Striatal dopamine in attentional learning: A computational model

This work describes a neural network model which simulates a discrete part of the dopaminergic striatal circuitry involved in reinforcement learning. We consider the proposal that learning by reinforcement is acquired by a he terosynaptic mechanism affecting plasticity of corticostriatal synapses, un der the modulatory control of DA neurons. The simulation results seem to be successful since they are consistent with the literature that shows a tran sfer of DA neuron firing from the primary reinforcement stimulus to the con ditioned stimulus after learning. We also present some simulations, using t he same model, of two learning paradigms, blocking and overshadowing, which may be mediated by this dopamine circuitry. (C) 1999 Elsevier Science B.V. All rights reserved.