Neuroprotective action of MK-801 administrated pre- and postischemically, in vivo or in vitro, respectively, was studied on hippocampal slices using decapitation ischemia model. Recovery of orthodromic population spikes in CA1 region was measured during postischemic incubation of the slices with oxygenated artificial cerebrospinal fluid (ACSF). The ability of postischemically applied MK-801 to restore the electrical activity dramatically depended on the timing of its application during the reoxygenation period. When applied in vitro, together with the start of reoxygenation, MK-801 was as effective as in the case of in vivo administration before the ischemia. The delay in the in vitro administration for only a few minutes led to a dramatic decrease in the drug effectiveness. When applied in 30 min after the start of reoxygenation, MK-801 was totally ineffective. The dose/response relationship between MK-801 concentration and the amplitude of recovered orthodromic population spikes of hippocampal pyramidal neurons is logarithmic. The ED(50) value for the action of "postischemic" MK-801 is ca. 10(-5) M. Preischemic in vivo application of the drug [intraperitoneal (i.p.) injection 15 min prior to decapitation] results in ED(50) ca. 0,2 mg/kg. The slope of both dose/concentration-response curves is similar. The time course of population spike recovery after 90-min ischemia is identical for pre- and postischemic action of MK-801 (estimated for ED(50) in both cases). These data allow to suggest that "preischemic" MK-801 is predominantly active as a neuroprotector only after ischemia, within a short therapeutic window at the start of the reoxygenation period.