Besides its well-established efficacy in epilepsy, vagus nerve stimulation (VNS) may be of potential interest in pain treatment. It has, however, not yet been assessed in animal pain models with the devices and stimulation protocols used in humans. We have therefore studied in awake rats the effects of left cervical VNS on trigeminal nociception using an implantable electrode and stimulator (NCP-Cyberonics). VNS was applied for 24h at 2 mA intensity, 20 Hz frequency, 0.5 ms pulse width and a duty cycle of 20s ON/18s OFF. As a nociceptive stimulus, we injected formalin into the left mystacial vibrissae, assessed behaviour for 45 min and sacrificed the animals 45 min later. Fos-immunoreactive (Fos-Ir) neurons were counted in laminae I-II of trigeminal nucleus caudalis (TNC) on both sides. We used three groups of control animals: VNS without formalin, formalin without VNS and sham VNS (implanted without stimulation or formalin). Whereas sham VNS had no significant effect, VNS alone increased Fos expression in ipsilateral TNC in addition to the expected increase in nucleus tractus solitarius. It also significantly attenuated the increase of Fos-Ir neurons observed in ipsilateral TNC laminae I-II after formalin injection. If the proper VNS effect on Fos-expression was subtracted, the reduction of formalin-induced nociceptor activation was 55%. VNS also reduced nociceptive behaviour on average by 96.1% during the early phase (0-6 min) and by 60.7% during the late phase (6-45 min) after the formalin injection. These results suggest that VNS applied with a device used in human therapy may have in awake rats a significant antinociceptive effect in a model of trigeminal pain.