The dynamic interplay between subjective and objective cognitive performance and their impact on the prognosis of major depressive disorder: A cross-lagged panel network analysis

Background Subjective cognitive impairments are highly prevalent in patients with major depressive disorder (MDD) and, consistent with objective cognitive impairments, are associated with greater symptom severity and poorer treatment outcomes. However, their interactions and shared and distinct contributions to depressive symptoms and acute-phase clinical course remain poorly understood. Methods We analyzed 621 patients with MDD at baseline and 373 at Week 8. Treatment comprised usual care in a real-world cohort and antidepressant therapy with perospirone or placebo augmentation in a randomized, double-blind trial. Cross-sectional and cross-lagged networks integrated subjective cognition, objective cognition, and three selected depressive symptom domains. Remission-stratified and sensitivity analyses were conducted. Results Subjective cognition, objective cognition, and depressive symptoms formed distinct interacting modules at both assessments, with global network strength increasing from baseline to Week 8 (ΔS = 0.65, Holm-adjusted p = 0.01). Subjective executive function was most central in non-remitters and objective processing speed in remitters. Accounting for baseline network variables, better baseline subjective executive function was associated with greater Week 8 energy ( β = 0.11), and better baseline subjective memory with greater energy and general interest (both β = 0.06). In contrast, better baseline objective immediate verbal memory was prospectively associated with worse Week 8 levels for sad mood, energy/fatigability, and general interest ( β = −0.12, −0.07, and − 0.05, respectively). Conclusions Subjective and objective cognition showed distinct associations with symptom course, supporting their combined assessment. Better baseline objective memory did not necessarily indicate fewer residual symptoms, but this association does not establish reduced treatment efficacy.