We present a hypothetical neurocomputational model that combines a set of neural circuits at the molecular, cellular, and system levels and accounts for several neurobiological and behavioral processes leading to nicotine addiction. We propose that combining changes in the nicotinic receptor response, expressed by mesolimbic dopaminergic neurons, with dopamine-gated learning in action-selection circuits, suffices to capture the acquisition of nicotine addiction. We show that an opponent process enhanced by persistent nicotine-taking renders self-administration rigid and habitual by inhibiting the learning process, resulting in long-term impairments in the absence of the drug. The model implies distinct thresholds on the dosage and duration for the acquisition and persistence of nicotine addiction. Our hypothesis unites a number of prevalent ideas on nicotine action into a coherent formal network for further understanding of compulsive drug addiction.computational model ͉ reward T obacco addiction is a multistage process involving persistent cycles of chronic smoking (1, 2). It is tied to long-lasting effects on mesolimbic dopaminergic (DA) pathways by nicotine, the main addictive substance in tobacco smoke (2, 3). Although the pharmacological target of nicotine is now well identified (4), how nicotine binding translates into addictive behavior remains enigmatic. Particularly puzzling is the ease with which nicotine addiction is acquired and resists despite its limited (or even negative) hedonic impact (3).Both reinforcement learning and opponent processes have been proposed to play a significant role in the development of addiction (5, 6). However, direct DA-dependent reinforcement learning does not conclusively treat the issue of persistence of addictive behaviors to extinction (5), and the opponent process or allostasis (6) theory does not provide a specific computational account of how addiction is acquired. To unravel two such aspects, the acquisition and persistence, we model the interplay between the phasic and the persistent effects of nicotine on the DA pathway and its assumed control of the plasticity in corticostriatal action-selection (A-S) circuits (7). We propose a hypothetical minimal computational circuit of neuronal and pharmacological processes (see Fig. 1) and apply it to a specific animal model of smoking in humans: self-administration of nicotine. Our framework implements the dynamics of interaction between the effect of nicotinic acetylcholine receptors (nAChRs) on the DA neuronal population and learning in a model of A-S. We specify the differential roles of nicotine in the positive (direct) DA response on the acquisition of nicotinetaking and the slow opponent process in its persistence.At the molecular level, nicotine actions are mediated by persistent changes in brain nAChRs involving, among others, the 2-subunit (8-11). These nAChRs modulate the excitability of the DA neurons in the ventral tegmental area (VTA) that project to striatal structures (e.g., nucleus accumbens and striatum) ...