The paired-stimulus paradigm, presentation of a pair of identical reflex-eliciting stimuli to the supraorbital nerve (SO) with an interstimulus interval of less than 2 s, evokes a response to the second, test, stimulus which is less than that elicited by the first, conditioning, stimulus. In this study, we investigated the site of this suppression and its pharmacology in the alert guinea pig. Both the early (R1) and the late (R2) component of the SO-evoked blink reflex exhibited suppression in the paired-stimulus paradigm. Initiation of suppression appeared to be specific to the afferent limb of the reflex rather than the result of motor activity generated by the conditioning stimulus. Neither acoustic conditioning stimuli nor air puffs that elicited blinks via another branch of the trigeminal nerve suppressed the test response. Extremely weak SO shocks, however, that did not directly elicit a reflex, caused suppression of the response to subsequent SO stimuli of normal intensity. Paired stimulus suppression of the R1 component appeared to involve activation of GABAB receptors within the spinal trigeminal nucleus. Both systemic injections and microinjections of baclofen into the spinal trigeminal nucleus enhanced R1 suppression, whereas identical injections of CGP35348, a GABAB antagonist, diminished R1 suppression. Furthermore, single-unit recordings in alert animals revealed that spinal trigeminal neurons exhibited suppression in the paired-stimulus paradigm that resembled that of the R1 component of the blink reflex. These findings showed that sensory gating underlies paired-stimulus suppression of the SO-evoked blink reflex and that activation of GABAB receptors plays an important role in this process.