Fig. 3: Objective sleep efficiency predicts antidepressant response following CBT-I.

Added variable plots illustrating the unique relationships between the change in depression symptom severity and pre-treatment A objective sleep efficiency and B self-reported insomnia symptoms. Following a residualized change model approach, the y-axes represent post-treatment depression symptoms after accounting for baseline depression levels, age, and sex. The x-axis represents pre-treatment sleep efficiency and pre-treatment insomnia symptom severity after accounting for the same covariates. The solid line indicates the partial regression slope, with the shaded area showing the 95% CI.