We have simulated behaviour of a rodent solving a spatial task using a simple artificial neural network. The model behaves similarly to the rodent; in particular it anticipates the reward accordingly.