Valeria Del Prete, King's College, 4/8/03
<br>
<br>
 A non-equilibrium statistical mechanics approach to the population
dynamics of spiking neurons.
<br>
<br>
Experimental evidence suggests that the precise timing
of spikes might well be used by neurons for processing and storing
information. Unfortunately, the mathematical analysis of recurrent
networks with spiking neurons is highly non trivial. Most
analytical studies have therefore focused on rate-based
models, whereas recurrent networks with more realistic neurons tend to be
studied numerically. In order to bridge this gap, I will propose
a realistic spiking neuron model where recurrent plastic synapses still
allow for the application of non-equilibrium statistical
mechanical techniques. The final goal is to derive the temporal evolution
for global quantities significant for the population dynamics
(like the 'overlap' of the network activity with the stored patterns in
the simpler case of binary autoassociative memories). The model is
flexible and its parameters can be varied in order to allow a comparison
with real data.
<br>
<br>

See also the <a
href="http://www.math.nyu.edu/~tao/neuroday/schedule.html">neuroday</a>
page for another talk Valeria will give while she is in town.