Abstract
Background: Questionnaire-based chronotype measures may be affected by reporting bias. In contrast, 24-h accelerometry objectively captures rest-activity timing phenotypes as a behavioral proxy for chronotype. We examined whether accelerometer-derived activity timing phenotypes are linked to the onset of major depressive disorder (MDD) and anxiety disorders. Methods: With the UK Biobank, 94,344 participants free of MDD and anxiety disorders with valid accelerometry data were included. Accelerometer-derived rest-activity timing phenotypes were classified via k-means clustering of 24-h activity patterns. Cox models calculated hazard ratios (HRs) for the onset of MDD and anxiety. Brain magnetic resonance imaging (MRI) biomarkers were analyzed in 17,571 participants. Results: A total of 2,361 MDD and 2,278 anxiety cases were identified over a median 6.8-year follow-up period. Compared with participants in the accelerometer-derived daytime-active group, the night-active group had higher risks of MDD [1.30; 95% confidence interval (CI): 1.15 to 1.48] and anxiety disorders (1.27; 95% CI: 1.11 to 1.45). The early-morning-active phenotype group showed a 19% (95% CI: 8% to 29%) reduced risk of MDD. High physical activity (PA) levels in the period of 00:00 to 04:59 showed adverse effects for MDD and anxiety disorders, while high PA levels in the period of 06:00 to 08:59 showed favorable effects. In addition, the night-active group was linked to lower volume of white matter and gray matter, lower frontoparietal gray matter volumes, and lower subcortical volumes. Conclusions: Objectively measured activity timing phenotypes were associated with incident MDD and anxiety, and showed exploratory MRI correlates that may help generate mechanistic hypotheses. Whether modifying activity timing reduces mental disorder risk requires randomized controlled trials.