Current models of addiction are converging on the idea that aberrant function and remodeling of neural circuits cause the compulsive drug use despite negative consequences that define the disease. According to this hypothesis, drug reward would trigger specific forms of synaptic plasticity, which in susceptible subjects would become persistent. I will test this model by characterising the behaviour and neural changes in mice, which have had the opportunity of self-stimulate VTA dopamine neurons. Several stages from acute reinforcement to cue-associated seeking behaviour and resistance to punishment will be tested. The special emphasis will be on manipulations that allow to establish links of causality.