Modelling the human brain by modelling its evolutionary emergence

mat mats_trash at hotmail.com
Thu Feb 28 08:24:48 EST 2002


> 
> $8 I may refer to $4 in my response to your other email. My approach has
> been in short:
> *   Globalize the neural structure of the brain in meaningful information
> handling entities,
>    based on, unavoidably somewhat speculative, interpretations of known
> detail
>    characteristics.
> *   recognize that evolution is primarily an information handling procedure
> (should I decide
>    for this action or that? What was the result in a previous encounter?
> Etc.)
> *   Model the evolution of that information handling procedure (i.e. virtual
> image handling),
>    using the known characteristics of present Homo as the ‘aiming point' and
> make use of all
>    available information on (cultural) characteristics of Man and its
> precursors.

So what are you going to do? Determine every type of situation a human
will find himself in and look at the 'information handling procedures'
they use in coming to decisions?  This is a useless approach which has
the same pitfalls as the AI approach you talk about later.  Your
talking as if the 'known characteristics' are some fixed list of
behaviours somewhere!

> 
> $10  We are both guessing of course on the functionality of the DNA part
> (largest) that does
> not code for proteins. 

Well no actally, I'm not guessing, I am getting this information from
many years of reseacrh by many scientists.  Just becuase something is
there it doesn't have to have a function (if you really did understand
evolution you could grasp this)

> My suggestion/question is:
> Semir Zeki in his book ‘Vision on the Brain' describes the procedure in
> ontogenesis that
> develops the visual cortex in the foetus. As long as the eyes are not yet
> functioning, there is a
> stream of information flows, generated by that developing brain, that
> stimulates the visual
> cortex to develop in a certain way. As soon as the eyes are sufficiently
> developed, these take
> over by sending that, or another, information stream.
> My point is that these information streams must also be genetically
> transferred to the next
> generation; the only possible vehicle is DNA. I do not see how this could be
> generated by
> some protein(s); their goal seems to be reserved for the physical part of
> ontogenesis. Could
> this ‘information imprinting' not be the function of the non-coding DNA
> part? I speculate that
> this could also be the repository for inbred behaviour and, if accepted as
> concept, archaic
> memories. It could create a rich field for experimental initiatives!
> 

Absolute rubbish.  You do not understand molecular biology or
developmental neurobiology to any extent it seems (and given thats
what your project is centred on, its a bit strange you are so ignorant
of what has been discovered).  Of course there is no gene which
specifically 'says' this neuron will connect to that one, and will
transmit action potentials at 14Hz.  However you do not appreciate the
possibilities for complex emergent activity based on a plethora of
relatively simple processes.  Genes encode proteins which can guide
axons, control electrical properties etc..  This, coupled with the
envrionment, can produce the embryogenesis you talk about.  There are
30,000 genes in the human genome and you conclude that none of these
encodes the development of the visual cortex.  Did it not occur to you
that all of them may be involved?  That it is the interaction of
proteins that is so important, and thats why having the genome may be
useful but is by no means the end of molecular biology.

Further, this 'hidden information' in the non-coding DNA - what
exactly does it code for then?  It can't be proteins becuase you've
already dismissed them, so now you're suggesting that DNA can also
code for a completely new type of biomolecule that has never been
observed.  Good luck.

> $11  What I have presented in ‘Biography of Man' is:
> *   a very global modelling of the information handling aspects in the
> emergence of single-
>    and initial multi-cell life forms 

And now you contradict yourself in the same sentence.  How can you
model globally when talking about cells?

> *   An evolutionary modelling of the emergence, starting from an initial
> multi-cell life form,
>    some 700 million years ago, into Homo sapiens sapiens in chapter 5 of
> ‘Biography of
>    Man'. The resulting decision making procedure in Homo sapiens sapiens is
> described in
>    chapter 6. This will be the starting point for the software developments
> I identified under
>    $7. (See my reply to your other email at  16.11 h)

I'm sorry but how the hell do you know the detail of the evolution of
man's brain (a specifc) when well-regarded scientists are still
striving to understand when we stood upright first ect. (i.e. slightly
less specific - your speculation on the evolution of mind would be
similar to them talking about the first time someone ran 100metres - a
total impossibility)

> 
> $12  ‘Asking questions' is applying the modelling procedure to actual real
> life events, e.g.
> does it explain the existence of infantile amnesia? It is discussed in
> section 7/1/3 in chapter 7
> on psychology. My answer to the question is: yes, it gives a correct
> explanation for this
> phenomenon; the modelling is therefore isomorph with regard to this aspect.
> 
As already stated, your modeling will unlikely be an accurate
reflection of how the brain iteslf processes.  It may produce the same
output from input, but thats no test.



> $13  Newton's laws are then also speculation, but we can very effectively
> work with it
> within certain limitations.
> 

Well for a start, Newton was considering a much less complex problem,
that of single or two-body motion under ideal conditions.  Also, at
the time there was no evidence from studies of finer resolution i.e.
he was looking at the most detailed knowledge of the time.  However,
you are 'modelling' global brain properties when we know a lot about
the fine structure.  So if your ideas don't fit in with the latter
evidence, expect to have it pointed out.  If Newton had been told the
speed of light was measured as constant, his ideas would not have
caught on.  But, as with any contemporary theory that encompasses all
current observations, it became a good model.  Your model will not
encompass all that is known of neuroscience and will blatantly
contradict in certain areas.  Thus it is of little value.  Your
project is like attempting to model the behaviour of a single cell in
a given envrionment and seeing what 'internal' structures/ processes
developed to match the beaviour of a real cell.  However, if the model
came up with a totally different way of producing the same behaviour
it would be considered totally worthless becuase it adds nothing to
our knowledge.  Similarly with your project.




More information about the Neur-sci mailing list