Fig. 5

Bayesian posterior distributions for migraine variables predicting sleep duration. Having an attack did not predict alterations to sleep (A M = -0.03, 95% highest density interval (HDI) [-0.07—0.01]), but greater pain intensity (B M = 0.13, 95% HDI [0.06—0.20]) results in more hours slept the same night. Density plots show posterior distributions for the two tested models of the population means of migraine-related predictor variables on total hours slept. Plots represent the posterior distribution, density estimate. The x-axis represents b-value estimates of the parameter. Error bars indicate the 95% highest density interval. VAS = visual analogue scale