Re: CO2; experiment design

["John S. Denker" <jsd@MONMOUTH.COM>]

At 03:11 PM 8/21/01 -0400, I wrote:
> The length of time you can hold your breath does not typically have
> much to do with running out of oxygen.  Rather, it has to do with
> overloading on CO2. This is not super-obvious to most people, because under
> most conditions, anything that produces CO2 consumes oxygen, so it's not
> easy to distinguish one process from the other.  Can you design a
> non-hazardous experiment to test this?

A few people have responded to this.

It's about time for me to spill the beans about what I had in mind.  It's
partly a point about CO2 but mostly a point about experiment design.

First of all, a few words about safety:  Don't do hold-your-breath
experiments alone, and don't do them standing up.  People sometimes have
surprising reactions to seemingly-mild hypoxia.

Anyway....  We want to understand how the body regulates O2 and/or
CO2.  This is tricky because typical physiological processes move us along
a line of more CO2 and less oxygen.  Roughly speaking we are moving up and
down the line in a graph like this:

        CO2
         |
         |
         |\
         |  \
         |    \
         |______\_________O2-->


1) As an example of an experiment which is not (by itself) very effective,
consider measuring how long you can hold your breath with your lungs full
versus lungs empty.  With lungs empty, there is little oxygen in the lungs,
AND little room to store CO2.  (The partial pressure of CO2 that you can
stand is very small compared to 1 ATM, so don't think you can just fill up
the initially-empty lungs with CO2.)  So this experiment just moves us
along the same line.  We need a way to get !off! the line.

2) Things get somewhat more interesting if we consider the following
experiment:  Inhale a lungful of pure oxygen and see now long you can hold
your breath.  This provides 5x more oxygen, without any extra room for
storing CO2.  I predict that you will not be able to hold your breath 5x as
long, indeed hardly any longer than with plain old air, because of the
dominant role of CO2.

This is interesting because it is quite contrary to the common unscientific
intuition that O2 is the dominant factor.

However, this experiment (2) by itself is almost as incomplete as
experiment (1) by itself, because there are so many competing
hypotheses.  One could explain away the result of experiment (2) by
hypothesizing that lung-storage volume is negligible compared to storage in
blood and tissues.  This is not a crazy hypothesis:  quite a lot of O2 is
stored in hemoglobin and myoglobin, and quite a lot of CO2 is buffered by
serum chemicals and intracellular proteins.  Note that elephant seals dive
after exhaling, with their lungs (and trachea etc.) collapsed!  And they
are the world's-champion divers, capable of staying down longer than whales
or any other creatures.

3) The clever thing to do would be to combine experiment (1) with
experiment (2), thereby establishing that lung volume is important and that
O2 is not the dominant factor in the perceived need to breathe.

4) One could get an even more spectacular result as follows.  Think about
the graph given above.  Think, how can we move radically off of the usual
physiological line?  How can we move in the northeast-southwest direction
on that graph?
  a) Inhale a lungful of pure nitrogen (or helium or whatever).  See how
long you can hold your breath.  Your lungs start out with no oxygen and no
CO2, but lots of room for storing CO2.  That moves strongly to the
southwest corner of the graph.
  b) Inhale a lungful of 50% oxygen and 50% CO2.  See how long you can
hold your breath.  Your lungs start out with an abundance of O2, but waaay
too much CO2.  That moves strongly to the northeast corner of the graph.

My prediction is that you can hold your breath much, much longer in case
(4a) than case (4b).

================================

Anyway, you can see the point about experiment design.  The line in the
graph above is an example of a "confusion curve".  An experiment that moves
us up and down the confusion curve doesn't eliminate the confusion.  You
need to design an experiment that moves us crisply away from the confusion
curve.

Another point -- a cute little side-issue:  To really explore the parameter
space requires three sources of gas:  O2, CO2, and an inert gas (nitrogen
or whatever).  Why three?  After all, the graph above describes a
two-dimensional space, %CO2 versus %O2, so you might think that two
gas-supplies would suffice.  Answer:  You need three because there is a sum
rule:  under usual laboratory conditions you can't inhale a sample that has
a partial pressure of 0.05 Atm CO2 and a partial pressure of 0.05 Atm O2
and nothing else.  You need a buffer gas to bring the total pressure up to
1.0 Atm.

To summarize:  Some guidelines for experiment design:
 -- Figure out what are the key variables.
 -- Make sure you have enough independently-controllable variables so that
you can explore the relevant parts of the space.  Drawing a graph of the
space may help.  Don't let yourself get trapped on a confusion curve.
 -- If a simple experiment doesn't do the job, maybe you need a compound
experiment consisting of two or more sub-experiments.
 -- Be skeptical of your own hypotheses.  Make a list of competing
hypotheses and see if your experiments suffice to rule them out.