Detailed Topic Index

Thursday, June 21, 2018

Fast Start Strategy and O2 monitoring

We all know that professional cyclists (as well as enthusiasts) will go to great lengths to shave a few seconds off a given course.  Thanks to nuances in bike tech and aerodynamics many advances have been achieved.  However, another factor that is generally not thought of affecting course completion time is the pacing strategy that is employed.  Pacing choice can be looked at from two different realms, one physics and the other physiology.  Studies have been done looking at the time saved by riding a bit harder up hills than down, and it is not trivial.  From the standpoint of exercise physiology, there may be benefits to whether one rides at an even pace of power or not.  So instead of cycling at 300 watts for 10 minutes, one could start with using more power, or start with less and have the same average power at the end.  The question then becomes, which method (if any) yields the best average power with the same effort or the fastest time at maximal effort.  This has been examined in both running and cycling.  This post will address the literature and some of my personal experiences with the different power start strategies.
To begin with, I stumbled on this subject purely by accident. Recently I was doing one of my usual "drop the rider" intervals.  The first 100 second piece was done at about 400 watts and as costal O2 dropped into my caution zone, power was reduced to about 280.  Instead of stopping at 3 minutes total time, I felt pretty good and just kept the same pace until the 5 minute mark.  Despite still feeling decent at that point, the average power was a personal best by far.
After a 2 day rest, I cycled the exact same course again but at a constant power within 1% of previous.  Although I was fine at 3 minutes, by 4 minutes I was starting to suffer.  At 4:30, I was ready to quit but wanted to finish in the name of science.  The end RPE was off the charts and costal O2 near historical lowest value.

This result is not surprising given some fascinating studies that have been done over the past few years. In a comparison of fast start, slow start and even start power pacing, the fast start strategy is usually associated with the best overall average power in the range of 1 to 6 minutes. The theory behind this is that starting well above the critical power speeds up the kinetics of reaching your VO2 max. In fact riders using the fast start method will consume more O2 than the slow/even start approaches. Whether or not this helps save anaerobic resources later in the effort is not clear.

Here is some data from an early study looking at the slow, even and fast start pacing with a 3 or 6 min time trial.  The trial was ended in each instant by a short sprint.
The exact power profiles were:
One of the parameters looked at was the MRT, mean response time of the VO2 max:
The results show that the fast start protocol lead to reaching VO2 max faster as well as improving the MRT:

Total work done was improved in the 3 min trial but not the 6 min one:
More recently the same group of investigators looked at a simulated 4km time trial with either an initial all out 12 seconds or just selt paced.  They also looked at the effect of "priming", which is adding an intense warmup interval before the bout.  The priming could be somewhat equivalent to a good warm up with some hills if you are trying this out for yourself.
As before, the fast start (all out) lead to faster VO2 kinetics and more O2 consumed.

The O2 and HHb tracings of a muscle (?vastus?) were also looked at:
There may have been a quicker desaturation in the fast start and primed groups.  However the amount and degree are close to noise levels of O2 sensor day to day variation to be of practical use.

The combination of using the fast start strategy with costal O2 monitoring may be a superior technique to maximize speed and average watt output in short time trials and parts of a longer race.  There have been some interesting proposals in modeling power reserve based on physiological parameters.  Perhaps in some later posts, this can be discussed in more detail.  What I would like to do here is just present some data and hope that other folks can try this for themselves and perhaps report back if it leads to better results.

Now for some of my personal observations.  To preface what is to follow, these were not done with the method above.  The intervals were all preceded by an hour of warm up including hills.  This could be similar to a "primed" strategy but since on a practical basis, everyone is going to warm up before intense efforts it seemed reasonable to do it this way.
The course, conditions were as consistent as possible (same hill, time of day, temp).  When the intervals were done I did not have a plan of testing prospectively.  For most of the tracings, the goal was to reach a near peak average without totally wiping myself out since I still needed to ride home about an hour away.  

Therefore, another way to look at the potential benefit of optimal power pacing (and VO2 max on kinetics) would be to minimize negative factors leading to fatigue and failure to maintain performance.  Hopefully by optimizing VO2 max, anaerobic reserves can be saved for later on, as well as minimizing the buildup of fatigue associated substances.

The first set of comparison tracings to be discussed are 5 min intervals, Fast vs Even start.  To begin with though lets look at a previous "even effort" at slightly less power:
The costal O2 does drop significantly but there is only a slight drop at the end and not into a zone of concern.  This degree of drop would be consistent with continuing the ride without major impairment

The Fast start:
The initial costal O2 drop from baseline is not much different than the even start despite 100 watts extra average power.  As the O2 sat dipped below 30%, I backed off to my 10 minute pace or so.  Costal O2 rose a bit before a slight dip at the end.  The deltoid data was very similar and rose promptly after pedaling stopped.  Avg speed was 20.6The costal O2 saturation was similar to the 307 watt average and performance was not impaired after.

