Abstract: Artificial neural networks (ANN) are excellent biological-like models to simulate nervous system functions.

Their spontaneous biological likeness can be further enhanced by letting biological-like procedures, such as

evolutionary selection, shape the investigated ANN models. This makes the investigator duties shift from

designing the networks to designing the environment where the networks behave, so reducing the impact by

possible investigator preconceptions on the models. In this paper we present a few examples about modeling

nervous system functions with ANN subjected to evolutionary selection by genetic algorithms. This

methodology, called Evolutionary Connectionism, proved feasible and profitable in studying both

sensorimotor and cognitive nervous functions through simulation.