This paper presents a new theoretical design of how an optimal synchronization is achieved for stochastic coupled neural networks with respect to a risk sensitive optimality criterion. The approach is rigorously developed ...