AbstractImpairments in central reward processing constitute an important aspect of the negative symptoms of schizophrenia. Despite its clinical relevance, the etiology of deficient reward processing in schizophrenia remains largely unknown. Here, we used an epidemiologically informed mouse model of schizophrenia to explore the effects of prenatal immune activation on reward-related functions. The model is based on maternal administration of the viral mimic PolyI:C and has been developed in relation to the epidemiological evidence demonstrating enhanced risk of schizophrenia and related disorders following prenatal maternal infection. We show that prenatal immune activation induces selective deficits in the expression (but not acquisition) of conditioned place preference for a natural reward (sucrose) without changing hedonic or neophobic responses to the reward. On the other hand, prenatal immune activation led to enhanced place preference for the psychostimulant drug cocaine, while it attenuated the locomotor reaction to the drug. The prenatal exposure did not alter negative reinforcement learning as assessed using a contextual fear conditioning paradigm. Our findings suggest that the nature of reward-related abnormalities following prenatal immune challenge depends on the specificity of the reward (natural reward vs drug of abuse) as well as on the valence domain (positive vs negative reinforcement learning). Moreover, our data indicate that reward abnormalities emerging in prenatally immune-challenged offspring may, at least in part, stem from an inability to retrieve previously established context–reward associations and to integrate such information for appropriate goal-directed behavior.