Wednesday, March 29, 2017

Some thoughts on running economy (or, don't run fast!)

Everyone knows that running faster is harder.  It just feels less efficient: breathing is labored, stride mechanics feel a little jaunty, and (if you go hard enough) lactic acid accumulates.  But dogma has told us that, per mile, there is no difference in energy consumption at various paces.  That is, if it costs you 120 calories to run a mile in 8 minutes, it also costs 120 calories to run it in 5 minutes.  While that 5 minute mile is very hard and feels inefficient, it is 3 minutes less effort than the 8 minute a way, that 5 minute mile just felt inefficient in part because you were incurring some oxygen debt, to be paid back later-- say, in those 3 minutes of extra running you're not doing.  This makes some intuitive sense.  While there are many physiological and biomechanical differences between these 2 efforts-- anaerobic vs. aerobic energy contributions, fat vs. carbohydrate usage, etc-- perhaps both efforts really do cost the same in terms of calories.

In spite of the logic above, I've always thought this was at least a little wrong.  Machines have more and less-optimal operating speeds.  I've heard that cars are most efficient at 55 mph*...why would we expect a nature-built machine to be any different?  Four-legged animals also have more-and-less efficient paces, and again, we shouldn't expect humans to be different.

*not that it stops me from speeding, because efficiency isn't always key when I want to get the hell to my destination.

This quasi-informed hunch has been vindicated by several studies over the last few years.  Here's what we now know:  Faster is costlier.  Yes, just as you might expect, you are less efficient at harder efforts.  This is probably true for most endurance/aerobic physical activities-- it's not just due to less optimal biomechanics that come into play at faster running paces.  By "efficient" I mean "cost in calories to run a given distance", also known as "cost of transport" or "cost of running", used interchangeably with "running economy".  Let's go with CoT, for "cost of transport". Let's delve into the research.

1.  This 2009 paper by a biologist who studies locomotion in animals and humans was, I think, the first to offer good evidence that cost of running varies with pace.  In fact, she found a "curvilinear" association between pace and CoT: each study subject was less efficient at slow and fast paces, and had a sweet spot pace. The graphs below show the CoT curves for 6 of the 9 study subjects.

from Steudel-Numbers and Wall-Sheffler, 2009.

For me this too makes intuitive sense.  Slower paces should also be less efficient--for one, because efficiency in running largely comes from elastic energy storage (and subsequent return to foward motion) in muscle and tendon; very slow paces, and indeed walking (and uphill running! the subject of another post) benefit very little from this elastic-energy-return effect.  For me, the problem with this study was in its ultimate conclusion.  As this was published in the Journal of Human Evolution, it had an evolutionary argument to make.  In brief: persistence hunting, in which human hunters run prey into heat stroke, works in part because humans can choose from a variety of paces, without penalty, while prey animals are constrained to a select few paces.  The authors interpreted their results to mean that early humans probably did not engage in persistence hunting because we cannot, in fact, select any pace we choose with equal efficiency.  This makes persistence hunting slightly less efficient than we had thought, they argue.  Well, the energy difference between paces is quite minimal, and on the caloric scale we're talking about for a persistene hunt (1- 2 thousand calories, ballpark) the extra caloric cost of running less optimal paces during a hunt is a drop in the bucket*.  Here, I think, the authors took really cool data and forced an evolutionary significance onto it that didn't fit.

*If persistence hunting worked at all it was because the reward-- a 100,000 calorie animal-- far outweighed the cost.  And anyway, persistence hunting, if it happened (and like modern hunting techniques by modern hunter gatherers) probably failed quite often, and was really a high-risk venture made possible by the more reliable foraging efforts of others (women?) and serving in part as a display of vigor. 

2) One or two other studies found a similar result- that running faster cost a bit more.  Shaw et al (2014) asked a related question (and definitively answered ours as well): can running economy, or CoT, be accurately assessed by measuring oxygen consumption, as many studies have traditionally done?  The thinking of this particular dogma is that any oxygen you take in will be used in the oxidation of carbohydrates and fats, and it's all used-- every drop.  This part is true.  But as studies using oxygen consumption to get at CoT routinely found no difference in CoT based on pace, clearly something was missing.

