Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images

Aaron Carass, Jennifer L. Cuzzocreo, Shuo Han, Carlos R. Hernandez-Castillo, Paul E. Rasser, Melanie Ganz, Vincent Beliveau, Jose Dolz, Ismail Ben Ayed, Christian Desrosiers, Benjamin Thyreau, José E. Romero, Pierrick Coupé, José V. Manjón, Vladimir S. Fonov, D. Louis Collins, Sarah H. Ying, Chiadikaobi U Onyike, Deana Crocetti, Bennett A. Landman & 3 others Stewart H Mostofsky, Paul M. Thompson, Jerry Ladd Prince

Research output: Contribution to journalArticle

Abstract

The human cerebellum plays an essential role in motor control, is involved in cognitive function (i.e., attention, working memory, and language), and helps to regulate emotional responses. Quantitative in-vivo assessment of the cerebellum is important in the study of several neurological diseases including cerebellar ataxia, autism, and schizophrenia. Different structural subdivisions of the cerebellum have been shown to correlate with differing pathologies. To further understand these pathologies, it is helpful to automatically parcellate the cerebellum at the highest fidelity possible. In this paper, we coordinated with colleagues around the world to evaluate automated cerebellum parcellation algorithms on two clinical cohorts showing that the cerebellum can be parcellated to a high accuracy by newer methods. We characterize these various methods at four hierarchical levels: coarse (i.e., whole cerebellum and gross structures), lobe, subdivisions of the vermis, and the lobules. Due to the number of labels, the hierarchy of labels, the number of algorithms, and the two cohorts, we have restricted our analyses to the Dice measure of overlap. Under these conditions, machine learning based methods provide a collection of strategies that are efficient and deliver parcellations of a high standard across both cohorts, surpassing previous work in the area. In conjunction with the rank-sum computation, we identified an overall winning method.

Original languageEnglish (US)
Pages (from-to)150-172
Number of pages23
JournalNeuroImage
Volume183
DOIs
StatePublished - Dec 1 2018

Fingerprint

Cerebellum
Magnetic Resonance Spectroscopy
Pathology
Cerebellar Ataxia
Autistic Disorder
Short-Term Memory
Cognition
Schizophrenia
Language

Keywords

  • Attention deficit hyperactivity disorder
  • Autism
  • Cerebellar ataxia
  • Magnetic resonance imaging

ASJC Scopus subject areas

  • Neurology
  • Cognitive Neuroscience

Cite this

Carass, A., Cuzzocreo, J. L., Han, S., Hernandez-Castillo, C. R., Rasser, P. E., Ganz, M., ... Prince, J. L. (2018). Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images. NeuroImage, 183, 150-172. https://doi.org/10.1016/j.neuroimage.2018.08.003

Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images. / Carass, Aaron; Cuzzocreo, Jennifer L.; Han, Shuo; Hernandez-Castillo, Carlos R.; Rasser, Paul E.; Ganz, Melanie; Beliveau, Vincent; Dolz, Jose; Ben Ayed, Ismail; Desrosiers, Christian; Thyreau, Benjamin; Romero, José E.; Coupé, Pierrick; Manjón, José V.; Fonov, Vladimir S.; Collins, D. Louis; Ying, Sarah H.; Onyike, Chiadikaobi U; Crocetti, Deana; Landman, Bennett A.; Mostofsky, Stewart H; Thompson, Paul M.; Prince, Jerry Ladd.

In: NeuroImage, Vol. 183, 01.12.2018, p. 150-172.

Research output: Contribution to journalArticle

Carass, A, Cuzzocreo, JL, Han, S, Hernandez-Castillo, CR, Rasser, PE, Ganz, M, Beliveau, V, Dolz, J, Ben Ayed, I, Desrosiers, C, Thyreau, B, Romero, JE, Coupé, P, Manjón, JV, Fonov, VS, Collins, DL, Ying, SH, Onyike, CU, Crocetti, D, Landman, BA, Mostofsky, SH, Thompson, PM & Prince, JL 2018, 'Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images', NeuroImage, vol. 183, pp. 150-172. https://doi.org/10.1016/j.neuroimage.2018.08.003
Carass A, Cuzzocreo JL, Han S, Hernandez-Castillo CR, Rasser PE, Ganz M et al. Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images. NeuroImage. 2018 Dec 1;183:150-172. https://doi.org/10.1016/j.neuroimage.2018.08.003
Carass, Aaron ; Cuzzocreo, Jennifer L. ; Han, Shuo ; Hernandez-Castillo, Carlos R. ; Rasser, Paul E. ; Ganz, Melanie ; Beliveau, Vincent ; Dolz, Jose ; Ben Ayed, Ismail ; Desrosiers, Christian ; Thyreau, Benjamin ; Romero, José E. ; Coupé, Pierrick ; Manjón, José V. ; Fonov, Vladimir S. ; Collins, D. Louis ; Ying, Sarah H. ; Onyike, Chiadikaobi U ; Crocetti, Deana ; Landman, Bennett A. ; Mostofsky, Stewart H ; Thompson, Paul M. ; Prince, Jerry Ladd. / Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images. In: NeuroImage. 2018 ; Vol. 183. pp. 150-172.
@article{94a93f4aa6574732ae64d01d0b4f5ad9,
title = "Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images",
abstract = "The human cerebellum plays an essential role in motor control, is involved in cognitive function (i.e., attention, working memory, and language), and helps to regulate emotional responses. Quantitative in-vivo assessment of the cerebellum is important in the study of several neurological diseases including cerebellar ataxia, autism, and schizophrenia. Different structural subdivisions of the cerebellum have been shown to correlate with differing pathologies. To further understand these pathologies, it is helpful to automatically parcellate the cerebellum at the highest fidelity possible. In this paper, we coordinated with colleagues around the world to evaluate automated cerebellum parcellation algorithms on two clinical cohorts showing that the cerebellum can be parcellated to a high accuracy by newer methods. We characterize these various methods at four hierarchical levels: coarse (i.e., whole cerebellum and gross structures), lobe, subdivisions of the vermis, and the lobules. Due to the number of labels, the hierarchy of labels, the number of algorithms, and the two cohorts, we have restricted our analyses to the Dice measure of overlap. Under these conditions, machine learning based methods provide a collection of strategies that are efficient and deliver parcellations of a high standard across both cohorts, surpassing previous work in the area. In conjunction with the rank-sum computation, we identified an overall winning method.",
keywords = "Attention deficit hyperactivity disorder, Autism, Cerebellar ataxia, Magnetic resonance imaging",
author = "Aaron Carass and Cuzzocreo, {Jennifer L.} and Shuo Han and Hernandez-Castillo, {Carlos R.} and Rasser, {Paul E.} and Melanie Ganz and Vincent Beliveau and Jose Dolz and {Ben Ayed}, Ismail and Christian Desrosiers and Benjamin Thyreau and Romero, {Jos{\'e} E.} and Pierrick Coup{\'e} and Manj{\'o}n, {Jos{\'e} V.} and Fonov, {Vladimir S.} and Collins, {D. Louis} and Ying, {Sarah H.} and Onyike, {Chiadikaobi U} and Deana Crocetti and Landman, {Bennett A.} and Mostofsky, {Stewart H} and Thompson, {Paul M.} and Prince, {Jerry Ladd}",
year = "2018",
month = "12",
day = "1",
doi = "10.1016/j.neuroimage.2018.08.003",
language = "English (US)",
volume = "183",
pages = "150--172",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Comparing fully automated state-of-the-art cerebellum parcellation from magnetic resonance images

