This study contrasted the effect of hypnosis on self-reported pain and changes in a nociceptive brainstem reflex, the blink reflex (BR), in 39 women with temporomandibular disorder. The patients were randomized to hypnosis or control (nonhypnotic relaxation). Pain intensity was assessed 3 times daily on a 0 to 10 numerical rating scale. BRs were elicited by electrical stimulation with a nociceptive-specific electrode and recorded before and after treatment at pain threshold (Ip) and supra threshold (2×Ip). Significant reduction of pain intensity was observed in the hypnosis group from 4.5±2.1 at baseline to 2.9±2.4 after treatment (P<0.001). The pain reduction was generally unrelated to changes in the BR, with the exception being a lowered ipsilateral R2 BR component at the right side supra threshold (P=0.034). Hypnosis thus seems to reduce complex temporomandibular disorder pain, most likely because of cortical changes with little, if any, involvement of brainstem reflex pathways.