Using a model network of spiking neurons, the present work invesigates the development of orientation maps in the primary visual cortex. Specifically, a very early Hebbian plasticity of intracortical connections is taken into account. Numerical simulations show that this intracortical plasticity can lead to the development of a lateral connectivity pattern whose global organization resembles very much the structure of orientation maps obtained experimentally. Once such an intracortical connectivity pattern has emerged, it can guide the subsequent Hebbian development of geniculocortical projections. In this way, the layout of the emerging orientation map in the primary visual cortex is predetermined very early by the structure of the intracortical connections. The proposed model provides a unique framework to explain several experimental results that were difficult to understand previously (like, e.g., the results of the so-called reverse-lid-suturing experiments). In order to obtain a better understanding of the underlying processes of synaptic plasticity, an extensive analytical investigation of the model is given in the second part of this work. First, a method is derived that allows to approximate spike-spike correlations in recurrent neural networks. This method is then used to describe the model's learning dynamics by means of eigenvalue problems. The analysis of these eigenvalue problems reveals how the development of the geniculocortical projections can interact with the intracortical connectivity pattern.
«
Using a model network of spiking neurons, the present work invesigates the development of orientation maps in the primary visual cortex. Specifically, a very early Hebbian plasticity of intracortical connections is taken into account. Numerical simulations show that this intracortical plasticity can lead to the development of a lateral connectivity pattern whose global organization resembles very much the structure of orientation maps obtained experimentally. Once such an intracortical connecti...
»