The Even start:
The costal O2 drops initially (although slower), and continues especially the last minute.  The last 30 seconds were quite difficult with a max RPE.  Deltoid tracing was very similar.  The costal O2 was near my nadir at the end.  As expected, the interval "cost me" significantly and I had reduced ability for the rest of the ride.  If this was a race situation I would have probably just dropped out.

The following are the tracings for the VL (a bit noisy on Fast start, also Tot Hb shown):

There does not appear to be any major difference in degree of relative desaturation.  The tot Hb does rise a bit initially in both tracings, perhaps from venous outflow compression.  It then drifts back down to baseline.  Despite the massive reperfusion and VL O2 overshoot, little is seen in the Tot Hb rise, making this a mediocre marker of flow (unless flow was going elsewhere).

What about at 3 minutes duration? 
Here I have many samples (the 3 min interval is my standard test I do 2x/week) but will only show 3 representative tracings of the Fast, Slow and Even starts.

Given that the 3 min average power is the same in all, there are some interesting differences.  As a preface, these do not represent an all out maximal effort, but a near max over the exact same hill.  Both the Slow and Even starts show a gradual decline on costal O2 sat.  Special attention is made to the 10 seconds or so after the pedaling stops with the circle in green.  In the slow start there is a continued fall even with coasting.  I have seen this with post 1 min max intervals possibly from severe acidosis, continued high work of breathing leading to persistent costal muscle use and hypoxia.  This was not seen in the Fast start.
Despite all three intervals delivering the same average power, the costal O2 status at the end is quite different.  In the Slow and Even trials, the O2 curve is downward, and would shortly reach an unsustainable nadir.  Here is a typical all out 1 min costal tracing showing what happens if the O2 keeps dropping.  
There is a steep drop (after a mild rise) and the desat continues to drop to nadir levels after the interval (green marker).  The power of 235 was the best I could do at that point and was forced to stop pedaling after 1 minute.

With the Fast trial, the O2 does drop at 2 minutes or so, but begins to come back up into a more steady state zone that potentially could continue.  A potential usage of costal/deltoid O2 monitoring is avoiding that steep drop into the untenable zone, thereby optimally judging power vs time vs ability to continue.  As mentioned above, I personally was in much better state after finishing the Fast start, confirmed by the higher costal/deltoid O2 values.

The studies do make a point that the Fast start strategy may not be effective for durations longer than 6 minutes (or even at 6 min).  I decided to take a look at this as well. Same course, time of day, temp.

My conclusion was that at similar RPE, costal O2 curves, I was able to get another 14 watt average, basically from the first 60 second fast start without harming my ability to perform an intense effort.  Incidentally, the Fast start 8 minute average above was a personal best over the past 20 years (yes I had a Powertap back then), and I was not even trying to do that.

The bottom line here is that at least for me the Fast start strategy seems to work over the span of 3 to 8 minutes potentially by speeding VO2 on kinetics.  After the fast start is over, continuing on at a stable power that keeps costal O2 from bottoming out seems to be effective.  Of course, if this was an actual race, running the O2 down at the end with a sprint or just keeping the pace higher throughout would yield a better time.

I asked my friend the cross country skier to try the slow, even and fast approaches as well.  He kindly ran 2 sets of the 3 different pace patterns on a 1 mile track.  He has done extensive physiological testing and knows his lactate numbers pretty well.  He did one set at a power associated with a lactate of about 2 and another at a power equivalent to 4 mmol lactate.  Power was measured with a Stryd and recorded on his Garmin watch.  I plugged the data in at  Both the time splits indicated by the Garmin watch, as well as manually highlighting the time under power were determined.  The average power readings were within a percent or 2 in each set.
Here are the "lactate 2" power intervals:
Both the manual measurements as well as the Garmin watch splits indicate the Fast start yields the shortest duration to the mile run.

Next, the lactate threshold run:
The Fast start "wins" by both estimations.  As in my tracings, this does not represent an all out maximum effort, but a significant amount of power near max. Measures of the speed/distance covered showed improvement seen with a Fast start protocol.  One could argue that by using the Fast start method he is getting a faster time without negative physiologic impact.

