IUBio Biosequences .. Software .. Molbio soft .. Network News .. FTP

single neuron activity

Oliver Wendel AG Richter wendel at uklirb.informatik.uni-kl.de
Thu Sep 9 08:10:00 EST 1993

Dear readers of bionet.neuroscience,

I'm involved in an interdisciplinary project where we try to model and simulate 
biologically realistic neural networks (in invertebrates; stick insects, to be 
more precise).

My interest lies in describing the patterns generated by single neurons of a 
network. We transform simulated or actually recorded (real) and digitized 
membrane potential traces into a symbolic behavioral description which is 
represented as a tree structure. This tree representation accounts for 
interesting phenomena like occurences of spikes, EPSPs, IPSPs and (nested) 
repetitions of these plus additional information (duration, amplitude, 
frequency,...). The tree representation can be further processed by a computer 
system. Currently we try to describe neural activity in terms of (tree) grammars 
that are to be induced by presenting the induction algorithm a couple of 
activity patterns a neurophysiologist considers as similar. The induced grammar 
defines their common language and eventually this language may be a 
classification basis for neural behavior of single neurons. It may be 
questionable though, if a single neuron really encodes a significant amount of 
information or if the focus should be more on the behavior of neuronal ensembles 
or even areas in the brain. But I think this is a question of scale and 
investigations on every level are justified and may contribute to a better 
understanding of the "real thing". 

So here goes my question: Does anybody have pointers to publications of similar 
syntactical approaches to the description of neural behavior (besides chaos 
theory, statistical approaches, hidden markov models etc.). For me, the idea of 
a neuron "speaking a language" in a certain context seems really appealing... 
Please let me know about any references that come to your mind. Thanks very much 
in advance. 


   -- oliver.

Oliver Wendel
AG Kuenstliche Intelligenz/Expertensysteme
FB Informatik
Universitaet Kaiserslautern
Postfach 3049
67653 Kaiserslautern

Tel.:     + 49 631 205 3327
Fax:      + 49 631 205 3357
email:    wendel at informatik.uni-kl.de

-- oliver.

More information about the Neur-sci mailing list

Send comments to us at biosci-help [At] net.bio.net