In article <38329829.99C7146B at cns.mpg.de> Axel Hutt, hutt at cns.mpg.de
writes:
>there seem to be a kind of forecasting of seizures. I have heard
>that brain signals show a transition from noisy and high-dimensional
>temporal behaviour to phase-locked signals means a state of ordering.
>>Can anybody tell me how sure this point is means is this just a
>hypothesis
>or even proofed?
I think this is still a very much developing area. Neurologists can
certainly tell by looking at an EEG that a seizure is taking place, but
I'm not so sure whether there are yet very robust ways of predicting them
beforehand.
One line of work that has a bearing on this, and may be of interest to
you, is the use of so-called "Chaos Control" methods to stabilize neural
activity either into periodic behavior (Control) or out of periodic
behavior (AntiControl). These methods are based on the Ott, Grebogi and
York (OGY) methods developed for chaos control in mechanical systems, and
have been used most notably in brain (i.e., hippocampal slice
preparation) by Stephen Schiff and colleagues, who had a Nature paper
about it several years ago. Try a medline search for "Schiff" AND
"Chaos". Pretty cool stuff. I tried it in a simple model simulation and
it actually worked (!).
Cheers,
Matt Jones
