Experimental studies have shown that cortical neurons activated by external stimulus could synchronize their spike trains. A network architecture inspired from biological is here presented which models part of the retina and parts of two visual cortical areas in opposite hemispheres. A probabilistic model is proposed to model the behaviour of neurons. This model takes into account both spatial and temporal integrations with interneural delays. Its own particularity is the model of the refractory period using for each neuron, a spike memory term. Neuron states are binary (0, 1), and dynamic is stochastic.
展开▼
