Consistency of magnetoencephalographic functional connectivity and network reconstruction using a template versus native MRI for co-registration

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Introduction: Studies using functional connectivity and network analyses based on magnetoencephalography (MEG) with source localization are rapidly emerging in neuroscientific literature. However, these analyses currently depend on the availability of costly and sometimes burdensome individual MR scans for co-registration. We evaluated the consistency of these measures when using a template MRI, instead of native MRI, for the analysis of functional connectivity and network topology. Methods: Seventeen healthy participants underwent resting-state eyes-closed MEG and anatomical MRI. These data were projected into source space using an atlas-based peak voxel and a centroid beamforming approach either using (1) participants' native MRIs or (2) the Montreal Neurological Institute's template. For both methods, time series were reconstructed from 78 cortical atlas regions. Relative power was determined in six classical frequency bands per region and globally averaged. Functional connectivity (phase lag index) between each pair of regions was calculated. The adjacency matrices were then used to reconstruct functional networks, of which regional and global metrics were determined. Intraclass correlation coefficients were calculated and Bland-Altman plots were made to quantify the consistency and potential bias of the use of template versus native MRI. Results: Co-registration with the template yielded largely consistent relative power, connectivity, and network estimates compared to native MRI. Discussion: These findings indicate that there is no (systematic) bias or inconsistency between template and native MRI co-registration of MEG. They open up possibilities for retrospective and prospective analyses to MEG datasets in the general population that have no native MRIs available.

Original languageEnglish
Pages (from-to)104-119
Number of pages16
JournalHuman Brain Mapping
Volume39
Issue number1
DOIs
Publication statusPublished - 1 Jan 2018

Cite this

@article{12cd0784ccda47f2ba234f89e9179a54,
title = "Consistency of magnetoencephalographic functional connectivity and network reconstruction using a template versus native MRI for co-registration",
abstract = "Introduction: Studies using functional connectivity and network analyses based on magnetoencephalography (MEG) with source localization are rapidly emerging in neuroscientific literature. However, these analyses currently depend on the availability of costly and sometimes burdensome individual MR scans for co-registration. We evaluated the consistency of these measures when using a template MRI, instead of native MRI, for the analysis of functional connectivity and network topology. Methods: Seventeen healthy participants underwent resting-state eyes-closed MEG and anatomical MRI. These data were projected into source space using an atlas-based peak voxel and a centroid beamforming approach either using (1) participants' native MRIs or (2) the Montreal Neurological Institute's template. For both methods, time series were reconstructed from 78 cortical atlas regions. Relative power was determined in six classical frequency bands per region and globally averaged. Functional connectivity (phase lag index) between each pair of regions was calculated. The adjacency matrices were then used to reconstruct functional networks, of which regional and global metrics were determined. Intraclass correlation coefficients were calculated and Bland-Altman plots were made to quantify the consistency and potential bias of the use of template versus native MRI. Results: Co-registration with the template yielded largely consistent relative power, connectivity, and network estimates compared to native MRI. Discussion: These findings indicate that there is no (systematic) bias or inconsistency between template and native MRI co-registration of MEG. They open up possibilities for retrospective and prospective analyses to MEG datasets in the general population that have no native MRIs available.",
keywords = "Beamforming, Co-registration, Consistency, Functional connectivity, Magnetoencephalography, MNI template, MRI, Network analyses",
author = "Linda Douw and Dagmar Nieboer and Stam, {Cornelis J.} and Prejaas Tewarie and Arjan Hillebrand",
year = "2018",
month = "1",
day = "1",
doi = "10.1002/hbm.23827",
language = "English",
volume = "39",
pages = "104--119",
journal = "Human Brain Mapping",
issn = "1065-9471",
publisher = "Wiley-Liss Inc.",
number = "1",

}

TY - JOUR

T1 - Consistency of magnetoencephalographic functional connectivity and network reconstruction using a template versus native MRI for co-registration

AU - Douw, Linda

AU - Nieboer, Dagmar

AU - Stam, Cornelis J.

AU - Tewarie, Prejaas

AU - Hillebrand, Arjan

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Introduction: Studies using functional connectivity and network analyses based on magnetoencephalography (MEG) with source localization are rapidly emerging in neuroscientific literature. However, these analyses currently depend on the availability of costly and sometimes burdensome individual MR scans for co-registration. We evaluated the consistency of these measures when using a template MRI, instead of native MRI, for the analysis of functional connectivity and network topology. Methods: Seventeen healthy participants underwent resting-state eyes-closed MEG and anatomical MRI. These data were projected into source space using an atlas-based peak voxel and a centroid beamforming approach either using (1) participants' native MRIs or (2) the Montreal Neurological Institute's template. For both methods, time series were reconstructed from 78 cortical atlas regions. Relative power was determined in six classical frequency bands per region and globally averaged. Functional connectivity (phase lag index) between each pair of regions was calculated. The adjacency matrices were then used to reconstruct functional networks, of which regional and global metrics were determined. Intraclass correlation coefficients were calculated and Bland-Altman plots were made to quantify the consistency and potential bias of the use of template versus native MRI. Results: Co-registration with the template yielded largely consistent relative power, connectivity, and network estimates compared to native MRI. Discussion: These findings indicate that there is no (systematic) bias or inconsistency between template and native MRI co-registration of MEG. They open up possibilities for retrospective and prospective analyses to MEG datasets in the general population that have no native MRIs available.

AB - Introduction: Studies using functional connectivity and network analyses based on magnetoencephalography (MEG) with source localization are rapidly emerging in neuroscientific literature. However, these analyses currently depend on the availability of costly and sometimes burdensome individual MR scans for co-registration. We evaluated the consistency of these measures when using a template MRI, instead of native MRI, for the analysis of functional connectivity and network topology. Methods: Seventeen healthy participants underwent resting-state eyes-closed MEG and anatomical MRI. These data were projected into source space using an atlas-based peak voxel and a centroid beamforming approach either using (1) participants' native MRIs or (2) the Montreal Neurological Institute's template. For both methods, time series were reconstructed from 78 cortical atlas regions. Relative power was determined in six classical frequency bands per region and globally averaged. Functional connectivity (phase lag index) between each pair of regions was calculated. The adjacency matrices were then used to reconstruct functional networks, of which regional and global metrics were determined. Intraclass correlation coefficients were calculated and Bland-Altman plots were made to quantify the consistency and potential bias of the use of template versus native MRI. Results: Co-registration with the template yielded largely consistent relative power, connectivity, and network estimates compared to native MRI. Discussion: These findings indicate that there is no (systematic) bias or inconsistency between template and native MRI co-registration of MEG. They open up possibilities for retrospective and prospective analyses to MEG datasets in the general population that have no native MRIs available.

KW - Beamforming

KW - Co-registration

KW - Consistency

KW - Functional connectivity

KW - Magnetoencephalography

KW - MNI template

KW - MRI

KW - Network analyses

UR - http://www.scopus.com/inward/record.url?scp=85030627302&partnerID=8YFLogxK

U2 - 10.1002/hbm.23827

DO - 10.1002/hbm.23827

M3 - Article

VL - 39

SP - 104

EP - 119

JO - Human Brain Mapping

JF - Human Brain Mapping

SN - 1065-9471

IS - 1

ER -