AU - Carass, Aaron

AU - Cuzzocreo, Jennifer L.

AU - Han, Shuo

AU - Hernandez-Castillo, Carlos R.

AU - Rasser, Paul E.

AU - Ganz, Melanie

AU - Beliveau, Vincent

AU - Dolz, Jose

AU - Ben Ayed, Ismail

AU - Desrosiers, Christian

AU - Thyreau, Benjamin

AU - Romero, José E.

AU - Coupé, Pierrick

AU - Manjón, José V.

AU - Fonov, Vladimir S.

AU - Collins, D. Louis

AU - Ying, Sarah H.

AU - Onyike, Chiadikaobi U

AU - Crocetti, Deana

AU - Landman, Bennett A.

AU - Mostofsky, Stewart H

AU - Thompson, Paul M.

AU - Prince, Jerry Ladd

PY - 2018/12/1

Y1 - 2018/12/1

N2 - The human cerebellum plays an essential role in motor control, is involved in cognitive function (i.e., attention, working memory, and language), and helps to regulate emotional responses. Quantitative in-vivo assessment of the cerebellum is important in the study of several neurological diseases including cerebellar ataxia, autism, and schizophrenia. Different structural subdivisions of the cerebellum have been shown to correlate with differing pathologies. To further understand these pathologies, it is helpful to automatically parcellate the cerebellum at the highest fidelity possible. In this paper, we coordinated with colleagues around the world to evaluate automated cerebellum parcellation algorithms on two clinical cohorts showing that the cerebellum can be parcellated to a high accuracy by newer methods. We characterize these various methods at four hierarchical levels: coarse (i.e., whole cerebellum and gross structures), lobe, subdivisions of the vermis, and the lobules. Due to the number of labels, the hierarchy of labels, the number of algorithms, and the two cohorts, we have restricted our analyses to the Dice measure of overlap. Under these conditions, machine learning based methods provide a collection of strategies that are efficient and deliver parcellations of a high standard across both cohorts, surpassing previous work in the area. In conjunction with the rank-sum computation, we identified an overall winning method.

AB - The human cerebellum plays an essential role in motor control, is involved in cognitive function (i.e., attention, working memory, and language), and helps to regulate emotional responses. Quantitative in-vivo assessment of the cerebellum is important in the study of several neurological diseases including cerebellar ataxia, autism, and schizophrenia. Different structural subdivisions of the cerebellum have been shown to correlate with differing pathologies. To further understand these pathologies, it is helpful to automatically parcellate the cerebellum at the highest fidelity possible. In this paper, we coordinated with colleagues around the world to evaluate automated cerebellum parcellation algorithms on two clinical cohorts showing that the cerebellum can be parcellated to a high accuracy by newer methods. We characterize these various methods at four hierarchical levels: coarse (i.e., whole cerebellum and gross structures), lobe, subdivisions of the vermis, and the lobules. Due to the number of labels, the hierarchy of labels, the number of algorithms, and the two cohorts, we have restricted our analyses to the Dice measure of overlap. Under these conditions, machine learning based methods provide a collection of strategies that are efficient and deliver parcellations of a high standard across both cohorts, surpassing previous work in the area. In conjunction with the rank-sum computation, we identified an overall winning method.

KW - Attention deficit hyperactivity disorder

KW - Autism

KW - Cerebellar ataxia

KW - Magnetic resonance imaging

UR - http://www.scopus.com/inward/record.url?scp=85051711072&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051711072&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2018.08.003

DO - 10.1016/j.neuroimage.2018.08.003

M3 - Article

VL - 183

SP - 150

EP - 172

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -