BACKGROUND Diffuse noxious inhibitory control (DNIC) is a phenomenon whereby wide dynamic range neurons are selectively and powerfully inhibited through the central nervous system by noxious stimuli heterotopically applied to a body area distant from their excitatory receptive fields. Previous work has shown that systemic administration of an alpha1-adrenoceptor agonist, phenylephrine (PE), blocked the DNIC. We hypothesized that descending inhibitory pathways mediate the DNIC mechanism and that the neural network of the DNIC loop exists in the middle brainstem, likely in a more rostral part than formerly assumed, possibly the nucleus raphe magnus (RMg). The aim of this study was to determine whether DNIC is directly modulated by PE when administered close to the RMg. METHODS The experiments were performed on anesthetized male Sprague-Dawley rats. For administration of different drugs close to the RMg, the tip of a 33-gauge cannula was placed into an area close to the RMg as determined using the atlas of Paxinos and Watson. Single square-wave electrical stimuli were applied to the digits of the left hindpaw. The C-fiber reflex response elicited by electrical stimulation within the receptive field of the ipsilateral sural nerve was recorded from the biceps femoris muscle in the absence and presence of noxious tail immersion in warm water at 50 degrees C. The DNIC effect was calculated from a recorded electromyogram as the "inhibition rate." Saline (0.05 microL) or PE (0.05 microg/0.05 microL) was microinjected close to the RMg through the cannula. The C-fiber reflex evoked by electromyographic activity was recorded the same way. The inhibition rate of the C-fiber reflex was compared before and after administration of drugs. A paired t test was used for statistical comparison between same drug administration groups, and 1-way analysis of variance and Bonferroni multiple comparison were used for statistical analysis between different drugs. At the end of all experiments, the tissue-contacting end of the cannula tip was cauterized with an electric current to localize the drug administration site. The brain was removed, sliced in coronal sections, and stained with hematoxylin and eosin. RESULTS The C-fiber reflex inhibited by noxious thermal stimuli (DNIC) was significantly blocked after the injection of PE close to the RMg. CONCLUSION Direct administration of PE close to the RMg inhibited DNIC, thereby affecting and modulating the intrinsic pain inhibition system. These findings suggest that the RMg may be involved in the regulation of DNIC.