Quantitative MRI at 7-Tesla reveals novel frontocortical myeloarchitecture anomalies in major depressive disorder

Major Depressive Disorder (MDD) has been consistently linked with perturbations in cortical morphology, particularly within prefrontal and anterior cingulate cortices. Large meta-analyses (e.g., ENIGMA MDD) report cortical thinning in the orbitofrontal cortex (OFC) and rostral anterior cingulate cortex (rACC). However, conventional macrostructural metrics (thickness, surface area, volume) capture boundary shape rather than microstructural processes within the cortical mantle. Intracortical myelination and iron concentration are key microstructural candidates relevant to cortical morphology, neural circuit integrity, and network function. The study set out to test whether MDD diagnosis and clinical characteristics are associated with intracortical myelin and iron abnormalities in frontocortical regions implicated by prior meta-analyses, using ultra-high field (7T) quantitative MRI. The primary hypothesis predicted reduced intracortical myelin (indexed by higher T1) in OFC and rACC in MDD, with exploratory tests of associations with severity, medication, comorbid anxiety, and childhood trauma, and to examine whether iron content (T2*) is altered.

Prior structural MRI studies and meta-analyses implicate prefrontal and anterior cingulate cortices in MDD, with ENIGMA reporting robust cortical thinning in OFC and rACC. Intracortical myelination, concentrated in deeper cortical layers, supports circuit integrity and functional networks and is sensitive to environmental influences across the lifespan. Preliminary postmortem and in vivo indications suggest intracortical demyelination in prefrontal regions in depression and a possible link between demyelination and reduced cortical volumes. Iron is essential for oligodendrocyte myelin synthesis; however, excess iron can impede remyelination and induce oxidative stress and cell toxicity. Imbalanced iron can affect neurophysiology including neurotransmitter synthesis. Together, these lines of evidence motivate in vivo mapping of myelin and iron to clarify cortical microstructural substrates of MDD-related macrostructural abnormalities.

Design: Cross-sectional case-control study using ultra-high field (7T) quantitative MRI to assess intracortical myelin (T1) and iron concentration (T2*). Participants: Recruited 73 individuals; final analyzed sample reported in sample characteristics included MDD patients (N = 48) and healthy controls (HC; N = 10), matched on age, sex, and education. Inclusion (MDD): primary DSM-5 current MDD (within 6 months) via CIDI, referral to mental health care. Inclusion (HC): no depression history/treatment, no psychopathology, normal/subclinical dimensional scores. Exclusion: psychosis/mania/Tourette’s/OCD; major internal/neurological disorders; traumatic head injury; current substance abuse/dependence requiring treatment; acute suicidal risk; MRI contraindications; deafness; pregnancy; inadequate Dutch; age <20 or >55. Clinical measures: IDS for symptom severity; CTQ for childhood trauma; medication and comorbid anxiety disorders recorded. MRI acquisition: Philips Achieva 7T, 32-channel head coil. Simultaneous T1 and T2* via MP2RAGE multi-echo sequence. Key parameters: TI1 = 670 ms, TI2 = 3675 ms; multi-echo in second inversion with TE1 = 3.9 ms and additional echoes (e.g., 11.5, 19.85 ms); flip angles 14° and 47°; voxel size 0.7 mm isotropic; FOV 205×205×164 mm3; total acquisition ~19.3 min; motion correction via interleaved fat navigators. T1 maps computed via lookup table from sequence model; T2* from least-squares exponential fit across multi-echo images. T1 is considered a proxy for myelin content; T2* a proxy for iron concentration. Regions of interest (ROIs): rACC, medial OFC (mOFC), lateral OFC (lOFC) from Desikan-Killiany atlas, defined in FreeSurfer space. Laminar sampling: Quantitative values sampled across 10 cortical depths per ROI per hemisphere using a profile sampling module based on a laminar model. Hemispheric values averaged after confirming no lateralization differences (p > 0.05). Derived metrics: - Average T1 and T2*: mean across 10 depths. - AUC (area under curve) of depth profile: 3rd-order polynomial fit across depths; AUC reflects depth-dependent distribution. - Offset (intercept) parameter from polynomial fit as an index at the WM/GM transitional zone. Statistical analysis: ANCOVA compared MDD vs HC on ROI metrics, controlling for age and sex, with alpha = 0.05 and effect sizes reported. Within-patient analyses (MDD only) tested associations of IDS severity, antidepressant use, comorbid anxiety disorders, and childhood trauma (CTQ) with T1/T2* metrics in ROIs showing group differences. Control analyses: Primary somatosensory cortex (S1) tested to evaluate specificity; VBM and white matter T1 examined to rule out confounding by gray matter volume or underlying WM changes.

