Time-series data and the "migraine generator"

Objective.-To examine circadian and menstrual patterns of migraine frequency in order to identify properties of the putative migraine generator. Methods.-Analysis of circadian and menstrual migraine attack frequency distributions, using between-interval differences as estimates, at each time point, of the derivative of the equation that would, theoretically, model the observed oscillating functions. Results.-Circadian and menstrual analyses exhibit many similarities that are consistent with a single migraine generator in the final common pathway for expression of the phenotype. These analyses are inconsistent with a role for cortisol or ACTH to be activators of the putative migraine generator, could be consistent with visual stimulation as a deactivator of the generator, and may also be consistent with a low threshold concentration of estradiol and/or progestagens as inactivators of the migraine generator. Conclusions.-Analysis of time-series data using differentials uncovers some previously unidentified chronobiological properties of migraine, helps select amongst various candidate provocative factors, and suggests some properties of a putative migraine generator.