BackgroundThe mechanisms associating sleep and migraine are unknown. No previous polysomnographic (PSG) or pain-threshold (PT) study has compared patients with sleep-related migraine attacks (SM), non-sleep related migraine attacks (NSM) and healthy controls.MethodsWe have performed a blinded, prospective exploratory study with case–control design. Thirty-four healthy controls, 15 patients with SM and 18 patients with NSM had interictal PSG heat-, cold- and pressure PT (HPT, CPT, PPT) recordings and completed diary- and questionnaire on sleep and headache related aspects.ResultsNSM patients had more slow-wave sleep (SWS) and more K-bursts than SM patients (K-bursts: p = 0.023 and SWS: p = 0.030) and controls (K-bursts: p = 0.009 and SWS: 0.041). NSM patients also had lower HPT and CPT than controls (p = 0.026 and p = 0.021). In addition, SM patients had more awakenings and less D-bursts than controls (p = 0.025 and p = 0.041).ConclusionSM- and NSM patients differed in objective-, but not subjective sleep quality. NSM patients had PSG findings indicating foregoing sleep deprivation. As foregoing sleep times were normal, a relative sleep deficit might explain reduced PT among NSM patients. The SM patients had signs of slightly disturbed sleep.