fi@grebyn.com (Fiona Oceanstar) (04/01/91)
[This is the 2nd posting of this, which didn't appear to get through
on the 1st pass--apologies to those who may have gotten duplicates.]
Since I'm a newcomer to comp.ai.philosophy, let me start by introducing
myself. I'm a psychiatrist (M.D.) with experience in the following
areas: 1) teaching of neurosciences to medical students, 2) research in
neurophysiology (of rat hippocampus), 3) clinical psychiatry in general,
and 4) psychoanalytic psychotherapy (practice, teaching, research). I'm
actively exploring the realm of ideas loosely classifiable under the
rubric of mind/brain philosophy, but I have no experience in artificial
intelligence--other than that of an observer. (Please let me know, by
the way, if this is the right electronic forum for my interests, or if
there's another newsgroup or mailing list that would be more
appropriate.)
-------------------------------------------------------------------------
Not surprisingly, I've been paying close attention to the discussion of
the function of dreams. I'd like to offer up for consideration a
hypothesis put together by neuroscientist Jonathan Winson, as described
in his 1985 book, _Brain_&_Psyche_:_The_Biology_of_the_Unconscious_.
The central problem faced by the mammalian brain, says Winson, is how to
integrate experience over time--how to construct and modify neural
representations of the environment so as to guide the animal's future
behavior. *Learning*, in other words. Winson breaks this process down
into three components:
1) assimilating new information,
2) associating new info. with memories of past experiences,
and 3) formulating strategies to govern future behavior.
Now if learning is seen only as a process that goes on when the animal
is awake and interacting with the environment, there would be no need
for the dreaming brain to participate in this process. And indeed,
observations of an Australian animal called the echidna (spiny
anteater)--the one other living monotreme in addition to the duck-billed
platypus--confirm that it's possible to construct a brain that doesn't
need REM sleep.
What's different about the echidna brain? It has a very large
prefrontal cortex--the largest in relation to the size of the brain, in
fact, of any species--including human beings. Indeed, if we had as
proportionally large a prefrontal cortex as echidnas do, we would need
a wheelbarrow to carry it around.
Rather than going into depth on the subject of the prefrontal cortex,
suffice it to say that this area of the brain has been found to process
sensory information--through a continual sifting and updating of data--
with the eventual product of developing a cognitive scheme for behavior.
The prefrontal cortex has been called the "executive of the brain"
because of how its processes lead to *action*.
Winson uses the strange case of the echidna to make a specific
connection between the role of the prefrontal cortex, and the role of
dream sleep. Because the echidna does *not* dream, he says, it needs a
very large prefrontal area to carry out the complex functions of
learning.
So what is the scheme used by marsupial and placental mammals? Winson
says, "I propose that it [is], in computer terms, off-line processing"--
where off-line processing entails the acquisition of input information
and its temporary storage in memory, until such time as processing
components become available. And when is the brain "off-line"? When
it's asleep. This scheme would allow for a smaller prefrontal cortex
because the task of integration, or learning, would not have to take
place simultaneously with the processing of new information. The
integration, association, and action-planning stages of learning, could
take place during REM sleep.
I'm grossly oversimplifying his argument, of course, but here are a few
hints about the evidence he has to support this hypothesis:
--observations regarding theta rhythm, which signals processing of
sensory information;
--neuronal gating in the hippocampus, which allows for different
functional connectivity during theta vs. non-theta states, awake vs.
slow-wave sleep vs. dreaming states, and so on;
--lesion experiments in cats in which the "activities" of REM sleep
are found to be attack and fear behaviors--thus giving rise to the
notion that the laying down, integration, and reheasal of predatory and
territorial behaviors take place in the dreaming brain;
--and, of course, the large pool of data from psychoanalysis about
the function of dreams.
Winson's book is quite provocative. I recommend it highly to all who
are interested in brain, mind, dreams, and the unconscious.
--Fiona Oceanstarpetersja@debussy.cs.colostate.edu (james peterson) (04/03/91)
In article <1991Apr1.143027.12963@grebyn.com> fi@grebyn.com (Fiona Oceanstar) writes: > >So what is the scheme used by marsupial and placental mammals? Winson >says, "I propose that it [is], in computer terms, off-line processing"-- >where off-line processing entails the acquisition of input information >and its temporary storage in memory, until such time as processing >components become available. And when is the brain "off-line"? When >it's asleep. This scheme would allow for a smaller prefrontal cortex >because the task of integration, or learning, would not have to take >place simultaneously with the processing of new information. The >integration, association, and action-planning stages of learning, could >take place during REM sleep. > > --Fiona Oceanstar Hmm. The notion that dreams are actually the mind working overtime to manufacture symbols and assimilate new information is not new. I am reminded of Suzanne Langer's *Philosophy in a New Key* from the late 60's or early 70's. -- james lee peterson petersja@CS.ColoState.edu dept. of computer science colorado state university "Some ignorance is invincible." ft. collins, colorado (voice:303/491-7137; fax:303/491-2293)
prune@athena.mit.edu (Paul Berland) (04/05/91)
In article <1991Apr1.143027.12963@grebyn.com>, fi@grebyn.com (Fiona Oceanstar) writes: |> .... |> The central problem faced by the mammalian brain, says Winson, is how to |> integrate experience over time--how to construct and modify neural |> representations of the environment so as to guide the animal's future |> behavior. *Learning*, in other words. Winson breaks this process down |> into three components: |> 1) assimilating new information, |> 2) associating new info. with memories of past experiences, |> and 3) formulating strategies to govern future behavior. |> .... |> So what is the scheme used by marsupial and placental mammals? Winson |> says, "I propose that it [is], in computer terms, off-line processing"-- |> where off-line processing entails the acquisition of input information |> and its temporary storage in memory, until such time as processing |> components become available. And when is the brain "off-line"? When |> it's asleep. This scheme would allow for a smaller prefrontal cortex |> because the task of integration, or learning, would not have to take |> place simultaneously with the processing of new information. The |> integration, association, and action-planning stages of learning, could |> take place during REM sleep. Q1: Does this theory imply that while awake, only 1 and 2 can operate? (Perhaps only tentative "strategies" can be formed while awake that become incorporated only during dreaming? Can we consciously reference these tentative strategies and in what way do they subjectively differ from fully formulated strategies?) Q2: Does this theory imply that lucid dreaming (having an awake mind while dreaming) may have significant effect on subconscious motivations? (I know that lucid dreaming has occasionally been used in psychotherapy and "self-psychology" but TO WHAT EXTENT is lucidity potentially causing havoc to the proper functioning of the brain? How comprehensible is the "strategy formulation process" to the conscious mind?)