Scholarlink

<p>The structural neuroanatomy underlying Major Depressive Disorder (MDD) remains elusive and even large-scale neuroimaging studies and meta-analyses produced inconsistent outcomes with marginal effect sizes. Furthermore, the replicability of brain-psychopathology associations is under scrutiny, attributable to inadequately powered sample sizes and inflated effect sizes. The present study aimed to examine the effect sizes, replicability, and generalizability of brain structural alterations in MDD, harnessing the wealth of three large-scale clinical cohorts. We used three independent cohorts totaling <i>n</i> = 4021 deeply characterized participants (aged 15–65 years), encompassing MDD patients (<i>n</i> = 1764), and healthy controls (HC; <i>n</i> = 2257) with individual MRI data. The diagnosis was confirmed by structured clinical interviews. Brain-wide case-control differences in voxel-based morphometry were tested. We conducted pooled analyses to maximize statistical power, and further investigated cross-cohort replicability and generalizability by inspecting 1) correlations between t-maps from each cohort, 2) converging significance across the three cohorts, and 3) by using a cross-validation framework, iterating through cohorts as independent test sets. Pooling all individuals together yielded reduced gray matter volumes in patients with MDD in widespread bilateral clusters, covering the insula, thalamus, orbitofrontal cortex, and a parahippocampal-fusiform-lingual complex. No evidence was found for differences within the hippocampus. The largest effect size was found in the bilateral anterior insula (partial R<sup>2</sup> = 0.01). Analyzing cohorts separately yielded 1) t-map correlations between cohorts (r = 0.212 to r = 0.281), 2) replicability indicated by overlapping significance in multiple areas exceeding thresholds in each single cohort, and 3) generalizability of case-control differences in clusters that were deemed non-significant in single cohorts, implicating further frontal, temporal and cerebellar regions. Results indicate that gray matter correlates of MDD are subtle but nevertheless replicable and generalizable. These alterations are localized in an array of regions involved in emotion regulation and sensory processing. Our investigation underscores the need to investigate the replicability and generalizability of mental health neuroimaging findings and provides a tangible framework for this endeavor.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The neuroanatomy of depression: weak but replicable effects in 4021 individuals from three clinical cohorts

  • 期刊论文

  • 发表日期：

    2026年7月10日

查看全文
  • Udo Dannlowski,
  • Nils R. Winter,
  • Susanne Meinert,
  • Dominik Grotegerd,
  • Anna Kraus,
  • Kira Flinkenflügel,
  • Elisabeth Leehr,
  • Joscha Böhnlein,
  • Tiana Borgers,
  • Lukas Fisch,
  • Michael Bauer,
  • Andrea Pfennig,
  • Maike Richter,
  • Nils Opel,
  • Jonathan Repple,
  • Marius Gruber,
  • Heike Minnerup,
  • Marco Hermesdorf,
  • Robert Nitsch,
  • Joachim Groß,
  • Andreas Jansen,
  • Nina Alexander,
  • Frederike Stein,
  • Igor Nenadić,
  • Klaus Berger,
  • Tilo Kircher,
  • Tim Hahn,
  • Janik Goltermann

摘要

The structural neuroanatomy underlying Major Depressive Disorder (MDD) remains elusive and even large-scale neuroimaging studies and meta-analyses produced inconsistent outcomes with marginal effect sizes. Furthermore, the replicability of brain-psychopathology associations is under scrutiny, attributable to inadequately powered sample sizes and inflated effect sizes. The present study aimed to examine the effect sizes, replicability, and generalizability of brain structural alterations in MDD, harnessing the wealth of three large-scale clinical cohorts. We used three independent cohorts totaling n = 4021 deeply characterized participants (aged 15–65 years), encompassing MDD patients (n = 1764), and healthy controls (HC; n = 2257) with individual MRI data. The diagnosis was confirmed by structured clinical interviews. Brain-wide case-control differences in voxel-based morphometry were tested. We conducted pooled analyses to maximize statistical power, and further investigated cross-cohort replicability and generalizability by inspecting 1) correlations between t-maps from each cohort, 2) converging significance across the three cohorts, and 3) by using a cross-validation framework, iterating through cohorts as independent test sets. Pooling all individuals together yielded reduced gray matter volumes in patients with MDD in widespread bilateral clusters, covering the insula, thalamus, orbitofrontal cortex, and a parahippocampal-fusiform-lingual complex. No evidence was found for differences within the hippocampus. The largest effect size was found in the bilateral anterior insula (partial R2 = 0.01). Analyzing cohorts separately yielded 1) t-map correlations between cohorts (r = 0.212 to r = 0.281), 2) replicability indicated by overlapping significance in multiple areas exceeding thresholds in each single cohort, and 3) generalizability of case-control differences in clusters that were deemed non-significant in single cohorts, implicating further frontal, temporal and cerebellar regions. Results indicate that gray matter correlates of MDD are subtle but nevertheless replicable and generalizable. These alterations are localized in an array of regions involved in emotion regulation and sensory processing. Our investigation underscores the need to investigate the replicability and generalizability of mental health neuroimaging findings and provides a tangible framework for this endeavor.