Vertex-wise meta-analysis of 12,000 MRI scans identifies cortical thinning in MDD concentrated in parietal, orbitofrontal and cingulate regions

The study aggregated T1-weighted scans processed with standardized FreeSurfer pipelines and applied random-effects meta-analysis at each vertex. Cohen’s d values for thickness reductions ranged from –0.10 to –0.18 in the identified clusters. Adolescent subgroups showed no detectable differences, while adult patients in current depressive episodes and those receiving SSRIs or SNRIs exhibited modestly larger effect sizes. Surface-area maps yielded null results after multiple-comparison correction. These patterns extend earlier ENIGMA reports by resolving sub-regional specificity and clarifying that medication status and illness phase, rather than lifetime diagnosis alone, modulate the magnitude of thinning.

Prior smaller studies often conflated thickness and area or lacked power to stratify by age and treatment. The current map aligns with task-based fMRI meta-analyses implicating the same territories in cognitive control and reward processing, suggesting structural change may index functional inefficiency rather than serve as a diagnostic marker. Because the design remains cross-sectional, reverse causation from chronic stress or antidepressant exposure cannot be excluded; the absence of surface-area effects further implies that any neurodevelopmental contribution is selective to radial cortical architecture.

Longitudinal imaging embedded in ongoing pragmatic trials is required to test whether baseline thinning predicts non-response to first-line pharmacotherapy or recurrence after remission. Integration with peripheral inflammatory markers and polygenic risk scores could clarify whether the observed structural signature reflects state-dependent atrophy or a stable trait conferring vulnerability.