Hyperdopaminergic states in mental disorders are associated with disruptive deficits in decision-making. However, the precise contribution of topographically distinct mesencephalic dopamine pathways to decision-making processes remains elusive. Here we show, using a multidisciplinary approach, how hyperactivity of ascending projections from the ventral tegmental area (VTA) contributes to faulty decision-making in rats. Activation of the VTA-nucleus accumbens pathway leads to insensitivity to loss and punishment due to impaired processing of negative reward prediction errors. In contrast, activation of the VTAprefrontal cortex pathway promotes risky decision-making without affecting the ability to choose the economically most beneficial option. Together, these findings show how malfunction of ascending VTA projections affects value-based decision-making, providing a mechanistic understanding of the reckless behaviors seen in substance abuse, mania, and after dopamine replacement therapy in Parkinson's disease.Impaired decision-making can have profound negative consequences, both in the short and in the long term. As such, it is observed in a variety of mental disorders, such as mania 1,2 , substance addiction [3][4][5][6] , and as a side effect of dopamine (DA) replacement therapy in Parkinson's disease 7,8 . Importantly, these disorders are associated with aberrations in DAergic neurotransmission 9,10 , and DA has been implicated in decision-making processes [11][12][13] . However, ascending DAergic projections from the ventral mesencephalon are anatomically and functionally heterogeneous [14][15][16] and the contribution of these distinct DA pathways to decision-making processes remains elusive.The mesocorticolimbic system, comprising DA cells within the ventral tegmental area (VTA) that mainly project to the nucleus accumbens (NAc; mesoaccumbens pathway) and medial prefrontal cortex (mPFC; mesocortical pathway), has an important role in value-based learning and decision-making [14][15][16] . When an experienced reward is better than expected, the firing of VTA DA neurons increases, thereby signaling a discrepancy between anticipated and experienced reward to downstream regions. Conversely, when a reward does not fulfill expectations, DA neuronal activity decreases. This pattern of DA cell activity is the basis of reward prediction error (RPE) theory [17][18][19][20] , which describes an essential mechanism through which organisms learn to flexibly alter their behavior when the costs and benefits associated with different courses of action shift. Although the relevance of RPEs in value-based learning is widely acknowledged, little is known about how different VTA target regions process these DA-mediated error signals, and how this ultimately leads to adaptations in behavior.Here, we used projection-specific chemogenetics combined with behavioral tasks, pharmacological interventions, computational modelling, in vivo microdialysis and in vivo neuronal population recordings to investigate how different ascending V...