We also have some VL O2 saturation data that is of interest:
The Fast start tracing has a faster desaturation pattern that remains stable.  Whether or not this is an indicator of better O2 extraction, is hard to say given lack of knowing the blood flow (not tot Hb).  What we can say is that the curve is different than the Slow/Even pacing pattern.
Although this is only a brief look at data without rigorous statistical validation, the results are intriguing and seem to agree with the literature (except for the time of 7 minutes being longer than the study addressed).  My cycling data was done for a fixed time, the running tests were done at fixed distance.  However as shown below the new priming and pacing study did show better time to completion with priming and all out fast start:

Terrain/Wind considerations:
There is an additional practical consideration when employing a fast start strategy, namely aerodynamics, wind and climbing. In a very elegant study of physics in cycling, modulation of power based on terrain will yield tangible time trial benefits:

From the paper:
Swain (1997) used the equation of motion of a cyclist presented by Di Prampero, Cortili, Mognoni, and Saibene (1979) to model performance on undulating and windy time-trial courses. Faster times were predicted when a cyclist increased power in slower uphill or headwind sections even if this higher power was compensated for by reducing power in faster downhill or tailwind sections so that total work done was unchanged. 

Adopting a variable power profile based on terrain/wind can improve time trial results:

 And the net results in a simulated 40 km TT:
I underlined the curves for different power averages.  Interestingly, the least time saved was in the intermediate power group, the most in the weak riders.
The authors concluded that this type of intervention was potentially worthwhile compared to other methods:

An obvious extrapolation is to combine elements of fast start with terrain/wind considerations to fully optimize a time trail, break away or ascent.
Going back to my 5 min Fast vs Even start tracings (the second and third of my data graphs), the initial 2+ minutes are up a hill (encompassing the Fast start distance).  The Fast start average speed was 20.6 vs 19.8 mph with an Even start at the same net power.  Interesting.

A last comment:
A model of athletic performance was recently discussed in regards to Chris Froome's power profile in various race situations.  This model does not incorporate any power modeling VO2 alterations (especially fast start) in the equations and discussion.  There is also a significant degree of variation in each assumption that could add up to some major significance.  So when modeling such as INSCYD is presented, it may or may not be precise for the given athlete.  Using this approach to provide a clue to doping, motors and such can be very problematic in my opinion.

Final thoughts:
  • A fast start strategy can lead to faster "on kinetics" of VO2 max, allowing a higher power to be averaged over a duration of 1 to 6 minutes (perhaps longer) at max effort or the same average power with less negative physiologic impact.
  • The use of either costal or deltoid NIRS monitoring may help indicate the degree and duration of each segment of the power partitioning strategy to inform the rider how to proceed in real time.  Avoidance of a given riders' personal nadir costal/deltoid O2 range would best allow them to continue at optimal power levels.
  • Consideration of the terrain and wind conditions may also play a role in longer time trials for best race times.
  • Athletic modeling should take into account the actual tracing of power since starting strategy alters VO2 max assumptions.

 VO2 max related posts

Tuesday, June 5, 2018

Muscle fiber transformation, HIT and NIRS observations

Over the last few posts we have explored how different sports as well as variation in genetic markers for endurance/strength impact local and systemic muscle oxygenation.  In addition, it has been postulated that an individual approach to training based on those genetic markers, costal desaturation patterns may be helpful.  The large variety of training protocols and HIT routines in the literature generally show improvements in many attributes of endurance exercise.  So which is the best one for a given athlete?  And conversely, are certain specific programs potentially detrimental depending one one's status?

One of the attributes of successful endurance performance is the presence of a large proportion of "slow twitch", type 1 muscle fibers.  Baseline genetic endowment is advantageous but aggressive endurance training is potentially helpful.  To that end, multiple studies have shown that both short high intensity as well as long moderate intensity exercise can lead to fiber transformation.  Many of the interventions have shown that if there is not a true fiber type switch, at least there will be a change of the type 2x (strength, fatigable, glycolytic) to type 2a (strength, fatigue resistant, oxidative).  In a previous post we looked at genetic SNP markers that correlate to muscle fiber type.  In particular if you have ACTN3, ACE DD positivity, chances are you will have a relative low proportion of type 1 and a higher amount of type 2 fibers.  Therefore if you are born with this genetic pattern it would not be advantageous to train with a protocol that tends to increase type 2 fibers even further.  You would want to employ a method that would best make your type 2 fibers fatigue resistant/oxidative (type 2a) and hopefully some conversion to a true type 1 over the years.  On the other hand, lets say you have homozygous markers for all endurance parameters (ACTN3 neg, ACE II) and have noted that your long distance performance is fine but sprint power is lacking.  Then it may be helpful to convert some of your type 1 fibers to the stronger, fast twitch variety to improve that issue.  So the question becomes what is the best type of HIT method to either cause a fiber shift if wanted, as well as what to avoid if endurance enhancement is your goal.
A study was just published looking at a high intensity sprint training method that indeed will lead to a slow to fast fiber transformation.  This post will discuss the study, training method and a look at NIRS tracings in RF, VL and costal areas. 

