Thoughts on Consciousness

Harry Erwin erwin at trwacs.fp.trw.com
Mon Mar 22 07:47:58 EST 1993


Over this last weekend (March 20-21), I made a 1000-mile round trip to
pick up a son at college. (This is the son who did that early GA work
five years ago.) He had some questions about models of consciousness that
are worth following up.

(In the following, I will use "single-threading" in contra-distinction to
parallel processing.)

1. Is consciousness associated with single-threading?

2. We know of at least two processing modules in the brain that
single-thread: attention and memory-retrieval. 
2.a Are there any others?
2.b Is planning distinguishable from memory-retrieval in the same way
attention is distinguishable from memory-retrieval?
2.c Is planning single-threaded?
2.d Desimone's work distinguishes two types of attention: search for a
known object and scanning for novelty, with separate functional modules in
the brain. Can these be done in parallel?
2.e Has anyone studied the orientation response in detail? (I assume so,
but I'm looking for references.)
2.f Is there any aspect to creative cognition that is not subsumed in
planning?

3. I know of one mechanism that necessarily single-threads--Freeman's
model of olfactory processing. Are there any others for which evidence is
forthcoming.

By the way, for those of you who like to speculate about chaos in the
brain, here are a few points:
Chaos plus noise can result in two phenomena: noisy chaos and chaotic
noise. Noisy chaos looks like chaos, with the strange attractor
dominating, but "fuzzed up" some. Chaotic noise is mostly noise, with
hints of pattern. Noisy chaos is associated with fairly robust non-linear
dynamics, with a limited number of hyperbolic points. In a parametric
system approaching chaos, the noise can cause the system to become chaotic
early, but doesn't change the nature of the chaos. In chaotic noise,
things are much less robust, with many, many hyperbolic points and the
emergence of transient domains not seen in the deterministic system.
Chaotic noise is much less useful to an engineer than noisy chaos. Noisy
chaos can free the system from being locked into limit cycles and fixed
points. Chaotic noise is so indeterminant that the predictability needed
for engineering (and natural selection) is not available on any time
scale.
Freeman's results suggest that chaos in the brain is noisy chaos, not
chaotic noise.

Cheers,
-- 
Harry Erwin
Internet: erwin at trwacs.fp.trw.com
Looking for interesting work...



More information about the Neur-sci mailing list