EPISeg: Automated segmentation of the spinal cord on echo planar images using open-access multi-center data

Rohan Banerjee; Merve Kaptan; Alexandra Tinnermann; Ali Khatibi Tabatabaei; Alice Dabbagh; Christian Büchel; Christian W. Kündig; Christine S.W. Law; Dario Pfyffer; David J. Lythgoe; Dimitra Tsivaka; Dimitri Van De Ville; Falk Eippert; Fauziyya Muhammad; Gary H. Glover; Gergely David; Grace Haynes; Jan Haaker; Jonathan C. W. Brooks; Jürgen Finsterbusch; Katherine T. Martucci; Kimberly J. Hemmerling; Mahdi Mobarak-Abadi; Mark A. Hoggarth; Matthew A. Howard; Molly G. Bright; Nawal Kinany; Olivia S. Kowalczyk; Patrick Freund; Robert L. Barry; Sean Mackey; Shahabeddin Vahdat; Simon Schading; Stephen B McMahon; Todd Parish; Véronique Marchand-Pauvert; Yufen Chen; Zachary A. Smith; II Kenneth A. Weber; Benjamin De Leener; Julien Cohen-Adad

doi:10.1162/IMAG.a.98

EPISeg: Automated segmentation of the spinal cord on echo planar images using open-access multi-center data

Rohan Banerjee
, Merve Kaptan
, Alexandra Tinnermann
, Ali Khatibi Tabatabaei
, Alice Dabbagh
, Christian Büchel
, Christian W. Kündig
, Christine S.W. Law
, Dario Pfyffer
, David J. Lythgoe
, Dimitra Tsivaka
, Dimitri Van De Ville
, Falk Eippert
, Fauziyya Muhammad
, Gary H. Glover
, Gergely David
, Grace Haynes
, Jan Haaker
, Jonathan C. W. Brooks
, Jürgen Finsterbusch

Katherine T. Martucci, Kimberly J. Hemmerling, Mahdi Mobarak-Abadi, Mark A. Hoggarth, Matthew A. Howard, Molly G. Bright, Nawal Kinany, Olivia S. Kowalczyk, Patrick Freund, Robert L. Barry, Sean Mackey, Shahabeddin Vahdat, Simon Schading, Stephen B McMahon, Todd Parish, Véronique Marchand-Pauvert, Yufen Chen, Zachary A. Smith, II Kenneth A. Weber, Benjamin De Leener, Julien Cohen-Adad

École Polytechnique de Montréal c
Stanford University
University Medical Centre Hamburg-Eppendorf
University of Zurich
King's College London
King’s College London
University of Geneva
University of Oklahoma
University of East Anglia
University Hospitals Bristol NHS Foundation Trust
University of Bristol
Duke University
Northwestern University
University College London
Harvard University
McGill University
Sorbonne Université

Research output: Contribution to journal › Article › peer-review

3 Citations (SciVal)

51 Downloads (Pure)

Abstract

Functional magnetic resonance imaging (fMRI) of the spinal cord is relevant for studying sensation, movement, and autonomic function. Preprocessing of spinal cord fMRI data involves segmentation of the spinal cord on gradient-echo echo planar imaging (EPI) images. Current automated segmentation methods do not work well on these data, due to the low spatial resolution, susceptibility artifacts causing distortions and signal drop-out, ghosting, and motion-related artifacts. Consequently, this segmentation task demands a considerable amount of manual effort which takes time and is prone to user bias. In this work, we (i) gathered a multi-center dataset of spinal cord gradient-echo EPI with ground-truth segmentations and shared it on OpenNeuro https://openneuro.org/datasets/ds005143/versions/1.3.1 and (ii) developed a deep learning-based model, EPISeg, for the automatic segmentation of the spinal cord on gradient-echo EPI data. We observe a significant improvement in terms of segmentation quality compared with other available spinal cord segmentation models. Our model is resilient to different acquisition protocols as well as commonly observed artifacts in fMRI data. The training code is available at https://github.com/sct-pipeline/fmri-segmentation/, and the model has been integrated into the Spinal Cord Toolbox as a command-line tool.