The protocol
All training sessions started with 5min at rest on the cycle
ergometer, followed by a running-in for 2min at 10 Watt (W),
a 10min warm-up at 50% maximum power (Pmax), a 45min
interval phase with 90 intervals of 6 s at 250% Pmax, each followed
by a 24 s pause at 10W, and a cool-down of 5min at 50%
Pmax. Subjects were asked to keep cadence at 70-90 min−1.
Subjects performed three training sessions per week, resulting
in 18 training sessions in 6 weeks

To start with I was a bit confused by the Pmax parameter.  Is this the maximum power reached in 1 revolution (Cycling Peaks definition) or related to power at VO2 max, steady state lactate levels etc?  Well, it appears to be defined as the last 1 min stable power on a VO2 max ramp test.  This value can vary according to the ramp protocol but you get the idea.  It is not your 1 min max average power(would be very hard or impossible to do 2.5x that number).  The study used (2.5) x (the ramp 1 min peak power).  Given that I have no way of knowing what that value would be for myself, I decided that I would just do an almost all out 6 sec sprint.

The training sessions were split into 2 x 3 weeks periods, each week having 3 interval days, each session consisting of 45 mins, which translates to 90 intervals.  So, 270 intervals in 1 week!

Testing was done at various times including pre, post and during the intervention.  Muscle biopsies were done of the VL, enzymes involved in fat and carbohydrate oxidation assessed and various exercise tests done (lactate levels at various power outputs, VO2 max).

Without getting too detail oriented lets look at some of the results.  

Despite no difference in VO2 max pre and post, the ability to handle lactate was markedly improved.  This was felt to be possibly related to improved whole body lactate clearance (liver and non involved muscles).
Time to exhaustion at both 65 and 80% of the Pmax was improved:

However, the proportion of Type 1 fibers dropped significantly during the study:

Consistent with the fiber type shift, oxidative enzyme levels dropped as well:

In the discussion, other studies were reviewed that either were consistent or not with the findings.  One study of note done many years ago found the opposite, a fast to slow fiber transformation but used a training protocol much different (fewer sprints and longer rests).
This is data from that paper:

It is strange that the slow twitch area dropped while the % slow twitch fibers increased.  That said, the two studies are not comparable given the very significant difference in training volume (we will see my NIRS data in a bit to elaborate).
From the paper:
Sprint training. The sprint training programme lasted for 7 weeks
at the rate of four sessions each week. Each session consisted of
two series of maximal intermittent sprint cycling (5-s sprint, 55-s
rest) at 80% Fmax (i.e. 130-140 rpm). The rest period between the
two series was 15 min. The first week sessions included 8 sprints
in each series (see Fig. 1). This number was increased by 1 sprint
every week so that each series of the last weeks sessions included
13 sprints.

As noted the rest interval is 55 seconds, number of sprints was much less as well.

I was curious as to see what behavior both leg and costal O2 would display during the 6/24 sec training method.  My guess was that we would see little in the way of costal hypoxia but profound intermittent changes in working leg muscles since the interval time was so short.  If so, the short intense intervals may not be a good stimulus for cardiac output, contractility, venous return.  Therefore, for folks with problems in cardiac output distribution from lower stroke volumes, it may pay to avoid this type of training.  Conversely, in those who don't have a good peak power, sprint capability, transforming fibers to type 2 could be quite helpful.
Some clues...
Here is the usual 1 min maximum effort interval with sensors on the VL and costal:
Note that at about 12 seconds, the VL has it's nadir O2, while the costal takes far longer to drop.  Remember, the intense muscular contraction is also going to cause external compression (as in the weight training tracings) limiting blood flow making local muscular hypoxia even more pronounced.  Since this has been my usual pattern, I certainly did not believe that 6 seconds followed by a rest would lead to systemic costal hypoxia.

In relation the above study, one could almost compare this to doing a weight training session with heavy loads, few sets but short rests of 24 sec.

Now to the 6/24 protocol.  
This was done with sensors on the L RF, R VL and R costal areas.  Warm up was about 1 hour, ambient temp was a bit high, near 90 F.
The first tracing is that of the VL, with no real surprises, there is a desaturation on each 6 sec interval and a return to baseline quickly during the rest:

If we magnify a few of the 6 second bouts, the total Hb also drops nicely (purple curve) with each effort.  This certainly makes sense as there is significant external muscular compression (as in weight training).

