Statistics

We compared the outcome variables using linear mixed-effects models (fixed factor: condition, random factor: subject). Overnight differences were calculated as morning–evening values. If the linear mixed-effects model was significant for condition, we derived post hoc P-values for ISI1_High and ISI1_Low each compared with SHAM corrected for multiple comparisons for each model separately using the Hochberg method. Visual inspection of the residual plots of the linear models did not reveal any obvious deviations from normality or homoscedasticity. P-values <.05 were considered statistically significant.