Since blockers of excitatory transmission have been shown to reduce anoxic and ischemic neuronal damage, augmentation of inhibitory transmission by agents such as midazolam might have a similar protective effect. Rat hippocampal slices were maintained in vitro and used to determine whether and by what mechanism midazolam improves recovery of evoked responses after anoxia. The Schaffer collateral pathway in the slice was stimulated electrically, and an extracellular potential, the evoked population spike, was recorded from the CA1 pyramidal cells, which are postsynaptic. The slices were made anoxic by substituting artificial cerebrospinal fluid aerated with 95% nitrogen-5% carbon dioxide for fluid aerated with 95% oxygen-5% carbon dioxide. Percentage recovery was expressed as the amplitude of the evoked population spike 60 min after anoxia divided by its preanoxic amplitude. Protection in this model is defined as a significant (P less than 0.05) improvement in percentage recovery compared to the recovery of untreated slices. There was no recovery of the response recorded from CA1 pyramidal cells after 5 min of anoxia (4 +/- 2%) (mean +/- standard error of the mean [SEM]). Slices were treated with midazolam 10 min before, during, and 10 min after anoxia. Midazolam (1 microM) did not enhance recovery after anoxia when dissolved either in water (3 +/- 3%) or in dimethyl sulfoxide (DMSO) (1 +/- 1%). A higher concentration of midazolam (100 microM) did enhance recovery when dissolved in DMSO (27 +/- 7%) but not when dissolved in water (5 +/- 2%). To test whether prolonged pretreatment with midazolam dissolved in water would enhance recovery, slices were treated for 30 min prior to anoxia.(ABSTRACT TRUNCATED AT 250 WORDS)