Sleep–wake behavioral characteristics associated with depression symptoms: findings from the Multi-Ethnic Study of Atherosclerosis

Abstract Study Objectives To help prioritize target/groups for experimental intervention studies, we characterized cross-sectional associations between 24-hour sleep–wake measures and depression symptoms, and evaluated if similar sleep–wake–depression relationships existed in people with and without higher insomnia severity. Methods Participants had ≥3 days of actigraphy data (n = 1884; mean age = 68.6/SD = 9.1; 54.1% female). We extracted 18 sleep, activity, timing, rhythmicity, and fragmentation measures from actigraphy. We used individual and multivariable regressions with the outcome of clinically significant depression symptoms (Center for Epidemiologic Studies Depression Scale ≥ 16). We conducted sensitivity analyses in people with higher insomnia severity (top quartile of the Women’s Health Initiative Insomnia Rating Scale total score). Results From separate models in the overall sample, the odds of having depression symptoms were higher with: later timing (e.g. activity onset time odds ratio OR/1 SD = 1.32; 95% confidence interval CI: 1.16 to 1.50), lower rhythmicity (e.g. pseudo-F OR/1 SD = 0.75; 95% CI: 0.66 to 0.85), less activity (e.g. amplitude OR/1 SD = 0.83; 95% CI: 0.72 to 0.95), and worse insomnia (OR/1 SD = 1.48, 95% CI: 1.31 to 1.68). In multivariable models conducted among people with lower insomnia severity, later timing, lower rhythmicity, and higher insomnia severity were independent correlates of depression. In people with higher insomnia symptom severity, measures of later timing were most strongly associated with depression symptoms. Conclusions These correlative observations suggest that experimental studies are warranted to test if: broadly promoting 24-hour sleep–wake functioning reduces depression even in people without severe insomnia, and if advancing timing leads to depression symptom reductions in people with insomnia.