Treatment with non-competitive N-methyl-d-aspartate (NMDA) antagonists such as phencyclidine or ketamine have been shown to induce schizophrenia-like psychotic and cognitive symptoms in humans and animals. However, there have been a number of contradictory findings regarding the effects of repeated treatment with these drugs on working memory in experimental animals. We hypothesized that processes dependent on dopamine transmission in the medial prefrontal cortex (PFC) may be more sensitive to disruption following these treatment. We assessed the effects of repeated treatment with ketamine on working memory performance using a delayed spatial win-shift procedure conducted on a radial-arm maze, dependent on a neural circuit linking hippocampal and dopamine inputs to the medial PFC. Rats were trained on the task prior to drug exposure, after which they were subjected to one of two dosing regimes of ketamine (30 mg/kg twice a day for either 5 or 10 days). After a 10 day withdrawal period, they were re-tested on the task for 15 days. Ketamine treatment for 10 days, but not 5 days, increased the number of errors and days to re-achieve the criterion on the delayed task. However, in a separate group of rats, subchronic ketamine treatment (10 days) did not affect performance of the non-delayed random foraging task, dependent on the hippocampus, but not the PFC. These results indicate that working memory performance assessed with these procedures is sensitive to disruption following repeated exposure to ketamine. Impairments in working memory induced by these treatments are not attributable to dysfunction of motivational, motor, short-term or spatial memory processes. The use of these procedures may prove useful in modeling impairments in this executive function observed in schizophrenia.