Treatment-characteristic brain structural changes in depression: A multi-center longitudinal observational study.

BACKGROUND

The neurobiological correlates of remission/recovery in depression remain unclear, particularly across treatments. We investigated temporal gray matter volume (GMV) changes across four modalities: antidepressant medication (ADM), cognitive behavioral therapy (CBT), electroconvulsive therapy (ECT), and repetitive transcranial magnetic stimulation (rTMS).

METHODS

This multi-center longitudinal observational study included 307 participants in the final analyses: 155 patients with depression (DEP; ADM: 53, CBT: 60, ECT: 22, and rTMS: 20) and 152 healthy controls (HC). DEP underwent magnetic resonance imaging and clinical evaluations at three timepoints (pre-treatment, post-treatment, and 6-month follow-up), and HC at two timepoints. Voxel-based morphometry identified baseline GMV differences between DEP and HC, assessed longitudinal GMV changes per modality, and correlated GMV changes with symptom changes. ComBat harmonization minimized scanner effects.

RESULTS

At baseline, DEP showed smaller GMV in the bilateral hippocampus, amygdala, thalamus, left insula, and middle frontal gyrus. Longitudinal changes occurred only in ECT and ADM. ECT showed GMV increases from pre- to post-treatment in the bilateral hippocampus, amygdala, and left insula, maintained at 6-month follow-up in the right hippocampus and amygdala. In ADM, the left insula initially increased then decreased toward baseline. No significant changes occurred in CBT or rTMS. Later-phase left insular GMV reduction correlated with worsening in ADM (r = -0.342).

CONCLUSIONS

All four modalities achieved clinical improvement, yet GMV changes differed substantially across groups, suggesting treatment-characteristic patterns of structural change within this observational framework. Later-phase insular GMV reduction correlated with worsening in ADM, highlighting its potential as a biomarker for long-term response.