The study examined the effects of cocaine on learning and performance of a classically conditioned heart rate (HR) discrimination in rats involving two auditory conditioned stimuli (CSs). In the discrimination protocol, one CS (CS+) was paired with the shock unconditioned stimulus (US) on a consistent basis and the other CS (CS-) was always presented alone. Four groups received an IP injection of 1, 3, 10, or 30 mg/kg cocaine and a fifth group received saline. Shortly after the injections, all groups were given six CS-alone trials, followed by 24 randomly sequenced discrimination conditioning trials (12 CS+ and 12 CS-). Approximately 72 h later, all groups were given six test trials with each CS in the absence of cocaine to evaluate the presence or absence of discrimination learning. All cocaine groups showed impaired discrimination performance on the discrimination conditioning trials, reductions in early pretest CS-alone responses, and reductions in resting HR. However, on the non-drug test trials discrimination performance was normal in all cocaine groups. The results established that in spite of major changes in HR dynamics, learning of the HR discrimination was not affected by cocaine but that cocaine did interfere with the performance of the discrimination. Except for the highest 30 mg group, the performance decrement appeared to be related to a cocaine-produced reduction in the capacity to inhibit bradycardia responding to the safe CS-. It was suggested that this loss of inhibitory control may have been due to cocaine changes in a corticothalamic pathway that controls inhibition of bradycardia to a safe CS-.