The enormous graph above shows the CoT at 4 differnet paces, slower to faster (left to right).  Over a hundred trained runners were tested and these graphs show the range (the bars) and the average (middle points on each line) values.  The top graph shows actual energy cost of running, which they calculated by analyzing CO2 exhaled during running.  From this they calculated the proportion of carbohydrate vs. fat that was used as fuel, and then worked backwards to figure out actual energy cost.  Cool.  See how it goes up with increasing pace? Faster is more expensive.  The middle graph shows oxygen consumption.  No change with increasing pace, just like every other study using oxygen consumption as a proxy for energy consumption has found. Why wasn't oxygen use going up, given that cost of transport was?

Look at the bottom graph, which shows "RER", the "respiratory exchange ratio".  It's a measure of how much CO2 vs. oxygen is exhaled.  This ratio changes based on the proportion of carbohydrates to fats that are being used to power the muscles.  As exercise intensity increases, a greater proprortion of carbohydrates is used, and relatively less fats are used.  (This is a known thing and it happens because, during hard exercise, energy is needed faster, and carbohydates oxidize faster than fats.)  An RER of 1.0 means that all muscle power is coming from carbohydrate oxidation...and at the harder pace, at the far right of the graph, runners' RER approached 1.

One more piece here.  While fats are a great energy source--they actually net a greater energy return than carbohydrates, and they can be stored in excess all over the body, making them great long-term fuel-- they require more oxygen for their breakdown, into water and CO2 (this is called oxidation).  Why is this?  On the right is a simple carbohydate called glucose, and at left is a fat molecule.  

The black atoms are carbon.  The carbohydate, glucose, has a 1-1 ratio of carbon to oxygen, the red atoms.  Basically this means that, to break down glucose and turn it into water and CO2, you only need one more oxygen per carbon.  Boom, fast energy!  The fat, however, has way more carbons than hydrogens, and each of those carbons needs 2 oxygens-- from the air you inhale--in order to be broken down for energy.

Back to the study.  At faster running paces, carbohydates become the predominant fuel source for muscles.  Because carbohydrate breakdown requires less oxygen than fat breakdown, the shifting carb/fat ratio with increasingly hard exercise presents an unchanging oxygen consumption per mile, masking the fact that energy consumption is increasing.  Here, then, is why dogma told us for decades that it cost the same to cover a mile fast or slow.

So... you are right to feel that different paces are more or less efficient! (Noting, please, that energy differences between paces are tiny.)  Does any of this have any bearing on evolutionary questions?  I'd say likely not. Running is enormously costly, regardless of pace, and for early humans living on the margins of energy balance, such a behavior would have needed a damn good purpose, just as all animals who move a lot have a good reason for it.  I have some ideas about that...

Thursday, January 26, 2017

The Thrifty Human Machine: "Constrained Total Energy Expenditure"

(Now we'll see if the small and mixed readership of this blog can tolerate my continued trend towards an esoteric and narrowed focus.)

In the past decade, Herman Pontzer of Hunter College in NY has made enormous contributions towards our understanding of how human metabolism is tuned. His work from the last few years is presented in digestible fashion in the latest issue of Scientific American (article here ) and, if you find this post interesting, I highly recommend you try to get past the paywall and read this thing.  Below I'm going back to the research he summarizes in this Sci American article to make more sense of it and consider its evolutionary relevance.  Note: I love Pontzer's work so may this critical deconstruction not be misinterpreted as anything but a thought exercise from a fan and grateful student of evolutionary physiology.

"Constrained total energy expenditure"

Ponzter's big contribution is the finding that, contrary to the widespread belief that exercise boosts your metabolism, it actually does the opposite.  How did he reach this anti-dogmatic conclusion?  By carefully measuring the total energy expenditure (TEE) per day in various groups of people of varying activity levels.  For some extra evolutionary relevance and panache, he gathered data from the Hadza, a traditional foraging peoples who we like to think live in a manner similar to our earlier ancestors.  Compared with data from farming peoples and modern Westerners, here's what he found:

What matters is total energy expenditure per kg. of body weight (a 6 ft 5, 300 lb man will burn more calories than me-- but that doesn't necessary tell us anything about his metabolic rate, just that he has more tissue consuming energy).  The solid line in the above graph shows the regression for TEE vs. body weight in Western (American) men; the dashed line is for Western women.  Notice that if we were to make such lines for the Hadza men and women (the closed and open red circles, respectively), these lines would be similar to the Western ones.  Per kg of body mass, the Hadza aren't burning many more calories than Westerners, despite having more a more active lifestyle than we do.  A puzzling finding.

