Re: "analog computer" = useless hypothesis?

From: Anders Sandberg (
Date: Sun Apr 01 2001 - 13:46:13 MDT

"J. R. Molloy" <> writes:

> > No. Please wait. Let us ponder.
> > (Is J.R.'s mind digital or not?)
> > - s.
> You mean J. R.'s *brain* right?
> In that case, perhaps Anders (cognitive neuroscience specialist) can
> help to clear this up. Do human brains compute digitally or
> analogically? (Or neither, or both.)

There is currently a big debate among us neuroscience people about
rate coding vs. spike coding. Neurons in mammals usually send signals
by action potentials, quick voltage spikes that cause synapses to
transmit the signal to the recipient neurons. This is highly digital
in itself, although various nonlinearities in the system make it
possible that the interval between spikes might matter for some
processes. The rate coding view thinks that what really matters is
just the average rate of firing, while the spike coding view thinks
that it is the intervals that matter. The truth is likely somewhere in
between; there are evidence for several forms of coding in various
parts of the system. Some neurons appear to prefer an on/off approach
to activity, either being silent or firing away at a high rate, while
others are more gradual in their responses or show complex
interactions. Personally I tend to lean towards the rate position more
than the spike position, but things are not clear cut or simple.

In any case, different codes are limited by the noise levels in the
system which appear to be fairly high. Synapses sometimes accidentally
release vesicles, neurons sometimes misfire or do not fire when they
should, cells die etc. I would say the human brain is a largely
digital system employing special purpose analog components; the analog
parts do not have direct contact with each other, so whatever analog
information is around either exists as spike timing or possibly
chemical gradients. But as I said, it is all noisy.

Anders Sandberg                                      Towards Ascension!                  
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y

This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:59:44 MDT