Re: Further to neurons as networks

On Jun 10, 12:34 pm, Wolf Kirchmeir <wolf...@xxxxxxxxxxxx> wrote:


You're continuing to beat a dead horse - playing word games. Everyone
already knows what's given below, except for those who arbitrairily
over-simplify it - eg, check out the thread on AGI, in regards people
who say things like "neurons are simple" ...

http://www.listbox.com/member/archive/303/=now

In actuality, individual "real-world" neurons are extremely complex
analog-digital "computational" devices in their own right, and make
many 1000s of connections with other neurons .... as I already said.
Those who play with ANNs use extremely simplified "neuron" simulacra,
and stick them into extremely simplified networks.

Since the advent of new neuropharmacological techniques in the past
couple decades, realization of the complexity of individual neurons
has only increased. Identical looking neurons in identical looking
neuropil can have vastly different neuropharmacological markers.

To call individual neurons = networks, and individual synapses =
networks, etc, is just to play muddy-the-water wordgames. However, it
is clear their are complex chemical and genetic dynamics that underly
all of this. You're purposely confusing different levels of analysis.



Following from a report about differences in the synapses of different
species, in the NYT:

"A whole new dimension of evolutionary complexity has now emerged from a
cross-species study led by Dr. Seth Grant at the Sanger Institute in
England.
.....
The computing capabilities of the human brain may lie not so much in its
neuronal network as in the complex calculations that its synapses
perform, Dr. Grant said. Vertebrate synapses have about 1,000 different
proteins, assembled into 13 molecular machines, one of which is built
from 183 different proteins."

See:http://www.nytimes.com/2008/06/10/health/research/10brai.html?th=&adx...

Original article in Nature Neuroscience, where I searched on Seth Grant,
and found a reference to the original article (which may be purchased
for $32):http://www.nature.com/neuro/journal/vaop/ncurrent/full/nn.2135.html
will take you to the relevant page, where you can choose to spend your
money. Or not. ;-)

As I understand it, a synapse is a mini-network, where inputs (chemical
messengers) trigger reactions that amount to data processing/learning. A
synapse is a chemical network: the number and type of reagents determine
the reaction sequences triggered by the messenger molecules. Within the
neuron, further reaction sequences are triggered by the outputs from the
synapses.

Thus a neuron is in effect a micro-computer, possibly as complex and
capable as the ones that operate our electronic devices (watches, cell
phones, VCRs, microwave ovens, washing machines, clock radios, ....)
Which would suggest that we have already achieved AI, at least at a
level equivalent to that of neurons and simple neural networks.

Just ruminatin'. ;-)

--
wolf k.

.

Relevant Pages

  • Re: Gould vs. Dawkins
    ... however has about 100 million neurons ... with hundreds of dendrites and hundreds of synapses (outputs) ... form between specific populations of cells in specific arrangements. ... Roughly, there are some dozens of known neurotransmitters, and the same ...
    (talk.origins)
  • Re: Gould vs. Dawkins
    ... however has about 100 million neurons ... with hundreds of dendrites and hundreds of synapses (outputs) ... form between specific populations of cells in specific arrangements. ... Roughly, there are some dozens of known neurotransmitters, and the same ...
    (talk.origins)
  • Re: Statin reduction: Worthwhile approach?
    ... There are lots of laboratory arguments pro and con statins (and other ... Although synapses were first identified 100 years ago, their fomation, ... ....Neurons in culture form few synapses unless glial cells called ... These particles are internalized by neurons, ...
    (sci.med.cardiology)
  • Re: machine thinking
    ... >> The synapse can provide a basis for a complex logic system. ... >> For this complex logic system exhibit, the inhibitory synapse starts ... one looks to rhythmic neurons. ... Although I was considering inhibitor synapses, ...
    (sci.cognitive)
  • Is the brain behaviour chaotic or predictable?
    ... Neuroscience and the Max Planck Institute for Dynamics and Self- ... They have found that, under certain conditions, neuronal networks are ... Neurons process information represented as electrical impulses. ... theoretical study that the observed irregular neuronal activity of the ...
    (uk.philosophy.humanism)