A similar phenomenon was found with a big study looking at the activity habits and TEE in Americans.  Let's make sense of the graph below: 
from Pontzer et al., 2016
Look at the top, black line.  This shows how total energy expenditure (TEE) was related to how physically active the subjects were; physical activity was measured with little devices that record movement.  This is a crude measure-- "counts per minute, per day".  More on this later.  Anyway, notice that for the most active people in the study, the line stops climbing and levels off.  There is NOT the linear relationship between activity and total calories burned that we'd expect.  At higher activity levels, people are burning less calories than you'd expect.

The key finding from these 2 studies is that doing a moderate amount of exercise does not significantly increase how many total calories you burn in a day (TEE).   This was true in Hadza vs. Americans, and within the American sample too.  This is counter-intuitive.  Pontzer hypothesized that our bodies somehow compensate for the energy we use during exercise by reducing energy spent elsewhere in other body tissues.  A reduction in basal metabolic rate (BMR; the calories needed to run the body, without any physical activity) has been observed in trials of people doing moderate exercise, and Pontzer's data is a mega-confirmation of this.  Where exactly this reduction in BMR comes from is unknown, but in women, we know that ovarian and estrogen functions are lowered with exercise.  Likely, the body's metabolic engine tunes down energetic needs just about everywhere in response to a moderately-active lifestyle.  Pontzer calls this "constrained energy expenditure": Humans (and other primates, and maybe many animals?) have evolved through natural selection to conserve calories by turning down overall energetic needs when activity levels increase.  

from Pontzer et al., 2016.
This is a big finding, and actually it makes a lot of sense. It's especially interesting to me because primates overall are characterized by a slower metabolism (lower BMR, lower TEE) as compared with other mammals.  But humans have reverted back to the more primitive mammalian characteristic of a hotter engine: we have a higher TEE than other primates (see graph to the right).  So we are unique among primates in that we are high-octane.  But, it still makes sense that our bodies respond to moderate activity by reducing calories spent elsewhere. We evolved to require and burn lots of calories--more than other primates-- probably because a) we have big brains;  and b), as I've been harping on for a while, we are endurance creatures.  We evolved to be active, to cover lots of ground, for whatever reason (and these reasons are hotly debated).  But it makes sense that, in response to exercise, our bodies will reduce energy expended elsewhere.

Together this work has been interpreted to mean that exercise is a rather crappy way of losing weight, and that our sedentary lifestyles are not to blame for diabetes and obesity.  Rather, it's that we eat too much.  Mostly that makes sense: someone who hits the eliptical for 30 minutes 3 days a week is only burning an extra few hundred calories per week, calories that are a) probably made up for in extra food intake and b) compensated for by this "constrained energy expenditure" phenomenon.

What's missing?

As awesome as this new finding is, there are a few caveats, I think.  First, while clearly our bodies do reduce overall metabolism to compensate for exercise, we also compensate behaviorally.  I find myself unwilling to move much at all after a 2.5 hour trail run. (This is changing now that I consciously choose to remain active despite being tired from training.)  Ponzter's group made similar observations in Americans and Hadza: we lie down more, move less, and are generally lazier after we exercise.  This helps to account for the fact that active people don't burn as many calories as you'd expect.  Pontzer notes this in his papers but I think it's a bigger factor than is admitted.

Also, look at the 2nd graph.  "Counts per day"-- the number of movements picked up by the accelerometers worn by subjects-- is only poorly reflective of activity level.  See the red data points at the bottom of that graph? They show the relationship between counts per day and calories actually burned through physical activity.  The relationship is very weak-- less than 1% of the variation in calories burned can be explained by counts per day.  The Pontzer group touts this as evidence for their constrained-energy hypothesis: higher activity level does not equal proportionally-greater total energy expenditure.  Perhaps, though, this is evidence that their activity tracking system doesn't accurately measure physical activity.

