Cortical Architectures

Harry Erwin erwin at trwacs.fp.trw.com
Sun Mar 6 09:42:24 EST 1994


Has anyone tried to develop a Katchalsky network model of the neocortex?
(Here I am following Freeman in using a 'two-layer' functional view of
the neocortex, with the two functional layers each consisting of axons,
pyramidal cells, and inhibitory (GABAergic) interneurons as sublayers.)
The key 'design' features would be the reduced latency and latency
variance between the two layers and the close topological relationship
between the two layers. Hence the upper layer would have pyramidal cells
competing (through inhibitory interneurons and reciprocal synapses--
which operate as switches, BTW, since transmission in one direction
inhibits transmission in the other direction) to recognize the feature
being presented locally. The lower functional layer would then have
pyramidal cells competing to respond to the pattern of upper layer
pyramidal cells that have successfully responded to the afferent pattern.

Some questions:
1. Are the two layers of pyramidal cells in series, or is there immediate
feedback?
2. Do both layers of pyramidal cells have external afferents? If this were
the case, I'd suspect the two-layer structure of being a way of generating
'vectors' (as opposed to the 'points' generated by the paleocortex). Is
there any evidence for differences in cognition between mammals and other
tetrapods that would point towards a cognitive 'atom' present in mammals
and not in other tetrapods?
3. Do the inhibitory interneurons have external afferents? external
efferents? Are there connections between the inhibitory interneurons in
one functional layer and the cells in the other functional layer?
4. Does this system typically interact with a nucleus to create a
Hebb-Marr autoassociator?
5. Anyone have a clue as to the function of the cortical neurons that use
neuropeptide transmitters? 
-- 
Harry Erwin
Internet: herwin at gmu.edu or erwin at trwacs.fp.trw.com
Working on Katchalsky networks....



More information about the Neur-sci mailing list