Well, not everybody went
home. Some people were persistent enough to look at the reasons, took
the possible golden nugget to see what went wrong, and did a study in
animal models that concluded with the comment that
They used radioactive
tracers. I replotted the radioactive tracers distribution on a semilog
plot. Those red dots are the relative concentration as a function of
the distance from the source. The distance from the source is measured
in millimeters.
And I looked at this
thing and say: I've seen that curvature. That curvature is a diffusion
curve. I go back to my books, look up the diffusion curve, plot it and
it does look like the data. If it was convection, the concentration
would be much more even. So the data do not look like convection.
A more
recent experiment will give you further indications of that.
The target areas are
indicated in red. The cannulas which you can’t see, are the two vertical
lines. One of them, it turns out, misses the target area. The other one
doesn't. What happens when we look at the video of putting an MRI
observable tracer is that we see completely different behavior between
the two sides.
This is
representative of many, many observations, now that this kind of MRI
observation can be done intra-procedurally. And we find that sometimes
we get this, other times we get that, and God knows why one thing or
another. We certainly don't know in advance. We can hypothesize
afterwards, but hypothesizing afterwards is nothing short of conjecture,
and my conclusion is that convection is not much more than conjecture.