Cocaine administration has been shown to affect several sites in the limbic forebrain. The nucleus accumbens has been implicated as an important site for the reinforcing aspects of this drug whereas, the amygdala and hippocampus may be more involved in drug conditioning and/or drug induced changes in the perception of stimuli. In the present study, auditory event-related potentials (ERPs) were utilized to explore the effects of cocaine on sensory processing in several limbic sites. Eleven adult male Wistar rats were stereotaxically implanted with electrodes aimed at the dentate gyrus, dorsal hippocampus (CA1-CA2), amygdala, and nucleus accumbens. The rats received intraperitoneal injections of either saline, 2.5, 5.0, 10.0 or 20 mg/kg of cocaine. The ERPs were recorded in response to an auditory 'oddball' paradigm consisting of frequently and infrequently presented tones. Cocaine was found to produce a dose dependent decrease in the variance of the amplitude of several ERP components in the amygdala and the hippocampal formation but not in the nucleus accumbens. Cocaine also produced a decrease in the latency of the N1 component in the amygdala. No significant changes were observed in the amplitude of the ERP components following any of the cocaine doses studied, suggesting that cocaine may not increase the salience of auditory stimuli in these structures. The observed reduction in the variance of the ERPs may suggest an increased focussing of the animals' attention to the auditory stimuli. These studies also suggest that the sensory/cognitive aspects involved in cocaine's actions, as quantified by ERPs, may primarily involve the hippocampal formation and the amygdala, but not the nucleus accumbens.