The Rectus femoris also behaves as the VL, but with much deeper desaturations:
Although power is not shown in this plot, the RF/VL are essentially identical in time course.  However, the RF has nadir O2 values of near 15% and an exaggerated rebound.
To see how these values compare to a 1 min max effort (which usually produces the best nadirs) this was done later in the same ride session:
And all three sensors without showing power:
The RF and VL nadirs (15% and 54% for each) are about the same as during the 12 x 6 sec interval session.  Note that the costal drops from 67 to 4% and recovery is much delayed.

Now for the surprise.  I originally thought that brief 6 sec bouts with 24 sec rest would not produce much systemic effect.  Well, I was very wrong.  As noted below, the costal drops from a baseline of 60% to 12%.  This was after just 12 intervals, not 90.  In fact, I was fairly winded at 6 minutes and decided to call it quits.

The VL does desaturate during the 6 seconds, with a return to baseline during the rest.  However, the costal area has a notable delay and the actual nadir is well into the rest period.  In addition, as the intervals progressed, there was a downsloping of both the recovery and nadir costal O2 that was not seen in the VL.

A close up of the costal O2 also is interesting, the nadir is well into the rest and the O2 actually rises for the 6 sec effort.  I can't explain the initial rise, but that pattern is also present on the 1 min max above and on other high intensity tracings.  In regards to the O2 drop during "rest", several factors could be causing this.  The rebound blood flow rise into the legs needs to come from somewhere.  As the legs seem higher priority than the respiratory muscles, the costal area may be getting a perfusion drop.  This does not seem to be the case.  The following tracing (data points only every 5 sec, a Garmin quirk) shows the costal tot Hb rising during the rest phase (even as O2 drops):
The pattern is the same for each cycle.  Other potential reasons for the drop could be two complimentary effects.  There is both an acidosis leading to respiratory rate increase and therefore costal muscle activity may be even higher than the first few seconds.  Although there does not appear to be a drop in costal flow, these muscle are still a lower priority.

Looking at the above, the 6/24 study results now makes more sense.  The leg muscles are challenged by a "resistance training like effort" repeatedly but at the same time there appears to be significant systemic lactate elevation, acidosis and respiratory threshold effects (high resp rate, work of breathing related to the acidosis).  This is certainly a major training stimulus.
On a practical level, I have concerns of possible overtraining, as well as knee strain.  I don't think I could do the study protocol as written but perhaps if I was 25 again I could.

Where does this leave us from a practical standpoint?Should we use this training modality or not?
On one hand it improves lactate clearance, time to exhaustion parameters but on the other we convert type 1 fibers to type 2, as well as lose oxidative enzyme effect.

Training based on SNP markers and historical demographics?
Previously, I have shared some of my data and genetic markers, all clearly in line with strength rather than endurance.  Interestingly, my data also agrees with that assessment.  Over the years I have amassed a large database on that site and a graph exists that compares your best times with other athletes who use the site.  Here is what my curve looks like:
Everything less than 10 min is valid (I don't do longer intervals than that).  Note my ranking is proportionally much better at 30 to 60 seconds versus the 8 or 10 min times.  Even at 3 min, I fall off the curve despite doing my best efforts at those interval times.

Back to the study....
One important point about the study is it did not compare how other training modalities affect lactate clearance, fiber type and time to exhaustion.  This is not a criticism in any way, but for those of us who either can't tolerate such intervals (bad knees), or who are concerned with fiber shifts (that they do not want) there could be alternate interval types that may produce even the same or even more benefit.
The other question to ask, is whether or not there was a difference in training effect in each individual (good and poor responders to this method within the whole group).  If there were, it would be interesting to stratify them according to SNP markers, VO2 max, stroke volume, etc.

I think the decision to employ this training modality should be based on personal strengths and weakness.  
  • For those with genetic strength markers, evidence of limited stroke volume, performance characteristics more in line with sprinting, the above type of training may not be appropriate.  In this case longer intervals of 30 to 300 seconds, different ramp intensities may work to better enhance their deficiencies.
  • However, individuals with great genetic endurance markers, limited sprint strength, a performance curve that does not look like mine(better rating at longer duration), then taking advantage of the study technique is potentially of benefit.  
  • The concept of limiters based on local muscle O2 desaturation curves continues to have no support in the literature.
  • A better view of ones limitations should be based on genetics, systemic (costal, deltoid) desaturation as a surrogate of cardiac output limitation, demographic comparisons in power vs time.