“…We used the data coded by the UKB study and removed the subjects with responses "do not know" or "prefer not to answer". In our phenotypic analysis, we used the white British imaging individuals in UKB phases 1 to 3 data release (released up through 2020, average n = 29,025, mean age range = (45,82), mean = 64.15, standard error = 7.67, and proportion of female was 51.6%). We fitted linear models for each pair of sleep and imaging traits, in which we adjusted for the effects of age (at imaging), age-squared, sex, age-sex interaction, age-squared-sex interaction, imaging site code, the top 40 genetic principal components (PCs) 123 , volumetric scaling, head motion, head motion-squared, brain position, and brain positionsquared 53,55 .…”