Exponential convergence of BAM neural networks with time-varying coefficients and distributed delays

In the current paper, BAM neural networks with time-varying coefficients and distributed time delays are studied. Sufficient conditions guaranteeing the exponential componentwise convergence and existence of one unique periodic solution are obtained by the comparison principle, continuation theorem of topological degree and inequality techniques. The boundedness and differentiability of activation functions are removed. The obtained sufficient criteria are easy to verify and are hence very useful in applications.