>> Incidentally, electronic neural network computers designed to learn
>> continually [i.e., electronics that completely mimic the brain structure]
>> require dreaming in order to function. They can also have fairly
>> convincing analogs of the replay-life version of a near-death experience.
>> [some snippet in a 1995/1996 Scientific American]
>
>This is hogwash. Dreaming is not necessary in order for our minds
>to function.
My take on dreams is that dreams are memes/ideas not subject to (very much)
selection pressure from sensory input. Your eyes are closed, your limbs
aren't moving, and so ideas can arise in your mind which normally would be
outcompeted by more everyday ideas, which are supported by neural firing
patterns from sensory information processing. These ideas, the output of
your neural processing system, are then the only input from which to build
the next set of outputs.
The main thrust of the article I read (in New Scientist, in my case) was
that some guy (I think he's called Steve Thaler) has built a computer that
functions like a neural network. He'd give it various inputs, and parallel
processing in the machine would give rise to brain-like outputs. Meaning, he
could provide it with data about lots of structures for conducting ceramic
molecules, and the machine could, after a while, generate novel molecular
structures, proposals for novel ceramics with the desired qualities. And
apparently, it worked, and he's being paid for the work by chemical
companies. He's also used the machine to compose and copyright thousands of
pop tunes.
The "near-death" thing is about what happened when he started turning off
chunks of the machine's processing resource. The output of the neural net
circuitry looked like the machine was generating (almost unconnected)
memories. Thaler liked this, because as a child (he says) he almost drowned,
and did the life-flashbacks trick while his brain was being starved of
oxygen. The inference is that you might remember your childhood home (for
instance) either by talking about it, visiting it, seeing photos of it, or...
looking at something unrelated to it and KNOCKING OUT bits of processing
resource. If the neural activity balance of your brain, MINUS the activity
that would have occured in the knocked-out bits, is the same as that for you
looking at your childhood home WITH A FULLY FUNCTIONING BRAIN, you'll have
the experience of remembering your childhood home.
I don't have an opinion about whether dreaming is required for brains to
function or not. BUT I think that the ideas and memories that you experience
when dreaming can be thought of as memes interacting unfettered by most of
the evidence that our senses usually provide.
Plus ca change, plus c'est la meme chose (The more things change, the more
it's a meme thing)
Phonecalls: 01494 461648 Phights: 10 Riverswood Gardens
High Wycombe
HP11 1HN
Britain