• Group differences (MDD vs HC): In all ROIs, T1 profiles were higher in MDD; effects reached significance only in the lateral OFC (lOFC). lOFC average T1: F(1,54) = 4.174, p = 0.046, Cohen’s D = 0.55; MDD mean = 1903.36 ms (SD = 62.52), HC mean = 1858.89 ms (SD = 98.9). lOFC depth-dependent profile (T1 AUC): F(1,54) = 4.369, p = 0.041, Cohen’s D = 0.60; MDD AUC mean = 1904.75 (SD = 677.24) vs HC mean = 1890.97 (SD = 956.38). lOFC WM/GM transitional zone (T1 offset): F(1,454) = 4.722, p = 0.034, Cohen’s D = 0.66; MDD mean = 1788.83 ms (SD = 59.73) vs HC mean = 1746.40 ms (SD = 61.23). Interpretation: Higher T1 indicates decreased intracortical myelin concentration in lOFC in MDD. - Iron (T2*): No significant MDD–HC differences in average or AUC T2* metrics across ROIs (all p > 0.05). - Specificity controls: No group effects in primary somatosensory cortex (S1) (p > 0.05). VBM and underlying white matter T1 analyses suggested effects were not driven by gray matter volume or WM confounds (p > 0.05). - Clinical associations within MDD (lOFC T1 metrics): Higher IDS severity correlated with higher T1 (indicative of lower myelin): average T1 r = 0.269, p = 0.008 (CI 0.054–0.485); T1 AUC r = 0.271, p = 0.008 (CI 0.056–0.486); T1 offset r = 0.267, p = 0.008 (CI 0.048–0.487). No significant associations with antidepressant use, comorbid anxiety disorders, or childhood trauma (all p > 0.05). Overall: Findings indicate decreased intracortical myelination localized to the lOFC in MDD, scaling with symptom severity, with no evidence for altered iron concentration.

The study addressed the gap between macrostructural findings in MDD and underlying microstructural processes by leveraging 7T quantitative MRI. Results demonstrate that MDD diagnosis and symptom severity are associated with reduced intracortical myelin (higher T1) specifically in the lateral orbitofrontal cortex (lOFC), a hub implicated in negative affect processing and hypersensitivity to punishment/non-reward. The depth-wise and WM/GM transitional zone effects suggest a broad and layer-sensitive demyelination pattern in lOFC. The absence of T2* alterations suggests iron concentration changes are not a prominent feature in these regions in this cohort, pointing instead to myelin-related processes. These findings align with models positing that intracortical demyelination can disrupt cortico-limbic circuitry and contribute to the macrostructural cortical thinning seen in meta-analyses. Potential mechanisms include decreased neural activity-dependent myelination and stress-induced neuroinflammation, both of which can impair oligodendrocyte function and myelin integrity. The lOFC’s central role within negative affect and salience networks dovetails with observed associations between myelin alterations and depressive symptom severity, supporting a pathophysiological link between microstructural disorganization and affective dysfunction in MDD.

Using ultra-high field 7T quantitative MRI, the study provides fine-grained evidence that MDD is associated with decreased intracortical myelin in the lateral OFC, with effect sizes in the moderate range and correlations with symptom severity. No differences were detected in iron concentration. These results suggest intracortical demyelination may underlie previously observed macrostructural abnormalities and contribute to circuit-level dysfunction in depression. Future research should employ larger, longitudinal cohorts to clarify causality, track illness progression and treatment effects, and determine whether intracortical myelination alterations can serve as biomarkers and therapeutic targets in MDD.

• Small and unbalanced sample with relatively few healthy controls, limiting power and generalizability. - Cross-sectional design precludes causal inference and assessment of illness duration effects. - The study did not aim to replicate large-scale ENIGMA macrostructural findings or comprehensively model micro–macro relationships; sample size and complexity of processes likely limited such analyses. - Quantitative MRI provides specific biophysical measures, but broader microstructural features beyond myelin/iron were not assessed; larger UHF studies with comprehensive features are needed. - Data are not publicly available due to ethical constraints, limiting external validation.