Neural networks through the hourglass

The effect of the synaptic plasticity on the dynamics of a large neural net work is studied. Our approach is analytic but inspired by the data, both si mulated and experimental. We explain formation of the small strongly connec ted assemblies within a dynamical network following Hebb's rule. Also, we f ind the conditions for the synchrony effect in the stochastic network in th e absence of large synchronized input. (C) 2000 Elsevier Science Ireland Lt d. All rights reserved.