From thoughts to symptoms: The temporal and dimensional dynamics of rumination in depression

Rumination, characterized by persistent, repetitive focus on negative thoughts or emotions, has long been implicated as a central transdiagnostic process in Major Depressive Disorder (MDD). Although many studies have examined the rumination-depression temporal relationship, findings vary depending on design, measurement, and analytic approach. The present study applies a multiscale, data-driven framework to describe how rumination and depressive symptoms unfold across different temporal modes. Using Empirical Mode Decomposition (EMD) and Granger causality in two longitudinal ecological momentary assessment (EMA) datasets, we examined predictive associations at fast, medium, and slow timescales. Our analyses used two high-frequency EMA datasets: (1) 39 adults providing four daily reports of rumination and depressive symptoms across 40 days, and (2) 40 adults providing four daily reports across two weeks that included five rumination dimensions (perseveration, negativity, self-criticism, replaying, brooding) and baseline depressive symptoms. Across both datasets, the most consistent patterns appeared at slower timescales, where rumination and depressive symptoms showed bidirectional associations even after accounting for individual mean levels. Faster and cross-scale influences were detectable but sparser and varied across dimensions. Network controllability analyses suggested that both slow and fast components shape overall system behavior, although these interpretations remain preliminary. Taken together, these findings demonstrate how a multiscale analytic approach can complement existing methods by describing temporal heterogeneity within the association between rumination and depression. While the findings align with cognitive models emphasizing enduring ruminative tendencies, they generate hypotheses about how different temporal patterns may relate to persistent versus momentary components of depressive experience.