The blockade of spinal glycine receptors with intrathecal (i.t.) strychnine produces segmentally-localized allodynia in the rat; a reversible and highly reproducible effect that is attained without peripheral or central nerve injury. We investigated the effect of i.v. mexiletine, an orally active congener of lidocaine, on strychnine allodynia and compared the dose-response relationship of mexiletine in normal (noxious paw pinch) versus abnormal (i.t. strychnine) nociceptive conditions. In addition, we determined the dose-response effect of i.t. AP-7 (an NMDA antagonist) on strychnine allodynia. Male, Sprague-Dawley rats, fitted with chronic i.t. catheters, were lightly anesthetized with urethane. Stimulus evoked changes in blood pressure and heart rate were recorded from the left carotid artery and cortical electroence-phalographic (EEG) activity was continuously monitored using subdermal needle electrodes. After i.t. strychnine (40 micrograms), repetitive brushing of the hair (hair deflection) evoked a progressive increase in mean arterial pressure and heart rate, an abrupt motor withdrawal response, and desynchronization of the EEG, equivalent to those elicited by the chemical nociceptive agent, mustard oil (without strychnine). Pretreatment with mexiletine (5-30 mg/kg i.v. 5 min before i.t. strychnine) dose-dependently inhibited the responses evoked by noxious hind paw pinch (no strychnine) and hair deflection (after i.t. strychnine) with equal potency (ED50's = 9.1-17 mg/kg). Below 30 mg/kg, this effect was achieved without a change in EEG synchrony (cortical activity reflecting the level of anesthesia) and without affecting motor efferent pathways. Strychnine allodynia was also significantly blocked by i.t. AP-7. The ED50's and 95% confidence intervals were 1.1 micrograms (0.7-1.8) for mean arterial pressure, 1.7 micrograms (0.5-6.0) for heart rate, and 0.4 microgram (0.07-2.0) for withdrawal duration. Cortical EEG synchrony was unchanged after i.t. AP-7 consistent with a spinal site of action. The data indicate that: (i) robust allodynia can be selectively induced with i.t. strychnine in animals whose somatosensory systems are otherwise normal; (ii) sub-anesthetic doses of i.v. mexiletine inhibit the abnormal responses to low-threshold (A-fiber) afferent input in the strychnine model of allodynia (i.e., in the absence of peripheral or central nerve injury) at doses which affect normal nociception; and (iii) in the presence of i.t. strychnine, low-threshold afferent input activates a spinal NMDA-receptor mediated process normally restricted to noxious afferent input. Systemic mexiletine may have an important spinal site of action in abnormal pain states.