Rewards serve several important behavioural functions related to motivation, pleasure and learning. At the cerebral level, reward processing is thought to rely on a well-defined set of brain regions known as the “reward system”, whose disruption has been linked to maladaptive behaviours such as addiction. However, a wide variety of rewards exists, and there is no reason to think that the brain responds equivalently to all of them. In order to test this assumption, we used functional Magnetic Resonance Imaging (fMRI), with three different perspectives. In a first experiment, we investigated the distinction existing between so-called primary (i.e. primitive and concrete) rewards and secondary (i.e. evolved and abstract) rewards, studied here through the examples of erotic pictures and monetary gains. In addition to a common brain network recruited regardless of reward type, our results revealed a functional dissociation within the orbitofrontal cortex (OFC), whose posterior part responded specifically to primary rewards, while its anterior part responded specifically to secondary rewards. Interestingly, this finding supports the idea of a complexity gradient along the postero-anterior axis of the OFC. In a second study, we used a quantitative meta-analytic approach to compare the brain activations reported in the literature in response to monetary gains, pleasant tastes and erotic pictures. The results are in line with the conclusions drawn from the first experiment, and confirm the existence of reward-type-specific responses in the brain. Finally, we conducted a third study focusing on pathological gambling, and aiming to test the hypothesis of an imbalance in the sensitivity to monetary versus non-monetary rewards. The results bring evidence supporting this view, and essentially demonstrate an impaired processing of non-monetary rewards in the ventral striatum of gamblers. Overall, this work sheds new light on the functional architecture of the reward system, both in healthy subjects and pathological gamblers
