Abstract Transient, global cerebral ischemia (TGCI) leads to hippocampal damage and disruption of spatial learning and memory. The immunosuppressant, tacrolimus (FK506), prevents TGCI-induced hippocampal neurodegeneration, but its effectiveness in promoting the recovery of learning and memory performance after TGCI has been little investigated. Here, we use a confined version of the aversive, non-food rewarded radial maze to evaluate further the effects of FK506 on TGCI-induced learning and memory deficits. In the first experiment, rats were rendered ischemic (15 min 4-VO) and 20 days later were tested for acquisition of the radial maze task over 15 consecutive days (post-operative training). In the second experiment, naive rats were trained for 10 days and subjected to TGCI (pre-operative training); retention of task performance was assessed on days 31, 35 and 39 post-ischemia. Acquisition and retention performances were expressed as a) latency to find a goal box, b) number of reference memory errors, and c) number of working memory errors. Data are presented both across daily training sessions (15 days, 3-day blocks) and as a total value (summed over the 15 days). Histological examination was performed on the day after behavioral testing. In both experiments, FK506 (1.0 mg/kg) was given i.v. at the beginning of reperfusion, followed by doses applied intraperitoneally (i.p.) 6, 24, 48 and 72 h post-ischemia. TGCI markedly disrupted both acquisition and retention performance ( p < 0.0001–0.05). Treatment with FK506 did not prevent the TGCI-induced acquisition and retention deficits, independently of whether performances were quantified ‘daily’ or as a ‘total’ value. In contrast, FK506 reduced hippocampal damage significantly compared to the vehicle alone ( p < 0.001–0.05). We conclude that the present study did not confirm our earlier behavioral data, and suggest that FK506 is not effective in treating the behavioral outcomes of TGCI, despite its efficacy in reducing CA1, hippocampal damage. However, further studies including other behavioral tasks and more extensive neurohistological analysis, are needed to better elucidate the effectiveness of FK506 in promoting functional recovery in models of transient, global cerebral ischemia.