Classification of major depressive disorder using vertex-wise brain sulcal depth, curvature, and thickness with a deep and a shallow learning model

Major depressive disorder (MDD) is a complex psychiatric disorder that affects the lives of hundreds of millions of individuals around the globe. Even today, researchers debate if morphological alterations in the brain are linked to MDD, likely due to the heterogeneity of this disorder. The application of deep learning tools to neuroimaging data, capable of capturing complex non-linear patterns, has the potential to provide diagnostic and predictive biomarkers for MDD. However, previous attempts to demarcate MDD patients and healthy controls (HC) based on segmented cortical features via linear machine learning approaches have reported low accuracies. In this study, we used globally representative data from the ENIGMA-MDD working group containing 7,012 participants from 30 sites (N=2,772 MDD and N=4,240 HC), which allows a comprehensive analysis with generalizable results. Based on the hypothesis that integration of vertex-wise cortical features can improve classification performance, we evaluated the classification of a DenseNet and a Support Vector Machine (SVM), with the expectation that the former would outperform the latter. As we analyzed a multi-site sample, we additionally applied the ComBat harmonization tool to remove potential nuisance effects of site. We found that both classifiers exhibited close to chance performance (balanced accuracy DenseNet: 51%; SVM: 53%), when estimated on unseen sites. Slightly higher classification performance (balanced accuracy DenseNet: 58%; SVM: 55%) was found when the cross-validation folds contained subjects from all sites, indicating site effect. In conclusion, the integration of vertex-wise morphometric features and the use of the non-linear classifier did not lead to the differentiability between MDD and HC. Our results support the notion that MDD classification on this combination of features and classifiers is unfeasible. Future studies are needed to determine whether more sophisticated integration of information from other MRI modalities such as fMRI and DWI will lead to a higher performance in this diagnostic task.

Authors

Institutions

Publication Details

Journal
Molecular Psychiatry
Published
2025-10-03
DOI
https://doi.org/10.1038/s41380-025-03273-w
Citations
2
Primary Topic
Functional Brain Connectivity Studies
Type
article
Field-Weighted Citation Impact
2.86

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Classification of major depressive disorder using vertex-wise brain sulcal depth, curvature, and thickness with a deep and a shallow learning model

Christopher R. K. Ching, Edith Pomarol‐Clotet, Aleks Stolicyn, Roberto Goya‐Maldonado et al.
2 citations
Molecular Psychiatry
Functional Brain Connectivity Studies
2.86
article

Classification of major depressive disorder using vertex-wise brain sulcal depth, curvature, and thickness with a deep and a shallow learning model

Christopher R. K. Ching, Edith Pomarol‐Clotet, Aleks Stolicyn, Roberto Goya‐Maldonado, Martin Walter, Kathryn R. Cullen, David M. A. Mehler, Laura K. M. Han, Lianne Schmaal, L. Fisch, Baptiste Couvy‐Duchesne, Anouk Schrantee, Henry Völzke, Aleix Solanes, Raymond Salvador, Ben J. Harrison, Amar Ojha, Ling‐Li Zeng, Mardien L. Oudega, Sophia I. Thomopoulos, Neda Jahanshad, Alec J. Jamieson, Colm G. Connolly, Marı́a J. Portella, Dominik Grotegerd, Ali Saffet Gönül, Robin Bülow, Susanne L. Meinert, Annemiek Dols, Dick J. Veltman, Benson Mwangi, Romain Colle, Joaquim Raduà, Alyssa R. Amod, Ramona Leenings, Hans Jörgen Grabe, Elena Rodríguez‐Cano, Nils R. Winter, Yolanda Vives‐Gilabert, Francesco Benedetti, Elisa Melloni, Sheri‐Michelle Koopowitz, David E.J. Linden, Andriana Karuk, Jan Ernsting, Tiffany Cheing Ho, Yara J. Toenders, Margaret J. Wright, Christopher G. Davey, Brenda W.J.H. Penninx, Tim Hahn, Ian H. Gotlib, Igor Nenadić, Zeynep Başgöze, Eduard Vieta, Frederike Stein, Yasumasa Okamoto, Frank P. MacMaster, T. Lancaster, Udo Dannlowski, Tilo T. J. Kircher, Nynke A. Groenewold, Sarah Whittle, Bryon A. Mueller, Aslihan Uyar-Demir, Kang Sim, Tracy Erwin-Grabner, Dan Joseph Stein, Katharina Wittfeld, Emmanuelle Corruble, Katharina Brosch, Sara Poletti, Bianca Besteher, Bonnie Klimes‐Dougan, Jair C. Soares, J. Paul Hamilton, André Alemán, Carlos A. Zarate, Mon-Ju Wu, Paola Fuentes‐Claramonte, Heather C. Whalley, Meng Li, Paul M. Thompson, Tony T. Yang, Jennifer W. Evans, Matthew D. Sacchet
article en
2 citations

