Although correlations have been reported between maternal cocaine use and impaired attention in exposed children, interpretation of these findings is complicated by the many risk factors that differentiate cocaine-exposed children from SES-matched controls. For this reason, the present dose-response study (0, 0.5, 1.0, or 3.0 mg/kg cocaine HCl) was designed to explore the effect of prenatal cocaine exposure on visual attention in a rodent model, using an intravenous injection protocol that closely mimics the pharmacokinetic profile and physiological effects of human recreational cocaine use. In adulthood, animals were tested on an attention task in which the duration, location, and onset time of a brief visual cue varied randomly between trials. The 3.0 mg/kg exposed males committed significantly more omission errors than control males during the final 1/3 of each testing session, specifically on trials that followed an error, which implicates impaired sustained attention and increased reactivity to committing an error. During the final 1/3 of each testing session, the 0.5 and 1.0 mg/kg exposed females took longer to enter the testing alcove at trial onset, and failed to enter the alcove more frequently than control females. Because these effects were not seen in other tasks of similar duration and reinforcement density, these findings suggest an impairment of sustained attention. This inference is supported by the finding that the increase in omission errors in the final block of trials in each daily session (relative to earlier in the session) was significantly greater for the 1.0 mg/kg females than for controls, a trend also seen for the 0.5 mg/kg group. Unlike the cocaine-exposed males, who remain engaged in the task when attention is waning, the cocaine-exposed females appear to opt for another strategy; namely, refusing to participate when their ability to sustain attention is surpassed.