Extended Data Fig. 4: Sex difference of the association between sleep duration and mental health, cognitive function and brain structure.
The nonlinear association between sleep duration and anxiety symptom was more significant in female participants (F female = 622.6, n = 533, 2878, 15240, 36239, 26712, 4712 and 1000 participants respectively; F male = 417.2, n = 347, 2126, 12587, 29710, 18677, 3229 and 665 participants respectively), whereas mania symptoms showed more significant association with sleep duration for male participants (F female = 140.3, n = 550, 2928, 15466, 36672, 26988, 4774 and 1025 participants respectively; F male = 145.0 respectively, n = 354, 2148, 12673, 29893, 18780, 3252 and 670 participants respectively).Fluid intelligence were found to have a greater nonlinear association with sleep duration in females compared with males (F female = 272.7, n = 940, 3981, 16606, 32724, 25625, 5051 and 1502 participants respectively; F male = 205.4, n = 673, 3144, 14934, 29192, 20019, 4018 and 1223 participants respectively) while pair matching were more associated with sleep duration in males (F female = 85.8, n = 3087, 11892, 48704, 98567, 79070, 15934 and 4922 participants respectively; F male = 104.1, n = 2367, 9356, 44236, 88501, 61182, 12315 and 3980 participants respectively). For brain structure, female participants demonstrated a more significant association between sleep duration and cortical volumes (rh, F female = 29.1, n = 192, 991, 4221, 8375, 5746, 1158 and 249 participants respectively; F male = 14.7, n = 118, 631, 3445, 7523, 5592, 1231 and 238 participants respectively) while cortical thickness was more significantly associated with sleep duration for males (F female = 2.89, n = 192, 991, 4221, 8375, 5746, 1158 and 249 participants respectively; F male = 20.0, n = 118, 631, 3445, 7523, 5592, 1231 and 238 participants respectively). Lines are fitted nonlinear model indicating fitted mean value and error bar is standard error of the mean.
