TY - JOUR
T1 - Machine Learning for Large-Scale Quality Control of 3D Shape Models in Neuroimaging
AU - for the ENIGMA consortium
AU - Petrov, Dmitry
AU - Gutman, Boris A.
AU - (Julie) Yu, Shih Hua
AU - van Erp, Theo G.M.
AU - Turner, Jessica A.
AU - Schmaal, Lianne
AU - Veltman, Dick
AU - Wang, Lei
AU - Alpert, Kathryn
AU - Isaev, Dmitry
AU - Zavaliangos-Petropulu, Artemis
AU - Ching, Christopher R.K.
AU - Calhoun, Vince
AU - Glahn, David
AU - Satterthwaite, Ted
AU - Andreasen, Ole Andreas
AU - Borgwardt, Stefan
AU - Howells, Fleur
AU - Groenewold, Nynke
AU - Voineskos, Aristotle
AU - Radua, Joaquim
AU - Potkin, Steven G.
AU - Crespo-Facorro, Benedicto
AU - Tordesillas-Gutiérrez, Diana
AU - Shen, Li
AU - Lebedeva, Irina
AU - Spalletta, Gianfranco
AU - Donohoe, Gary
AU - Kochunov, Peter
AU - Rosa, Pedro G.P.
AU - James, Anthony
AU - Dannlowski, Udo
AU - Baune, Bernhard T.
AU - Aleman, André
AU - Gotlib, Ian H.
AU - Walter, Henrik
AU - Walter, Martin
AU - Soares, Jair C.
AU - Gur, Ruben C.
AU - Doan, N. Trung
AU - Agartz, Ingrid
AU - Westlye, Lars T.
AU - Harrisberger, Fabienne
AU - Riecher-Rössler, Anita
AU - Uhlmann, Anne
AU - Stein, Dan J.
AU - Dickie, Erin W.
AU - Pomarol-Clotet, Edith
AU - Fuentes-Claramonte, Paola
AU - Canales-Rodríguez, Erick Jorge
N1 - Publisher Copyright:
The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license.
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.
PY - 2017/7/21
Y1 - 2017/7/21
N2 - As very large studies of complex neuroimaging phenotypes become more common, human quality assessment of MRI-derived data remains one of the last major bottlenecks. Few attempts have so far been made to address this issue with machine learning. In this work, we optimize predictive models of quality for meshes representing deep brain structure shapes. We use standard vertex-wise and global shape features computed homologously across 19 cohorts and over 7500 human-rated subjects, training kernelized Support Vector Machine and Gradient Boosted Decision Trees classifiers to detect meshes of failing quality. Our models generalize across datasets and diseases, reducing human workload by 30-70%, or equivalently hundreds of human rater hours for datasets of comparable size, with recall rates approaching inter-rater reliability.
AB - As very large studies of complex neuroimaging phenotypes become more common, human quality assessment of MRI-derived data remains one of the last major bottlenecks. Few attempts have so far been made to address this issue with machine learning. In this work, we optimize predictive models of quality for meshes representing deep brain structure shapes. We use standard vertex-wise and global shape features computed homologously across 19 cohorts and over 7500 human-rated subjects, training kernelized Support Vector Machine and Gradient Boosted Decision Trees classifiers to detect meshes of failing quality. Our models generalize across datasets and diseases, reducing human workload by 30-70%, or equivalently hundreds of human rater hours for datasets of comparable size, with recall rates approaching inter-rater reliability.
KW - machine learning
KW - quality control
KW - shape analysis
UR - http://www.scopus.com/inward/record.url?scp=85095638058&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85095638058&partnerID=8YFLogxK
U2 - 10.1101/166496
DO - 10.1101/166496
M3 - Article
AN - SCOPUS:85095638058
SN - 0309-1708
JO - Unknown Journal
JF - Unknown Journal
ER -