More importantly, notice who's not included in these studies-- a group with potential evolutionary relevance: serious endurance athletes.  Look at the top graph again.  That black circle?  That's me-- I added roughly where I fall for an average training day.  As a fairly typical serious endurance athlete, I would be the highest calorie-burner in that study.  Even in the bigger study (see the 2nd graph) an endurance athlete would be the biggest calorie spender... these people were not included in either study, and I strongly suspect that a bunch of data points from us folks would make that body mass/TEE line look a lot more linear.  That is, while the body clearly compensates for activity by reducing energy spent elsewhere, there's only so much compensating the body can do when you're burning 1700 cal/day through exercise.  I think that the constrained energy model is, well, constrained at higher levels of activity, even though the studies didn't test the right people to demonstrate this.  What's the relevance of this?  Well, moderate exercise may not be good for losing weight, but 12 hrs/week of training certainly is.  And, because I think that early humans were (at least at times) more active than anybody in these studies, we're missing a piece of the story.

In short, these studies didn't include people who are highly active.  Even the Hadza only burned a modest number of calories per day through activity, and the most active Americans burned only 200 more than sedentary Americans.  That itself is a finding-- that hunter-gatherers don't really work all that hard physically (in terms of calories)--but it does stifle the main finding a bit, in my opinion.

"Life is essentially a game of turning energy into kids."  A great quote from the Scientific American article.  This recent work on metabolism is shaping our understanding of how our engines are tuned, no doubt...I hope future research here includes my relevant-if-weird subgroup of humans. 

March 2017 update: Pontzer has published a fantastic new paper on this topic with a bit more synthesis and discussion of evolutionary implications:


Pontzer, H., Brown, M. H., Raichlen, D. A., Dunsworth, H., Hare, B., Walker, K., ... & Plange-Rhule, J. (2016). Metabolic acceleration and the evolution of human brain size and life history. Nature.

Pontzer, H., Durazo-Arvizu, R., Dugas, L. R., Plange-Rhule, J., Bovet, P., Forrester, T. E., ... & Luke, A. (2016). Constrained total energy expenditure and metabolic adaptation to physical activity in adult humans. Current Biology26(3), 410-417.

Pontzer, H., Raichlen, D. A., Wood, B. M., Mabulla, A. Z., Racette, S. B., & Marlowe, F. W. (2012). Hunter-gatherer energetics and human obesity. PLoS One7(7), e40503.

Wednesday, January 11, 2017

The Man Bun: Energetic and Thermoregulatory Costs.
That's right-- it's time that everyone's burning question is finally explored: what are the costs associated with having a man bun?  Not the costs to his social standing, which are mixed, nor his manhood, also mixed.  But rather, energetic and thermoregulatory costs, particularly while locomoting.  I fully expect an NIH grant to explore this further.  You're welcome, science.

During particularly sweaty workouts I have questioned the wisdom of having extra hair attached to my dome.  Is it trapping heat?  Preventing sweating?  Adding excess weight? Let's try to get at some answers.

1. Energetic consequences of the man bun

All that extra hair only adds a trivial few ounces of weight.  It's not going to make much energetic difference to the whole-body cost of running.  It's like wearing slightly thicker shorts-- a few ounces there isn't going to add up to much.  However, the head moves around a bit when you run, so these extra few ounces on top of the head could count for more. Unlike extra weight at, say, the waist, which has to be carried forwards with each stride (plus a bit of up and down), weight at the end of a bouncing extremity incurs an extra cost through every cycle of its independent movement.  So a bouncing man bun is like wearing a heavier pair of shoes: there's a cost to carry the weight forwards, like any body weight, but also the cost of moving the part around during its extra movement cycle.

But it's not like heavier shoes.  Most of the work done to stabilize your head while running is performed by the nuchal ligament (right).  Unlike muscles, ligaments don't need energy to do their work, so the human head is restrained while running with little energy cost.  This structure is found in other animals that hold their head up all day, and in running animals; indeed, in humans, it may have developed largely because of running.  So, while a heavier shoe incurs a cost every time the foot is lifted, the man bun's weight is mostly controlled by the passive nuchal ligament and incurs little cost.  The extra weight of a sweaty man bun will add to this, but not much, and only if other muscles must be recruited.

