memory (and SOMs)

Matt Jones jonesmat at physiology.wisc.edu
Wed Feb 27 14:38:52 EST 2002


"yan king yin" <y.k.y@(dont spam)lycos.com> wrote in message news:<VUQe8.2378$XZ.351610 at news.xtra.co.nz>...
> "mat" <mats_trash at hotmail.com>
> Braitenberg and Schuz have done some statistical analysis of the cortex,
> and they found that most parts of the neocortex are organized into
> discrete "patches" with parallel projections connecting them. Therefore
> one thing for sure is that cortical connections are not diffuse (as least
> in the neocortex). My point is that these discrete connections between
> patches are specified by genes, ie not altered by experience under
> normal circumstances. For example in the visual cortex you have areas
> V1, V2, V4 and MT connected by specific pathways that do not vary
> from individual to individual (as least such has not been reported). As
> pointed out by Matt Jones, there are exceptional situations where
> cortical reorganization does take place -- see my reply to him.
> Personally I am inclined to the view that large scale cortical pathways
> connecting the "patches" are genetically determined, I'll explain this
> in Matt Jones thread as well.
> 
> Also Im still writing my reply to your SOM idea, please be patient.


Hi,

I have not yet seen the other posts you referred to, but I have two
brief comments about thislast one:

1) It's one thing to note that  V1 connetcs to V2, V2 connetcs to V4,
etc. It's quite another thing to say that these connections aren't
modified by experience. Such connections between large areas are made
of thousands of individual connections between individual neurons. For
example, is the -number- of axons contained in the projection from V1
to V2 fixed by genetics, or is it subject to refinement through
experience. Is the number and location of synapses that each axon
makes fixed? True, these issues don't change the fact there "is a
connection" but they can make all the difference to "what sort of
connection" it is.

2) Note that in SOMs and other ANNs,  no new connections are formed
during learning. Instead, the existing connections have their weights
adjusted.


In both instances above, which by the way are the norm, not special
exceptional cases, I don't think it is accurate to say " these
discrete connections between patches are specified by genes, ie not
altered by experience under normal circumstances."






Cheers,

Matt




More information about the Neur-sci mailing list