- Relapse to drug use in humans can be induced by exposure to drug-associated cues. The ability of drug cues to provoke 'relapse' has been studied in laboratory animals using a reinstatement model in which resumption of drug seeking is assessed after extinction of drug-reinforced responding. In this model, there are no adverse consequences of drug-seeking behavior. However, in humans, abstinence is often self-imposed, and relapse episodes likely involve making a choice between the desire for the drug and the negative consequences of pursuing it (a conflict situation). In this paper, we describe a conflict model of cue-induced relapse in rats that approximate the human condition. Rats were trained to lever press for cocaine; infusions were paired with a discrete light cue. An 'electric barrier' was then introduced by electrifying the floor area near the levers. Responding decreased over days with increasing shock intensities, until the rats did not approach the levers for 3 days. Subsequently, the effect of intermittent noncontingent light-cue presentations on resumption of lever responding (relapse) was assessed in extinction tests, with the electric barrier remaining activated; during testing, lever presses led to contingent light-cue presentations. Noncontingent cue exposure led to resumption of lever presses during the relapse tests in 14 of the 24 rats. Surprisingly, 24 h later, 11 of the 24 rats resumed lever responding in a subsequent post-noncontingent cue test under similar extinction conditions. Large individual differences in responding were observed during both tests. At its current stage of development, the conflict relapse model appears particularly suitable for studying individual differences in cue-induced relapse to cocaine seeking or factors that promote this relapse.