The rhythm of mental health: the relationship of chronotype with psychiatric trait dimensions and diurnal variation in psychiatric symptoms

Abstract

To advance the emergence of circadian-based therapies, this study characterized how psychiatric symptoms fluctuate across the day and vary between individuals. Using a dimensional approach, we determined how chronotype relates to 13 psychiatric traits, and modeled the temporal development of symptoms throughout the day using generalized additive mixed effects models. In this preregistered study, a subclinical sample completed 13 psychiatric trait scales and a chronotype scale at baseline (N = 515, n = 404 women, 109 men, n = 2 non-binary, M age = 32.4 years, range 18-77), followed by 22 psychiatric symptoms and behaviors rated repeatedly between ~08:00-00:00 (n = 410). Key findings are that 11 out of 13 psychiatric traits were associated with being an evening-type, ranging from depression to obsessive comulsive disorder, social anxiety, and delusional ideation, while only mania was associated with being a morning-type. Four distinct psychiatric trait factors were identified, each predicting worse symptom levels throughout the day. Fatigue-related symptoms exhibited strong time-of-day changes with evening-types experiencing worse fatigue in the morning and morning-types in the evening. Evening-types had considerably lower drive and motivation than morning-types from morning to early evening. Evening-types also had more pronounced negative emotional symptoms and ADHD-type symptoms in the evening, particularly among those high in psychiatric trait factors. These findings identified important research targets that hold promise for improving mental health outcomes, such as strategies to boost morning motivation. Furthermore, the results emphasize the relevance of incorporating circadian factors, including chronotype, into translational psychiatric research and interventions.

Fig. 1. Factor loadings.

Each panel shows the factor loadings for the constructs of the (A) transdiagnostic psychiatric traits and (B) psychiatric symptoms and behaviors. Higher factor loadings correspond to stronger loadings for the respective construct, with both positive and negative loadings. The scales are ordered according to strength of the factor loading when loading strongest on the respective construct. More saturated colored bars indicate the items that load ≥0.30 and load strongest on the respective construct. The items “would like to work out or exercise” and “would like to be by yourself” were not included in the factor calculations of (B) because these items did not significantly contribute to any of the factors (factor loading <0.30). The item “Would you like to be with a group of friends?” was removed because of insufficient sampling adequacy (Kaiser-Meyer-Olkin (KMO) < 0.60). Labeling of the factors was based on the scales/items that loaded most strongly on each respective construct.

Fig. 2. Relationships between chronotype, psychiatric traits and constructs, and diurnal symptom patterns.

A Coefficient plot of the relationships of chronotype with the psychiatric traits and constructs and B–E diurnal patterns of symptom constructs for chronotype and psychiatric constructs. The coefficient plot shows how chronotype relates to 13 psychiatric traits and four transdiagnostic psychiatric trait constructs, ordered according to coefficient strength (0 = no relationship). A lower score of chronotype (reduced Morningness-Eveningness Questionnaire; rMEQ) on the x-axis indicates a stronger evening-type. Error bars represent 95% confidence intervals. Filled dots indicate statistically significant relationships, open dots indicate statistically non-significant relationships. Diurnal patterns of B negative emotional symptoms, C fatigue-related symptoms, D low drive and motivation, and E ADHD-type symptoms for morning- and evening-types. The Y-axes represent mean scores of the item ratings with each item having a minimum of 0 “not at all” and a maximum of 8 “Very much/All the time”. Note the different start and end points of the axes. The axis for D drive and motivation was reversed for ease of interpretation, with worse symptom levels (lower drive and motivation) being illustrated as higher on the scales. Error bands represent 95% pointwise confidence intervals. Gray bars show the timepoints where morning- and evening-types differ significantly. See Supplementary Fig. S4 for illustrations that also include intermediate-types. ***p < 0.001; **p < 0.01; *p < 0.05.

Fig. 3. Diurnal patterns of psychiatric symptoms.

Diurnal patterns of A–D negative emotional symptoms, E–H fatigue-related symptoms, I–L low drive and motivation, and M–P ADHD-type symptoms, separated by Depression-Anxiety construct (1^st column), Downregulatory Problems construct (2^nd column), Compulsive Behavior and Intrusive Thoughts construct (3^rd column), and Social Dysfunction construct (4^th column). All plots are stratified by psychiatric construct (Low and High) and chronotype (Morning-type (solid) vs Evening-type (dashed)). For visualization purposes only, the constructs are divided into low and high construct: Low Construct = mean(construct) – 1.5*standard deviation(construct). High Construct = mean(construct) + 1.5*standard deviation(construct). Therefore, the difference between the low and high construct is 3 standard deviations. Brown (low construct) and blue (high construct) bars show the timepoints where morning- and evening-types significantly differ. The Y-axes represent mean scores of the item ratings with each item having a minimum of 0 “not at all” and a maximum of 8 “Very much/All the time”. Error bands represent 95% pointwise confidence intervals. Note the different start and end points of the scales. The presented values are predicted GAMM values.