Multi-site benchmark classification of major depressive disorder using machine learning on cortical and subcortical measures

Machine learning (ML) techniques have gained popularity in the neuroimaging field due to their potential for classifying neuropsychiatric disorders. However, the diagnostic predictive power of the existing algorithms has been limited by small sample sizes, lack of representativeness, data leakage, a...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2022-10
Main Authors: Belov, Vladimir, Erwin-Grabner, Tracy, Ali Saffet Gonul, Amod, Alyssa R, Ojha, Amar, Aleman, Andre, Dols, Annemiek, Scharntee, Anouk, Uyar-Demir, Aslihan, Harrison, Ben J, Irungu, Benson M, Besteher, Bianca, Klimes-Dougan, Bonnie, Brenda W J H Penninx, Mueller, Bryon A, Zarate, Carlos, Davey, Christopher G, Ching, Christopher R K, Connolly, Colm G, Fu, Cynthia H Y, Stein, Dan J, Dima, Danai, Linden, David E J, Mehler, David M A, Pomarol-Clotet, Edith, Pozzi, Elena, Melloni, Elisa, Benedetti, Francesco, MacMaster, Frank P, Grabe, Hans J, Völzke, Henry, Gotlib, Ian H, Soares, Jair C, Evans, Jennifer W, Sim, Kang, Wittfeld, Katharina, Cullen, Kathryn, Reneman, Liesbeth, Oudega, Mardien L, Wright, Margaret J, Portella, Maria J, Sacchet, Matthew D, Li, Meng, Aghajani, Moji, Mon-Ju Wu, Jaworska, Natalia, Jahanshad, Neda, Nic J A van der Wee, Groenewold, Nynke, Hamilton, Paul J, Saemann, Philipp, Bülow, Robin, Poletti, Sara, Whittle, Sarah, Thomopoulos, Sophia I, van, Steven J A, der Werff, Sheri-Michelle Koopowitz, Lancaster, Thomas, Ho, Tiffany C, Yang, Tony T, Basgoze, Zeynep, Veltman, Dick J, Schmaal, Lianne, Thompson, Paul M, Goya-Maldonado, Roberto
Format: Article
Language:English
Subjects:
Algorithms
Antidepressants
Benchmarks
Brain
Classification
Deep learning
Machine learning
Medical imaging
Mental depression
Support vector machines
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Machine learning (ML) techniques have gained popularity in the neuroimaging field due to their potential for classifying neuropsychiatric disorders. However, the diagnostic predictive power of the existing algorithms has been limited by small sample sizes, lack of representativeness, data leakage, and/or overfitting. Here, we overcome these limitations with the largest multi-site sample size to date (n=5,356) to provide a generalizable ML classification benchmark of major depressive disorder (MDD). Using brain measures from standardized ENIGMA analysis pipelines in FreeSurfer, we were able to classify MDD vs healthy controls (HC) with around 62% balanced accuracy, but when harmonizing the data using ComBat balanced accuracy dropped to approximately 52%. Similar results were observed in stratified groups according to age of onset, antidepressant use, number of episodes and sex. Future studies incorporating higher dimensional brain imaging/phenotype features, and/or using more advanced machine and deep learning methods may achieve more encouraging prospects.
ISSN:2331-8422