Neural adaptive filters for estimating brainstem auditory evoked potential

This paper presents Neural Adaptive Filters (NAF´s) for estimating brainstem auditory evoked potential (BAEP). The method is evaluated by using simulation and human data. It was observed that a significant improvement in waveform estimation, compared with the ensemble averaging and time varying adaptive filter (TVAF), can be achieved by the neural adaptive filters. In this work, a multilayer perceptron network (MLP) trained with the backpropagation learning rule and Radial Basis Function Network (RBFN) trained with stochastic gradient-based algorithm were employed for neural filter implementation. It was found that the RBFN give rises to improvements in BAEP estimation over the MLP