Speech hearing and neural network models door S. Nakgawa