2. Thermoregulatory consequences of the man bun

Except in very cool conditions, heat buildup is among the biggest factors limiting endurance performance (this is now the subject of my research).  Running generates massive amounts of heat; a trained athlete running a 1-hour race will elevate their metabolism ~19x over resting levels and heat is a byproduct. (In fact, you don't need to wait for summer to become heat acclimated: running even in moderate temperatures will get you partway there, as the endogenous (internally-produced) heat overwhelms your body even more efficiently than ambient heat.)  Because the body can only tolerate a small increase in internal temperature, much of our biology has been shaped in part by the need to dissipate heat: huge numbers of sweat glands (unusual and rare among animals), hairlessness (also weird), profusions of blood vessels supplying the skin, even our very body shape.  I'll note here that our unique heat dumping biology implies very strongly that our evolutionary past included strenuous heat producing activity.  Sustained running, or run-walking, could explain it, and few other things can.  Even carrying a heavy load in very hot weather, walking produces only half the sweat that running can; so arguing that our ancestors did not rely somehow upon running necessitates accepting that our cooling ability has huge safety margins, excess capacity.  And such excess capacities in complex and costly physiological systems are rare in the animal world.

Back to the head.  The brain lives here and is very temperature sensitive.  In fact, the body's chief heat-related concern during exercise is to cool the brain, as even a small temperature increase spells neural death.  In the cold, you lose roughly as much heat from the surface of the head as you do other parts of the body.  In hot conditions and during exercise the forehead sweats a lot, maybe even a bit more than other body regions, but overall the head isn't more of a radiator than you'd expect for a body part with its surface area to volume characteristics.  The brain is cooled primarily by taking hot blood elsewhere in the body for sweating, not simply by local sweating at the head, though other neat hypotheses abound: perhaps our larger sinuses or extra blood vessels permit brain cooling, or perhaps sweat on the scalp releases heat and then is re-absorbed into the skin instead of evaporating.

Machado-Moreira et al., 2008
So what about hair?  Does it prevent heat loss? Wearing a hat does, but hair?  You sweat everywhere on your head, though sweating under the hair isn't particularly effective because hair blunts evaporation and the sweat glands get clogged up (unless the weird hypothesis above holds true).   A 2008 experiment (left) confirmed earlier findings that the forehead sweats more than other head regions, with the top of the head having the lowest sweat rate.  Still, any sweat is helpful, the skin under head hair does do some sweating, and hair prevents sweat from doing its job (but again see the weird idea above).  However, this experiment involved shaved heads.  I suspect that the observed sweat rates would be lower with hairy subjects, maybe much lower, as sweat glands on a hairy scalp may clog and stop sweating very quickly.

The evidence is incomplete but it would seem that any hair on the head will blunt sweating to some degree.  But hair has another thermoregulatory purpose: it blocks solar radiation, shielding the brain from external heat.  In full sun, this heat-shielding mechanism might outweigh heat-increasing penalties incurred from reduced head sweating, having a net positive effect on body cooling.  In less sunny conditions I bet no hair is the best bet.  A man bun shouldn't be any worse than any length hair because it is up in a, well, bun, away from the forehead and neck and other sweat-producing skin regions.

What's the verdict?  A man bun will have a negligible effect on the energy cost of running and removes perhaps 10 square inches of skin real estate from the body's sweating arsenal.  This effect isn't any worse with any length hair, but in a competition where seconds count, this tiny added heat stress could matter.  For summer racing in the sun having some hair, even a man bun, is good, but a race entirely in the woods might call for a shaved head.

But none of this is as costly as the stigma that we man-bunners carry around.  Time to go to Whole Foods where I'll blend in.


Lieberman, D. E. (2015). Human locomotion and heat loss: an evolutionary perspective. Comprehensive Physiology.

Machado-Moreira, C. A., Wilmink, F., Meijer, A., Mekjavic, I. B., & Taylor, N. A. (2008). Local differences in sweat secretion from the head during rest and exercise in the heat. European journal of applied physiology104(2), 257-264.

Takeshita, K., Peterson, E. T., Bylski-Austrow, D., Crawford, A. H., & Nakamura, K. (2004). The nuchal ligament restrains cervical spine flexion. Spine29(18), E388-E393.

Wheeler, P. E. (1984). The evolution of bipedality and loss of functional body hair in hominids. Journal of Human Evolution13(1), 91-98.