Original language	English
Article number	IMAG.a.98
Journal	Imaging Neuroscience
Volume	3
Early online date	14 Jul 2025
DOIs	https://doi.org/10.1162/IMAG.a.98
Publication status	Published - 9 Sept 2025

Bibliographical note

Publisher Copyright:
© 2025 The Authors. Published under a Creative Commons Attribution 4.0 International (CC BY 4.0) license.

Funding

This study was funded by the Canada Research Chair in Quantitative Magnetic Resonance Imaging [CRC-2020-00179], the Canadian Institute of Health Research [PJT-190258], the Canada Foundation for Innovation [32454, 34824], the Fonds de Recherche du Québec—Santé [322736, 324636], the Natural Sciences and Engineering Research Council of Canada [RGPIN-2019-07244], the Canada First Research Excellence Fund (IVADO and TransMedTech), the Quebec BioImaging Network [5886, 35450], INSPIRED (Spinal Research, UK; Wings for Life, Austria; Craig H. Neilsen Foundation, USA), Mila—Tech Transfer Funding Program, the National Institute of Neurological Disorders and Stroke (K23NS104211, R01NS109450, K24NS126781, R01NS133305, and R01NS128478), the National Institute on Drug Abuse (K99/R00DA040154), the National Center for Complementary and Integrative Health (F32AT007800), and the National Institute of Biomedical Imaging and Bioengineering (R01EB027779 and R21EB031211), the Swiss National Science Foundation (SNSF; 32003B_204934). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Funders	Funder number
INSPIRED
National Institutes of Health
Canada First Research Excellence Fund
Canadian Institute of Health Research	PJT-190258
Canada Research Chairs	CRC-2020-00179
Institut de Valorisation des Données	35450
Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung	32003B_204934
National Institute of Biomedical Imaging and Bioengineering	R01EB027779, R21EB031211
Santé	324636, 322736
National Center for Complementary and Integrative Health	F32AT007800
National Institute of Neurological Disorders and Stroke	R01NS109450, K24NS126781, R01NS133305, K23NS104211, R01NS128478
Canada Foundation for Innovation	32454, 34824
National Institute on Drug Abuse	K99/R00DA040154
Natural Sciences and Engineering Research Council of Canada	RGPIN-2019-07244

Keywords

echo planar imaging
functional magnetic resonance imaging
segmentation
spinal cord

ASJC Scopus subject areas

Neuroscience (miscellaneous)
Medicine (miscellaneous)
Radiology Nuclear Medicine and imaging
Clinical Neurology

Access to Document

10.1162/IMAG.a.98Licence: CC BY

EPISeg (accepted author manuscript)Accepted author manuscript, 1.68 MBLicence: CC BY

Cite this

Banerjee, R., Kaptan, M., Tinnermann, A., Khatibi Tabatabaei, A., Dabbagh, A., Büchel, C., Kündig, C. W., Law, C. S. W., Pfyffer, D., Lythgoe, D. J., Tsivaka, D., Ville, D. V. D., Eippert, F., Muhammad, F., Glover, G. H., David, G., Haynes, G., Haaker, J., Brooks, J. C. W., ... Cohen-Adad, J. (2025). EPISeg: Automated segmentation of the spinal cord on echo planar images using open-access multi-center data. Imaging Neuroscience, 3, Article IMAG.a.98. https://doi.org/10.1162/IMAG.a.98

Banerjee, R, Kaptan, M, Tinnermann, A, Khatibi Tabatabaei, A, Dabbagh, A, Büchel, C, Kündig, CW, Law, CSW, Pfyffer, D, Lythgoe, DJ, Tsivaka, D, Ville, DVD, Eippert, F, Muhammad, F, Glover, GH, David, G, Haynes, G, Haaker, J, Brooks, JCW, Finsterbusch, J, Martucci, KT, Hemmerling, KJ, Mobarak-Abadi, M, Hoggarth, MA, Howard, MA, Bright, MG, Kinany, N, Kowalczyk, OS, Freund, P, Barry, RL, Mackey, S, Vahdat, S, Schading, S, McMahon, SB, Parish, T, Marchand-Pauvert, V, Chen, Y, Smith, ZA, Kenneth A. Weber, II, Leener, BD & Cohen-Adad, J 2025, 'EPISeg: Automated segmentation of the spinal cord on echo planar images using open-access multi-center data', Imaging Neuroscience, vol. 3, IMAG.a.98. https://doi.org/10.1162/IMAG.a.98

