Self-organized complex neural networks through nonlinear temporally asymmetric Hebbian plasticity

Abstract

Triggered by recent experimental results, the spike-timing-dependent plasticity (STDP) has been widely analyzed in the neuroscience. In this paper, we analyzed how spatial structure of neural networks will be organized through the STDP. In the experiments, we did use additive and multiplicative STDP rules as well as a nonlinear temporally asymmetric Hebbian rule. As a result, if the additive rule is applied, neural networks exhibit small-world properties. On the other hand, in the case of the multiplicative rule, the small-world networks are not constructed. In addition, we also found that the small-world properties of the neural networks become higher if the STDP rule is less dependent on current synaptic weights. © Springer-Verlag Berlin Heidelberg 2008.