Depression symptom network reorganization under internet-based emotion regulation intervention: Evidence from a large real-world cohort.

Depression can be conceptualized as a network of interacting symptoms. Although internet-based interventions are widely used, their associations with changes in symptom network structure remain unclear. Here, in a real-world observational cohort of 26,643 adults completing an internet-based emotion regulation intervention, we examined pre- to postintervention differences in depressive symptom network structure, global connectedness, and symptom centrality, stratified by reduction status. For the overall sample, analyses indicated that the two core depressive symptoms (i.e., depressed mood and anhedonia) showed reduced centrality after the intervention, whereas several noncore symptoms either increased or remained unchanged. The "symptom reduction" and "symptom nonreduction" subgroups showed largely comparable network characteristics at baseline. After the intervention, the symptom reduction subgroup displayed lower global network strength, a reorganized network topology, and reduced centrality of the core symptoms, whereas the symptom nonreduction subgroup did not show salient network alteration. Further analyses using residualized symptom change scores showed a similar pattern: Among individuals with comparable baseline depressive symptom severity, those with better-than-expected outcomes showed lower postintervention network connectivity than those with poorer-than-expected outcomes. Clinically anchored improvement analyses further indicated that reduced centrality of depressed mood was observed across alternative definitions of improvement. Although causal inference is precluded by the observational design, this study provides large-scale descriptive real-world evidence on how depressive symptom networks change alongside symptom reduction during routine use of a digital intervention. (PsycInfo Database Record (c) 2026 APA, all rights reserved).