Saturday, December 18, 2010

Is The Body Such a Chaotic System?

Art De Vany has an excerpt from his new book at Mark's Daily Apple, including this:
We tend to simplify what otherwise seems overwhelmingly complicated. But as we now know, our metabolic function is infinitely complex. I found myself using concepts from other scientific disciplines to help me understand and explain the human body’s inner workings.

According to chaos theory, certain systems that seem to be random in fact are not–it’s just difficult for us to perceive, at the outset, all the subtle factors that set the course and determine the outcome. One landmark of chaos theory is the “butterfly effect.” This says that even a very small, unseen occurrence in a far-off place can have a large eventual impact–that if a butterfly flaps its wings in Hong Kong, the resulting breeze can trigger a cascade of atmospheric events and cause a hurricane in Brazil.

This can be used to explain many of our bodies’ inner workings. Here’s a simple one: If you go to the gym several hours after your last meal (so that you’re on a relatively empty stomach), your body will quickly burn through whatever glycogen is in your muscles and then move on to burning fat, which is the desirable state. But if on your way to the gym you have a sports drink, one with lots of carbs, you’ll need to burn off the glucose first. And depending on your workout, you might never get around to burning fat at all. Same exact exercise routine, very different outcomes, all because of your choice of pre-exercise beverage.
It got me thinking about chaos theory, which I haven't thought much about since college. It also caused a little mental friction, and on closer examination, I decided it was just plain wrong.

Using the butterfly effect as an analogy is fine (if something of trite cliché by now). Perhaps it can also be used to explain some of our bodies' inner workings. But saying that the effect on your metabolism of an energy drink before a workout is "explained" by the butterfly effect is erroneous.

The butterfly effect describes systems that are extremely sensitive to small changes. I'm not sure if saying a butterfly flapping its wings 'causes' a hurricane is really an accurate way to describe it. The proverbial butterfly is not introducing millions of joules of energy into the Earth's atmosphere causing a hurricane. That hurricane, or one very much like it, is going to happen regardless.

When Lorenz (more-or-less) coined the term butterfly effect it was to describe how two non-linear systems, starting with slightly different initial conditions, could quickly diverge. At the time it was thought that long-term accurate weather prediction would soon be possible. Lorenz showed that it was essentially impossible unless highly accurate temperature sensors were placed two feet apart, up into the stratosphere, over the entire globe. Only if the initial conditions could be measured to an impossibly high degree of accuracy would it even be theoretically possible to create a long term predictive model for weather.

Is the body's metabolic function so non-linear, 'infinitely complex' as De Vany puts it? First of all, to say that metabolic function is infinitely complex seems rather silly. Everything is 'infinitely complex' at some level.

A rock sitting on a table is infinitely complex at the molecular level, but to us that rock is extremely stable. Push that rock off the table and it will hit the floor in a very predictable manner. At a much smaller level it becomes more complex. Can you predict the exact spot where it will hit? Well you'd need to know the exact amount of force with which it was pushed off the table but it is all pretty linear.

Compare that to a faucet that is dripping just fast enough so the drops overlap. Drip-drip....drip.....drip-drip-drip . . . very non-linear system. A slow drip is like a metronome but increase the flow a bit into non-linear territory and it becomes impossible to know exactly when the next drop will fall.

I think the body is a lot more like the rock than the weather. In other words, pretty linear. We have a lot, lot, lot of feedback mechanisms. We aren't even close to understanding them all, so they are definitely complex but they keep the body as a whole pretty damn linear. The job of all these feedback systems is to compensate for butterfly wings, and any other perturbations.

Take De Vany's example: a guy drinks an energy drink before a workout, his body burns glucose from the drink instead of in the muscles. Different outcomes.

This is NOT the butterfly effect.

The butterfly effect would be more like this: two identical twins with identical conditioning go to work out, one drinks an energy drink before his workout, the other works out fasted; ten years down the road one is winning a triathalon while the other has degenerated into a fat slob working on his fifth heart attack.

The body actually fights pretty hard to maintain its desired equilibrium. It takes years and years of crappy food and sedentary lifestyle to mess it up. Once it is deep into metabolic syndrome, it is probably much more sensitive to small changes, because the natural feedback mechanisms are in deep decline. But this isn't the point De Vany was trying to make.

If you look at identical twins separated at birth they usually look quite similar and have quite similar health despite living highly divergent lifestyles (certainly more divergent than whether or not they drank an energy drink or not before working out). There can be marked differences, of course, and this leads to the fascinating topic of epigenetics or gene expression. But in general, the body fights to stay the course. 

This goes both ways, of course. A butterfly flapping its metabolic wings is not going to turn Daniel Day-Lewis into Arnold Schwarzenegger no matter how much iron Day-Lewis pumps.


  1. What do you think of the General Adaptation Syndrome?

  2. I didn't even know what it was, just looked it up in Wikipedia.

  3. "If you look at identical twins separated at birth they usually look quite similar and have quite similar health despite living highly divergent lifestyles (certainly more divergent than whether or not they drank an energy drink or not before working out)."

    Sean you're aware of the Danish twin study, yes?:

  4. Skyler,

    No I'm not, thanks. But looking at the abstract, I'm a little confused.

    First they say
    "The number of hospitalizations in the previous 18 years was obtained through register linkage"

    Then they say
    "Structural equation modeling suggested that approximately a quarter of the variation in the liability to self-reported health and the number of hospitalizations could be attributed to genetic factors."

    So 25% of self-reported health (the 77% of twins that responded) was determined by their modeling to be genetic.

    But then

    "Analyses of the hospitalization patterns of proxy responders and nonresponders suggest that the estimates of the genetic influence on health outcomes in the study are conservative."

    So if they just use the hospitalization data it would seem that 25% genetic factor is too low (although they don't give any indication how low). Since self-reported stats are notoriously dodgy it would seem that this study isn't all that useful.

    I'm aware that identical twins can occasionally diverge quite a bit in health. I remember reading an article long time back about it, one twin only contracting Parkinson's or something, can't remember exactly but it was interesting and quite surprising to me.

    I can't claim even cursory knowledge on epigenetics but I don't think this discounts my claim that De Vany is spouting nonsense about the butterfly effect. Putting glucose in your system before a workout causing one to burn that instead of the stored liver and muscle glucose (if that is indeed what happens) is straightforward case of cause and effect.

    Only one twin contracting Parikinson's (if that's what it was--I really ought to look it up) could be labelled the butterfly effect, I suppose. Although since no one actually knows what caused that epigenetic divergence it might not be. It could be, say, that doing x will always cause this epigenetic divergence.