> "Richard Norman" <rsnorman at mediaone.net> wrote in message
> news:Kzul6.4744$v_1.474996 at typhoon.mw.mediaone.net...> > One major problem with your idea is that nerve cells are really quite
> > tightly packed. If you could, for example, eavesdrop on a peripheral
> > nerve from some distance and actually see all the action potentials,
> > you still have the monumental problem of separating the activity in
> > one cell from that of another which lies only a few hundredths of a
> > millimeter away. Somehow you would have to interpret what the
> > nerve activity "means" rather than identify activity purely by location.
> >
You will never figure out what the activity means by using probes or analyzing
the data generated by one nerve cell. The problem is similar to understanding
the non-linear dynamics of the activity of water/oil at one particular position
in a wave machine. As the water/oil goes up and down the mathematics of the
interaction is unpredictable at a particular point. The problem is you are
looking to close. If you back up you will find that you have a "correlational
opponent processing" or "associational reciprocal inhibition" machine which uses
wavelets and wavelet interference patterns.
Ron Blue
http://turn.to/ai
---
