JMIR Publications

ABSTRACT

Background:

While bidirectional associations among sleep duration, cognitive function, and depression are established, their symptom-level temporal interactions remain poorly characterized.

Objective:

We aim to employ a novel temporal network analysis to clarify these dynamics and guide targeted interventions.

Methods:

We conducted a longitudinal temporal network analysis on 3136 Chinese adults aged ≥45 years from the China Health and Retirement Longitudinal Study (CHARLS) across five waves (2011, 2013, 2015, 2018, and 2020). A graphical vector autoregressive model delineated the interdependencies among sleep duration, cognitive performance (assessed via the Mini-Mental State Examination), and depressive symptoms (evaluated with 10-item Center for Epidemiologic Studies Depression Scale). We also examined sex-specific differences in network structures.

Results:

The symptom “bothered” was found to predict all other CESD-10 symptoms. There were significant causal links between sleep and the CESD-10 node (i.e., Bothered, Drained, Depressed), along with sleep and the MMSE functions (i.e., Numerical ability). Furthermore, sleep duration serve as a bridge between depression symptoms and cognitive functions. There were significant differences in longitudinal network structure between genders. Sex-specific analyses revealed distinct network patterns. Among female participants, the “bothered” node significantly predicted several outcomes over time. In contrast, the temporal network for male participants was sparser, with the “stuck” node in depression domain being predominantly influenced by other nodes.

Conclusions:

Our study revealed that emotional distress, especially the "bothered" symptom, plays a central role in depressive symptoms and cognitive decline. Together with the bridging role of poor sleep duration bridges underscores the potential for interventions targeting both sleep and emotional distress to mitigate depression and delay cognitive deterioration in older adults.