@article{cdcbbedec71840d1967aa78da0bfcf66,

title = "EPISeg: Automated segmentation of the spinal cord on echo planar images using open-access multi-center data",

abstract = "Functional magnetic resonance imaging (fMRI) of the spinal cord is relevant for studying sensation, movement, and autonomic function. Preprocessing of spinal cord fMRI data involves segmentation of the spinal cord on gradient-echo echo planar imaging (EPI) images. Current automated segmentation methods do not work well on these data, due to the low spatial resolution, susceptibility artifacts causing distortions and signal drop-out, ghosting, and motion-related artifacts. Consequently, this segmentation task demands a considerable amount of manual effort which takes time and is prone to user bias. In this work, we (i) gathered a multi-center dataset of spinal cord gradient-echo EPI with ground-truth segmentations and shared it on OpenNeuro https://openneuro.org/datasets/ds005143/versions/1.3.1 and (ii) developed a deep learning-based model, EPISeg, for the automatic segmentation of the spinal cord on gradient-echo EPI data. We observe a significant improvement in terms of segmentation quality compared with other available spinal cord segmentation models. Our model is resilient to different acquisition protocols as well as commonly observed artifacts in fMRI data. The training code is available at https://github.com/sct-pipeline/fmri-segmentation/, and the model has been integrated into the Spinal Cord Toolbox as a command-line tool.",

keywords = "echo planar imaging, functional magnetic resonance imaging, segmentation, spinal cord",

author = "Rohan Banerjee and Merve Kaptan and Alexandra Tinnermann and \{Khatibi Tabatabaei\}, Ali and Alice Dabbagh and Christian B{\"u}chel and K{\"u}ndig, \{Christian W.\} and Law, \{Christine S.W.\} and Dario Pfyffer and Lythgoe, \{David J.\} and Dimitra Tsivaka and Ville, \{Dimitri Van De\} and Falk Eippert and Fauziyya Muhammad and Glover, \{Gary H.\} and Gergely David and Grace Haynes and Jan Haaker and Brooks, \{Jonathan C. W.\} and J{\"u}rgen Finsterbusch and Martucci, \{Katherine T.\} and Hemmerling, \{Kimberly J.\} and Mahdi Mobarak-Abadi and Hoggarth, \{Mark A.\} and Howard, \{Matthew A.\} and Bright, \{Molly G.\} and Nawal Kinany and Kowalczyk, \{Olivia S.\} and Patrick Freund and Barry, \{Robert L.\} and Sean Mackey and Shahabeddin Vahdat and Simon Schading and McMahon, \{Stephen B\} and Todd Parish and V{\'e}ronique Marchand-Pauvert and Yufen Chen and Smith, \{Zachary A.\} and \{Kenneth A. Weber\}, II and Leener, \{Benjamin De\} and Julien Cohen-Adad",

note = "Publisher Copyright: {\textcopyright} 2025 The Authors. Published under a Creative Commons Attribution 4.0 International (CC BY 4.0) license.",

year = "2025",

month = sep,

day = "9",

doi = "10.1162/IMAG.a.98",

language = "English",

volume = "3",

journal = "Imaging Neuroscience",

issn = "2837-6056",

publisher = "MIT Press",

}

TY - JOUR

T1 - EPISeg

T2 - Automated segmentation of the spinal cord on echo planar images using open-access multi-center data

AU - Banerjee, Rohan

AU - Kaptan, Merve

AU - Tinnermann, Alexandra

AU - Khatibi Tabatabaei, Ali

AU - Dabbagh, Alice

AU - Büchel, Christian

AU - Kündig, Christian W.

AU - Law, Christine S.W.

AU - Pfyffer, Dario

AU - Lythgoe, David J.

