Synaptic modification rules ?

Matthew Kirkcaldie m.kirkcaldie at removethis.unsw.edu.au
Thu May 20 00:46:50 EST 2004


In article <ec29a509.0405192012.5f68e41f at posting.google.com>,
 nettron2000 at aol.com wrote:

>  No need. Not run on at all, quite informative and interesting, some
> of your text leapt out especially when you mentioned AMPA and NMDA
> receptors, had jotted that down a few nights earlier. But first ive
> heard how they (AMPA, NMDA)were removed from or added to the membrane
> as a function of timing, interesting indeed.

Yes, it collapses the ideas into a concrete and visualisable form, which 
I'm all in favour of.

> I can see how this would modify the integrate and fire properties, as
> well as the time constant of the neurons but i guess what im getting
> at is more of the WHY of it. To use the cerebellum as an example: some
> experimental evidence ive rea about suggests that climbing fibers from
> the inferior olive "allow" modification to occur on the parallel
> fiber-purkinje cell synapses when certain conditions exists,
> signifying that either an error as occured and the neuron needs
> updating or the end of a movement/motor program is eminant.I guess
> this is more in-line to what i mean by "rules".

Oh dear, I think that's a whole deeper level of problem.  I understand 
what you were getting at now - why are some synapses more plastic than 
others?  That's really unanswerable at this stage I think.  The "how" is 
a bit simpler - if the biochemical machinery isn't there, the processes 
won't occur.

There are interesting ideas about other possibilities though - for 
instance LTD in the ventral tegmental area, which may be responsible for 
addiction, operates via a non-NMDA-mediated plasticity.  It may boil 
down to the presence or absence of a co-factor - such as stimulation of 
serotonergic synapses at the same time as glutamatergic (one of the 
protocols used in the synaptic capture stuff coming out of the Kandel 
lab) or the trafficking of appropriate biochemical machinery to 
appropriate places in the cell at certain times.

Of course these are only distractions in the face of the real question, 
"why do the properties and functions of neurons differ?", which, in its 
most basic form, is "why is the nervous system the way it is?"  Not too 
many (worthwhile) books on that subject!

>  Neuron function alone has proven to be rather difficult to pindown (
> atleast for me) and neuron learning concepts nearly in-excessable due
> mainly to the lingo used, lack of good lay-person reading material or
> buried in advanced mathematics.

Sadly, there's no good explanation of neuronal function, even buried in 
advanced mathematics.  There are interesting models which use those 
techniques, but the truth is that these questions are some of the 
hardest known to science.  I don't think it would be exaggeration to say 
that the function of a single neuron (actually, any complex cell) is 
nowhere near well-understood.  Anyone who tells you otherwise is 
deluding themselves. Funnily enough some of the problems get simplified 
when you consider groups of neurons, but nobody really knows what 
properties are safe to simplify, and which invalidate the whole 
exercise.  Donald Hebb had a pretty good stab at it though.

         Matthew.



More information about the Neur-sci mailing list