Male and female Long-Evans rats were trained to self-administer (SA) cocaine in a two-week SA experiment. Availability of cocaine was signaled by a discrete tone cue and drug administration was contingent upon a nose-poke response during the period when the tone was presented. Firing rates on trials preceding the first infusion (missed opportunities), were compared to the trial immediately leading to first self-administered infusion (first hit).

When movement-free, tone-evoked, firing rates of core and shell neurons were analyzed, it was observed that both core and shell showed robust task-related changes in firing. NAc neurons of both males and females showed differences in tone-evoked firing between all of the missed trials prior to the first infusion, vs the first hit of the day. Consistent with learning by core and shell neurons of association between tone and cocaine, only late sessions showed the difference between missed trials and the tone immediately leading to the first self-administered infusion of the day. For males the effect of neuron subregion did not modify the neural response to "Hits" vs "Misses", whereas for females there was a major increase for shell neurons, however only to "Missed" trials. Together these results suggest that the core is not the first and only stage of processing of limbic inputs necessary for cue-induced drug seeking in nucleus accumbens, but rather processing of those cues might involve NAc shell, which is anatomically upstream of the core.