The present study was designed to determine the effect of sleep on reflex pharyngeal dilator muscle activation by stimuli of negative airway pressure in human subjects. Intra-oral bipolar surface electrodes were used to record genioglossus electromyogram (EMG) responses to 500 ms duration pressure stimuli of 0 and -25 cmH2O applied, via a face-mask, in four normal subjects. Stimuli were applied during early inspiration in wakefulness and in periods of non-rapid-eye-movement (non-REM) sleep, defined by electroencephalographic (EEG) criteria. The rectified and integrated EMG responses to repeated interventions were bin averaged for the 0 and -25 cmH2O stimuli applied in wakefulness and sleep. Response latency was defined as the time when the EMG activity significantly increased above prestimulus levels. Response magnitude was quantified as the in ratio of the EMG activity for an 80 ms post-stimulus period to an 80 ms prestimulus period; data from after the subject's voluntary reaction time for tongue protrusion (range, 150-230 ms) were not analysed. Application of the -25 cmH2O stimuli caused genioglossus muscle activation in wakefulness and sleep, but in all subjects response magnitude was reduced in sleep (mean decrease, 61%; range, 52-82%; P = 0.011, Student's paired t test). In addition, response latency was increased in sleep in each subject (mean latency awake, 38 ms; range, 30-50 ms; mean latency asleep, 75 ms; range, 40-110 ms; P = 0.072, Student's paired t test). Application of the -25 cmH2O stimuli caused arousal from sleep on 90% occasions, but in all cases the reflex genioglossus muscle responses (maximum latency, 110 ms) always proceeded any sign of EEG arousal (mean time to arousal, 643 ms; range, 424-760 ms). These results show that non-REM sleep attenuates reflex genioglossus muscle activation by stimuli of negative airway pressure. Attenuation of this reflex by sleep may impair the ability of the upper airway to defend itself from suction collapse by negative pressures generated during inspiration; this may have implications for the pathogenesis of obstructive sleep apnoea.