher power).
RR starts at 21 (in the box), rises to 49 after the fast start is over, but continues to rise (slowly) throughout.  
Ventilation follows the same path, a gradual rise, peaking at the end, then rapidly drops after coasting. 
It does not appear there is anywhere near the degree of acidosis as in the 1 min interval where respiration rate and volume remained high (and was highest) after power was turned off. 
Also of interest, the RR and ventilation were not higher during the initial fast start segment.
It will be interesting to redo the even and slow starts to compare further.

Several days later a 3 min Fast start was done (also same place/hill).

I also put a sensor on my L costal area (all previous costal measurements are on the R) to double check accuracy since the Hexoskin may cause some motion as discussed above.
It was really nice to see that the tracing was within 1% of the R side!
In addition, there is a relatively smooth leveling out of the desat after 2 minutes (especially on the deltoid).
Will this reflect on the Hexoskin data?
The green arrows (activity data) indicate the interval. The start of the fast start intensity is manifested by the green cadence line.
The respiratory rate smoothly rises as does the minute ventilation.  The absolute value of the minute ventilation was near my historic max despite the respiratory rate being well below historic max (see above 1 and 5 min tracings).  Could this be an example of more efficient usage of respiratory muscles?  Since hypoxia of the respiratory muscles was minimal, perhaps the muscular system was more effective.
Something to look at in the future as well.

Usefulness for "next interval readiness"
One of the claims of both current muscle O2 sensor vendors is that by looking at leg muscle oxygen kinetics one would have a clue as to when to "hit it" hard again. That's misleading on many levels (previously covered) and certainly not supported by any literature.  But what about looking at respiratory parameters?  That may make more physiologic sense.  If you became acidotic, plus high O2 debt, your rate and volume will be high after the effort.  Although there are many other causes of fatigue (especially central factors), having your ventilation volume and rate near baseline probably heralds return to physiologic baseline (if that was what you are aiming for).
Here is a look at post interval data, first from the 3 min fast start:

Several interesting observations.  The costal O2 promptly returns to baseline, as does heart rate.  The respiratory rate and ventilation also return to "active baseline" within a couple of minutes, but it does take a full 10 minutes to get back to resting (coasting).
It seems that time to readiness would be relatively short in this case.

The one minute max data is more extreme:
The minute ventilation stays higher for a very prolonged time (O2 debt, acidosis) and even 8 min later despite almost no pedaling is still high.  Breathing rate does not appear to so affected, therefore the per breath volume must be higher during this time.  Time till next interval/effort in this situation would be far longer (if you wanted a more baseline condition to start with).

Now for some zoomed looks at the raw EKG data.
This is a tracing during coasting down a hill (before the interval-see orange circle on bottom), but still outside, in the heat and real world:
I was extremely impressed with the quality!
This is as good as a tracing done in an ER or hospital.  Well formed QRS complexes and stable baseline.

There is a bit of deterioration at max efforts(bottom circle in orange) but considering how I was moving, still very good.
The chest movement is also seen and appears physiologic, regular with clean contour.

A last tidbit
After my 3 hr ride in 100+ heat index conditions, I put the bike on a trainer (Elite Muin) and tried to do a 3 min interval with a lactate measurement(a future series of posts).  Unfortunately, I was pretty beat and could only do 2 minutes but here is the data:

Hexoskin data:
I circled the 3 min, 1 min and trainer intervals in green on the bottom time line to give an index of timing.
The activity sensor gives us the start, stop of the interval even on a trainer.
Heart rate started higher than usual and rapidly hit near max (dehydration, vasodilatation).
Both minute ventilation and resp rate smoothly rose, similar to the 3 min fast start (but this was an even start) and costal O2 fell relatively rapidly.  Lactate was 4.6 mM/L within a few seconds of finishing. I don't want to make too much of this, except as proof of concept.  I think the limit was more central fatigue with some component of dehydration mixed in.

Thoughts about Carre Technologies 
One of my major gripes of muscle O2 sensor companies is the over the top, distorted and many times baseless claims about what their product will do for you.  Since I covered this in prior posts I won't dwell on specifics.  How does the parent company Carre Technologies position the Hexoskin?  Is it the next miracle product that you must have?  Is there any innuendo that one will have some breakthrough in fitness, training, racing?  In one word NO!
Let's look at the web page:


What they claim is what you get.  The honesty is refreshing.  I also had several followup emails with tech support for helpful suggestions, background info and details of the web interface.  I tremendously respect this approach and hope that other outfits take note and follow this example.







Some concluding remarks:
  • The Hexoskin shirt accurately measures heart rate, HRV via a single lead EKG method.  Wave shape even with high rates and motion (with some deterioration at high activity) appear physiologic.
  • It has a 3 axis motion sensor, useful for figuring start and stop of interval efforts.
  • Breathing rate is measured and available either in the Phone(IOS, android) app or after download of the bluetooth module.  I have no benchmark for accuracy but studies have shown it is quite accurate.
  • Minute ventilation is calculated via the breathing rate and chest/abdominal expansion sensors.  It would be internally consistent in one person but conversion algorithms are needed for absolute values.
  • Looking at respiratory parameters alongside heart rate and costal O2 is of potential value.  One scenario is obtaining a better idea of steady state exercise intensity. The optimal values of both costal O2/power outputs along with breathing rate, minute ventilation when they are at their most efficient would be helpful.  As noted in comparing the 1 min vs 3 min fast start intervals, end respiratory parameters are very different.  Very high breathing rates with less than peak minute ventilation may correlate with reduced neuro muscular efficiency.  It will be interesting to compare the Fast/Slow/Even start protocols again with this approach, looking at parameters through the effort.
  • Another use could be in knowing when you are ready for another hard effort, at least from the respiratory standpoint.  We know from previous observations that leg muscle re-oxygenation happens quickly but costal is much slower.  This is partly due to the higher respiratory rate and minute ventilation, both of which can be observed with the Hexoskin in the field.  As illustrated above, the post 1 minute max effort with severe costal hypoxia leads to a prolonged period of high minute ventilation.  This can help plan proper rest periods between bouts, as well as provide potential data on improvement in fitness (shorter time till normalcy).  In addition, looking at post interval data of Fast, Slow and Even starts may also be different.
  • The parent company does not make any unproven claims which should be applauded.


Next: Can the Hexoskin be used as real time training tool during a road ride? 

See also:
Analysis of Hexoskin binary RR interval and respiratory .wav data 

And: An overlay for the Hexoskin app to overcome some of it's drawbacks