Abstract

Major depressive disorder (MDD) is a complex psychiatric disorder that affects the lives of hundreds of millions of individuals around the globe. Even today, researchers debate if morphological alterations in the brain are linked to MDD, likely due to the heterogeneity of this disorder. The application of deep learning tools to neuroimaging data, capable of capturing complex non-linear patterns, has the potential to provide diagnostic and predictive biomarkers for MDD. However, previous attempts to demarcate MDD patients and healthy controls (HC) based on segmented cortical features via linear machine learning approaches have reported low accuracies. In this study, we used globally representative data from the ENIGMA-MDD working group containing 7,012 participants from 30 sites (N=2,772 MDD and N=4,240 HC), which allows a comprehensive analysis with generalizable results. Based on the hypothesis that integration of vertex-wise cortical features can improve classification performance, we evaluated the classification of a DenseNet and a Support Vector Machine (SVM), with the expectation that the former would outperform the latter. As we analyzed a multi-site sample, we additionally applied the ComBat harmonization tool to remove potential nuisance effects of site. We found that both classifiers exhibited close to chance performance (balanced accuracy DenseNet: 51%; SVM: 53%), when estimated on unseen sites. Slightly higher classification performance (balanced accuracy DenseNet: 58%; SVM: 55%) was found when the cross-validation folds contained subjects from all sites, indicating site effect. In conclusion, the integration of vertex-wise morphometric features and the use of the non-linear classifier did not lead to the differentiability between MDD and HC. Our results support the notion that MDD classification on this combination of features and classifiers is unfeasible. Future studies are needed to determine whether more sophisticated integration of information from other MRI modalities such as fMRI and DWI will lead to a higher performance in this diagnostic task.

Molecular PsychiatryVol. 31(3)
Linköping University (SE), Florida State University (US), Hiroshima University (JP), University of Southern California (US), Leiden University (NL), Centre National de la Recherche Scientifique (FR), National Institutes of Health (US), University of Minnesota (US), University Medical Center Groningen (NL), Harvard University (US), Inserm (FR), University of Cape Town (ZA), Universitat de València (ES), Philipps University of Marburg (DE), University of California, Los Angeles (US), The University of Queensland (AU), The University of Melbourne (AU), National University of Singapore (SG), University of Calgary (CA), University of Groningen (NL), University of Pittsburgh (US), National University of Defense Technology (CN), Nanyang Technological University (SG), University of California, San Francisco (US), University of Münster (DE), Université Paris-Saclay (FR), Universitätsmedizin Greifswald (DE), German Center for Neurodegenerative Diseases (DE), Hospital de Sant Pau (ES), University Medical Center Utrecht (NL), Maastricht University (NL), Sorbonne Université (FR), Ege University (TR), Center for the Neural Basis of Cognition (US), Massachusetts General Hospital (US), Centro de Investigación Biomédica en Red de Salud Mental (ES), Orygen Youth Health (AU), Assistance Publique – Hôpitaux de Paris (FR), Centre de recherche en Epidémiologie et Santé des Populations (FR), MRC Centre for Regenerative Medicine (GB), Amsterdam Neuroscience (NL), Orygen (AU), Institut de Myologie (FR), Hospital Mare de Déu de la Mercè (ES), IRCCS Ospedale San Raffaele (IT), Jena University Hospital (DE), University of Minnesota Medical Center (US), GGZ inGeest (NL), Institute of Mental Health (SG), Amsterdam University Medical Centers (NL), Institut du Cerveau (FR), Melbourne Health (AU), National Institute of Mental Health (US), The University of Texas Health Science Center (US), Consorci Institut D'Investigacions Biomediques August Pi I Sunyer (ES), Bicêtre Hospital (FR), Phillips University (US), Universitat de Barcelona (ES), University of Göttingen (DE), Cardiff University (GB), RWTH Aachen University (DE), Erasmus University Rotterdam (NL), Leibniz Institute for Neurobiology (DE), Stanford University (US), University of Edinburgh (GB)
Bill and Melinda Gates Foundation, American Foundation for Suicide Prevention, University of Minnesota, National Alliance for Research on Schizophrenia and Depression, Children's Hospital Foundation, Universitätsmedizin Göttingen, South African Medical Research Council, Deutsche Forschungsgemeinschaft, National Natural Science Foundation of China, Bundesministerium für Bildung und Forschung, Generalitat de Catalunya, Alberta Children's Hospital Foundation, Fundação Bial, National Healthcare Group, Siemens Healthineers, Universität Greifswald, Fundación Bancaria Caixa d'Estalvis i Pensions de Barcelona, Ministry of Cultural Affairs, Ad Astra Chandaria Foundation, National Institutes of Health, University of Texas MD Anderson Cancer Center, Allergan, University of California, San Francisco, Medical Research Council, National Health and Medical Research Council, Instituto de Salud Carlos III, National Institute of Mental Health, National Institute of Child Health and Human Development, National Center for Advancing Translational Sciences, National Center for Complementary and Integrative Health, Weill Institute for Neurosciences, University of California, San Francisco
Openalex Percentile: Top 12%
Functional Brain Connectivity Studies
2.86
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.