This study examined whether the antinociception produced by glutamatergic stimulation of neurons in the nucleus raphe magnus (NRM) or nucleus reticularis gigantocellularis pars alpha (NGCp alpha) is mediated by activation of GABAA receptors in the spinal cord. Two approaches were used. The first approach determined the ability of intrathecally (i.t.) administered bicuculline, a competitive GABAA receptor antagonist, to attenuate the antinociception produced by microinjection of L-glutamate in the NRM or NGCp alpha. Bicuculline was selected on the basis of an initial study that determined that 0.3 micrograms i.t. bicuculline caused a 3.4-fold rightward shift in the dose-effect relationship of the i.t.-administered GABAA agonist, isoguvacine, without producing allodynia or alterations in nociceptive threshold. The GABAA antagonist SR 95531 was judged unsuitable because it caused only a modest 1.7-fold rightward shift in the dose-effect relationship of isoguvacine at doses that did not produce allodynia. The second approach determined the ability of i.t. administered diazepam, a benzodiazepine receptor agonist, to enhance the antinociception produced by microinjection of L-glutamate in these nuclei. Diazepam was selected because of its ability to enhance the actions of GABA at the GABAA receptor. Rats were pretreated with i.t. administration of 0.3 micrograms bicuculline, 40 micrograms diazepam, or vehicle, after which 30 nmol L-glutamate was microinjected into the NRM or NGCp alpha. Microinjection of L-glutamate into the NRM or NGCp alpha in vehicle-pretreated rats significantly increased tail flick latency. The antinociception produced by microinjection of L-glutamate in the NGCp alpha was antagonized by bicuculline and enhanced by diazepam. In contrast, the antinociception evoked from sites in the NRM was only partially attenuated by bicuculline and was not enhanced by diazepam. In an ancillary experiment, i.t. administration of 0.3 micrograms bicuculline or 40 micrograms diazepam did not alter tail skin temperature or nociceptive threshold, suggesting that their effects on glutamate-induced antinociception were not secondary to alterations in tail skin temperature. Taken together, these results support the hypothesis that the antinociception produced by activation of neurons in the NGCp alpha, but not the NRM, is mediated in part by an action of GABA at GABAA receptors in the spinal cord.