AU - Tsivaka, Dimitra

AU - Ville, Dimitri Van De

AU - Eippert, Falk

AU - Muhammad, Fauziyya

AU - Glover, Gary H.

AU - David, Gergely

AU - Haynes, Grace

AU - Haaker, Jan

AU - Brooks, Jonathan C. W.

AU - Finsterbusch, Jürgen

AU - Martucci, Katherine T.

AU - Hemmerling, Kimberly J.

AU - Mobarak-Abadi, Mahdi

AU - Hoggarth, Mark A.

AU - Howard, Matthew A.

AU - Bright, Molly G.

AU - Kinany, Nawal

AU - Kowalczyk, Olivia S.

AU - Freund, Patrick

AU - Barry, Robert L.

AU - Mackey, Sean

AU - Vahdat, Shahabeddin

AU - Schading, Simon

AU - McMahon, Stephen B

AU - Parish, Todd

AU - Marchand-Pauvert, Véronique

AU - Chen, Yufen

AU - Smith, Zachary A.

AU - Kenneth A. Weber, II

AU - Leener, Benjamin De

AU - Cohen-Adad, Julien

PY - 2025/9/9

Y1 - 2025/9/9

N2 - Functional magnetic resonance imaging (fMRI) of the spinal cord is relevant for studying sensation, movement, and autonomic function. Preprocessing of spinal cord fMRI data involves segmentation of the spinal cord on gradient-echo echo planar imaging (EPI) images. Current automated segmentation methods do not work well on these data, due to the low spatial resolution, susceptibility artifacts causing distortions and signal drop-out, ghosting, and motion-related artifacts. Consequently, this segmentation task demands a considerable amount of manual effort which takes time and is prone to user bias. In this work, we (i) gathered a multi-center dataset of spinal cord gradient-echo EPI with ground-truth segmentations and shared it on OpenNeuro https://openneuro.org/datasets/ds005143/versions/1.3.1 and (ii) developed a deep learning-based model, EPISeg, for the automatic segmentation of the spinal cord on gradient-echo EPI data. We observe a significant improvement in terms of segmentation quality compared with other available spinal cord segmentation models. Our model is resilient to different acquisition protocols as well as commonly observed artifacts in fMRI data. The training code is available at https://github.com/sct-pipeline/fmri-segmentation/, and the model has been integrated into the Spinal Cord Toolbox as a command-line tool.

AB - Functional magnetic resonance imaging (fMRI) of the spinal cord is relevant for studying sensation, movement, and autonomic function. Preprocessing of spinal cord fMRI data involves segmentation of the spinal cord on gradient-echo echo planar imaging (EPI) images. Current automated segmentation methods do not work well on these data, due to the low spatial resolution, susceptibility artifacts causing distortions and signal drop-out, ghosting, and motion-related artifacts. Consequently, this segmentation task demands a considerable amount of manual effort which takes time and is prone to user bias. In this work, we (i) gathered a multi-center dataset of spinal cord gradient-echo EPI with ground-truth segmentations and shared it on OpenNeuro https://openneuro.org/datasets/ds005143/versions/1.3.1 and (ii) developed a deep learning-based model, EPISeg, for the automatic segmentation of the spinal cord on gradient-echo EPI data. We observe a significant improvement in terms of segmentation quality compared with other available spinal cord segmentation models. Our model is resilient to different acquisition protocols as well as commonly observed artifacts in fMRI data. The training code is available at https://github.com/sct-pipeline/fmri-segmentation/, and the model has been integrated into the Spinal Cord Toolbox as a command-line tool.

KW - echo planar imaging

KW - functional magnetic resonance imaging

KW - segmentation

KW - spinal cord

UR - https://www.scopus.com/pages/publications/105017066781

U2 - 10.1162/IMAG.a.98

DO - 10.1162/IMAG.a.98

M3 - Article

SN - 2837-6056

VL - 3

JO - Imaging Neuroscience

JF - Imaging Neuroscience

M1 - IMAG.a.98

ER -

EPISeg: Automated segmentation of the spinal cord on echo planar images using open-access multi-center data

Abstract

Bibliographical note

Funding

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this