Retrospective motion artifact correction of structural MRI images using deep learning improves the quality of cortical surface reconstructions.
Adolescent
Adult
Artifacts
Autistic Disorder
/ diagnostic imaging
Cerebral Cortex
/ diagnostic imaging
Child
Databases, Factual
/ standards
Deep Learning
/ standards
Female
Humans
Image Processing, Computer-Assisted
/ methods
Magnetic Resonance Imaging
/ methods
Male
Middle Aged
Motion
Retrospective Studies
Young Adult
Cortical surface
Cortical thickness
Image quality
Motion artifact
Parkinson's disease
T1
Journal
NeuroImage
ISSN: 1095-9572
Titre abrégé: Neuroimage
Pays: United States
ID NLM: 9215515
Informations de publication
Date de publication:
15 04 2021
15 04 2021
Historique:
received:
15
09
2020
revised:
23
12
2020
accepted:
07
01
2021
pubmed:
19
1
2021
medline:
13
10
2021
entrez:
18
1
2021
Statut:
ppublish
Résumé
Head motion during MRI acquisition presents significant challenges for neuroimaging analyses. In this work, we present a retrospective motion correction framework built on a Fourier domain motion simulation model combined with established 3D convolutional neural network (CNN) architectures. Quantitative evaluation metrics were used to validate the method on three separate multi-site datasets. The 3D CNN was trained using motion-free images that were corrupted using simulated artifacts. CNN based correction successfully diminished the severity of artifacts on real motion affected data on a separate test dataset as measured by significant improvements in image quality metrics compared to a minimal motion reference image. On the test set of 13 image pairs, the mean peak signal-to-noise-ratio was improved from 31.7 to 33.3 dB. Furthermore, improvements in cortical surface reconstruction quality were demonstrated using a blinded manual quality assessment on the Parkinson's Progression Markers Initiative (PPMI) dataset. Upon applying the correction algorithm, out of a total of 617 images, the number of quality control failures was reduced from 61 to 38. On this same dataset, we investigated whether motion correction resulted in a more statistically significant relationship between cortical thickness and Parkinson's disease. Before correction, significant cortical thinning was found to be restricted to limited regions within the temporal and frontal lobes. After correction, there was found to be more widespread and significant cortical thinning bilaterally across the temporal lobes and frontal cortex. Our results highlight the utility of image domain motion correction for use in studies with a high prevalence of motion artifacts, such as studies of movement disorders as well as infant and pediatric subjects.
Identifiants
pubmed: 33460797
pii: S1053-8119(21)00033-1
doi: 10.1016/j.neuroimage.2021.117756
pmc: PMC8044025
mid: NIHMS1671804
pii:
doi:
Types de publication
Journal Article
Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't
Research Support, U.S. Gov't, Non-P.H.S.
Langues
eng
Sous-ensembles de citation
IM
Pagination
117756Subventions
Organisme : NIBIB NIH HHS
ID : P41 EB015922
Pays : United States
Organisme : NIBIB NIH HHS
ID : U54 EB020406
Pays : United States
Organisme : CIHR
Pays : Canada
Organisme : NIA NIH HHS
ID : U01 AG024904
Pays : United States
Organisme : NIBIB NIH HHS
ID : R01 EB028297
Pays : United States
Organisme : NIA NIH HHS
ID : P01 AG012435
Pays : United States
Organisme : NIA NIH HHS
ID : P30 AG066530
Pays : United States
Informations de copyright
Copyright © 2021. Published by Elsevier Inc.
Références
MAGMA. 2018 Apr;31(2):243-256
pubmed: 28932991
IEEE Trans Med Imaging. 1991;10(4):548-53
pubmed: 18222860
Mov Disord. 2015 Dec;30(14):1893-900
pubmed: 25759166
IEEE Trans Med Imaging. 1998 Feb;17(1):87-97
pubmed: 9617910
Magn Reson Med. 2013 Dec;70(6):1608-18
pubmed: 23401078
Mov Disord. 2010 Mar 15;25(4):496-9
pubmed: 20108369
Magn Reson Med. 2019 Sep;82(3):901-910
pubmed: 31006909
Med Image Anal. 2017 Feb;36:61-78
pubmed: 27865153
Nature. 2018 Mar 21;555(7697):487-492
pubmed: 29565357
Magn Reson Med. 2019 Oct;82(4):1527-1540
pubmed: 31081955
IEEE J Biomed Health Inform. 2019 May;23(3):923-930
pubmed: 30561355
Neuroimage. 2017 Jan 1;144(Pt A):128-141
pubmed: 27664827
PLoS One. 2015 Jul 30;10(7):e0133921
pubmed: 26226146
J Magn Reson Imaging. 2015 Oct;42(4):887-901
pubmed: 25630632
Magn Reson Med. 2012 Aug;68(2):389-99
pubmed: 22213578
Magn Reson Med. 2018 Jun;79(6):3055-3071
pubmed: 29115689
IEEE Trans Med Imaging. 2018 May;37(5):1253-1265
pubmed: 29727288
Hum Brain Mapp. 2012 Nov;33(11):2521-34
pubmed: 21898679
Comput Methods Programs Biomed. 2018 May;158:113-122
pubmed: 29544777
J Neurol Sci. 2019 Mar 15;398:31-38
pubmed: 30682518
PLoS One. 2013;8(1):e54980
pubmed: 23359616
Neuroimage. 2013 Oct 15;80:105-24
pubmed: 23668970
Magn Reson Med. 2009 Aug;62(2):365-72
pubmed: 19526493
Brain. 2015 Oct;138(Pt 10):2974-86
pubmed: 26173861
Proc Natl Acad Sci U S A. 2000 Sep 26;97(20):11050-5
pubmed: 10984517
Magn Reson Med. 2018 Mar;79(3):1365-1376
pubmed: 28626962
Mol Psychiatry. 2014 Jun;19(6):659-67
pubmed: 23774715
Magn Reson Med. 2010 Jan;63(1):91-105
pubmed: 20027635
IEEE Trans Med Imaging. 2015 Oct;34(10):1993-2024
pubmed: 25494501
Neuroimage. 2015 Feb 15;107:107-115
pubmed: 25498430
Neuroimage. 2006 Jul 1;31(3):968-80
pubmed: 16530430
IEEE Trans Med Imaging. 2000 Feb;19(2):143-50
pubmed: 10784285
Med Phys. 2019 Sep;46(9):3788-3798
pubmed: 31220353
IEEE Trans Med Imaging. 2010 Jun;29(6):1310-20
pubmed: 20378467
Mov Disord. 2011 Feb 15;26(3):436-41
pubmed: 21462259
IEEE Trans Med Imaging. 1997 Dec;16(6):903-10
pubmed: 9533590
IEEE Trans Image Process. 2004 Apr;13(4):600-12
pubmed: 15376593
Neuroimage. 2012 Mar;60(1):623-32
pubmed: 22233733
Neuroimage. 2018 Apr 15;170:456-470
pubmed: 28450139
Neuroimage. 2016 Sep;138:28-42
pubmed: 27184202
Magn Reson Med. 2020 Aug;84(2):
pubmed: 31898832
Pediatr Radiol. 2013 Jan;43(1):15-27
pubmed: 23288475
Mov Disord. 2016 May;31(5):699-708
pubmed: 27094093
Ann Neurol. 2015 Jan;77(1):154-62
pubmed: 25425403