Robot Dreams

George Bajszar gyuri at
Tue Sep 12 22:42:54 EST 2000

In article <8pm58i$buf$1 at>,
  "Sid Chaudhury" <direwolfc at> wrote:
> scroll down for response.

> > Everything that we "pay attention to" will get recorded into
> > our memory. This has been well demonstrated by various
> > neuropsychological experiments.
> >
> > In one test called the Recall Test, 15 cards with various
> > images or words are flashed, and than the subject is asked
> > to recall the images or words presented, regardless of their
> > original order of presentation. Only 6 or so on average is
> > recalled. That does not mean that the memory of all 15 cards
> > are not stored in the brain.
> >
> > In another test called the Recognition Test, the 15 cards
> > are flashed to the subject and then those cards are shuffled
> > in with another set of unseen 15 cards. The 30 cards then are
> > flashed one by one and the subject is asked to identify which
> > cards he has already seen. On average, 97% of the cards were
> > correctly recognized, even when pausing for over 20 minute
> > delay after presenting the first 15 cards.

>     I don't think that this recall test proves anything other
> than the fact that our brain stores a lot of memory, which you
> stated.

The recall/recognition tests demonstrate that a lot of memory is stored
into the short term memory.

> But this hardly supports a hypothesis that the brain
> uses the sleep period to 'clear out' useless information.

Think about this: People who miss out on proper sleep were shown to
develop bad memory. Something happens during sleep, particularly during
dreams that tends to have a memory cleanup effect. Now this can be
caused by other chemical reasons related to dreaming, or as I mentioned
in this small theory, perhaps dreaming is associated specifically for
filtering out garbage. It is natural for me to conclude that as the
sleep is governed by chemical processes, certain chemicals induce
firing of weakly established neurons in neural networks. The contents
of those neurons spark their visual contents and dreams begin, either
as a filtering process to provide some level of consciousness that is
required for determining useful data from unuseful one, or as a
sideeffect or random neuron firing, I prefer the first version that
dreams must have specific important purposes.

> Do we subconsciously sort out what is useful and what is not?

Clearly only the conscious mind can decide what is useful. A partly
concious mind that occurs during dreaming might be able to do the same.
Perhaps neurons begin firing random signals during sleep and depending
on the reaction of the other neurons the importance of a specific
neuron or small groups of neurons can be decided. Is that the actual
case, I don't know, I just presented a small theory that made some
sense to me.

> Subconsciously deciding importance and sorting seems a little
> extreme.  Also, i very much doubt, that even with the immense
> amount of information the brain 'records' during the day, that
> it will 'fill up' like the RAM on a computer that must then be
> emptied like a cache to make room for the next day.

The brain certainly has a limited capacity. I assume that either
neurons are rewritten as you said, or organized as I said during the
day and at night.

>     I think that memories are created and stored continuously.
> Also, the amount of 'importance' that a memory is from the
> hippocampus, which recent studies have shown deals with the
> allocation of 'attention' to memory, thus causing some events
> that use more attention to be able to be stored and recalled
> more easily.  The more attention was 'allocated' to the event,
> the stronger connection the memory has with the network.  This
> connection is strengthened through repeated activiation of such
> memories.  What happens to the 'useless' everyday memories is
> up for pure speculation.  It seems likely that 'camcorder'
> memory is stored  and lost at a constant rate, as in you
> will remember something that you passed by five minutes ago,
> and the further away from the time at which the mundane event
> occured, the less chance you will have of accessing that memory.
> Possibly the inactivity of such a neuron leads to reallocation
> of another more recent memory (again, speculation).

Sounds reasonable and very possible to me.

>     Whatever the case, it seems to me that you are treating the
> brain more like a computer than what it is.

That is correct. I know a lot about computers and very little about the
brain. Now I am trying to compare the two, there is a lot lot to learn
about the brain.

> While we cannot yet
> answer very much about memory, I don't know how much credibility
> your 'memory dump' theory has.

Again, think of dreams been shown to play significant role effecting

> The brain has an almost limitless capacity of storage,
> especially compared to computers.

In 20 years or so that limitation may be gone.

> I don't think the brain can be 'full' as you put it, and i don't
> think that we at a subconscious level sift through the gargantuan
> amount of information and sort what is 'needed' and what can be
> thrown out.

And perhaps the brain is always 100 percent full. How well organized
the neural nets may determine the functioning ability of memory.

> as i stated earlier, the hippocampus assigns importance to
> memories in 'real-time' based on the 'allocation' of 'attention'.
> So chances are, useless memories are just lost in the immense
> network, while useful memories remain connected.

I wonder what might be the role of dreams since without them neural
networks begin braking down, old and recent memories begin to fade away
quickly, and schizophrenic hallucinations begin.

George Bajszar

Sent via
Before you buy.

More information about the Neur-sci mailing list