Evaluating dynamic bivariate correlations in resting-state fMRI: A comparison study and a new approach

Martin Lindquist, Yuting Xu, Mary Beth Nebel, Brian S Caffo

Research output: Contribution to journalArticle

Abstract

To date, most functional Magnetic Resonance Imaging (fMRI) studies have assumed that the functional connectivity (FC) between time series from distinct brain regions is constant across time. However, recently, there has been an increased interest in quantifying possible dynamic changes in FC during fMRI experiments, as it is thought that this may provide insight into the fundamental workings of brain networks. In this work we focus on the specific problem of estimating the dynamic behavior of pair-wise correlations between time courses extracted from two different regions of the brain. We critique the commonly used sliding-window technique, and discuss some alternative methods used to model volatility in the finance literature that could also prove to be useful in the neuroimaging setting. In particular, we focus on the Dynamic Conditional Correlation (DCC) model, which provides a model-based approach towards estimating dynamic correlations. We investigate the properties of several techniques in a series of simulation studies and find that DCC achieves the best overall balance between sensitivity and specificity in detecting dynamic changes in correlations. We also investigate its scalability beyond the bivariate case to demonstrate its utility for studying dynamic correlations between more than two brain regions. Finally, we illustrate its performance in an application to test-retest resting state fMRI data.

Original languageEnglish (US)
Pages (from-to)531-546
Number of pages16
JournalNeuroImage
Volume101
DOIs
StatePublished - Nov 1 2014

Fingerprint

Magnetic Resonance Imaging
Brain
Volatilization
Neuroimaging
Sensitivity and Specificity

Keywords

  • Dynamic conditional correlations
  • Dynamics
  • FMRI
  • Functional connectivity
  • Resting state

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

Evaluating dynamic bivariate correlations in resting-state fMRI : A comparison study and a new approach. / Lindquist, Martin; Xu, Yuting; Nebel, Mary Beth; Caffo, Brian S.

In: NeuroImage, Vol. 101, 01.11.2014, p. 531-546.

Research output: Contribution to journalArticle

@article{f11aae3b58c848ff837dfcf57f9ddccc,
title = "Evaluating dynamic bivariate correlations in resting-state fMRI: A comparison study and a new approach",
abstract = "To date, most functional Magnetic Resonance Imaging (fMRI) studies have assumed that the functional connectivity (FC) between time series from distinct brain regions is constant across time. However, recently, there has been an increased interest in quantifying possible dynamic changes in FC during fMRI experiments, as it is thought that this may provide insight into the fundamental workings of brain networks. In this work we focus on the specific problem of estimating the dynamic behavior of pair-wise correlations between time courses extracted from two different regions of the brain. We critique the commonly used sliding-window technique, and discuss some alternative methods used to model volatility in the finance literature that could also prove to be useful in the neuroimaging setting. In particular, we focus on the Dynamic Conditional Correlation (DCC) model, which provides a model-based approach towards estimating dynamic correlations. We investigate the properties of several techniques in a series of simulation studies and find that DCC achieves the best overall balance between sensitivity and specificity in detecting dynamic changes in correlations. We also investigate its scalability beyond the bivariate case to demonstrate its utility for studying dynamic correlations between more than two brain regions. Finally, we illustrate its performance in an application to test-retest resting state fMRI data.",
keywords = "Dynamic conditional correlations, Dynamics, FMRI, Functional connectivity, Resting state",
author = "Martin Lindquist and Yuting Xu and Nebel, {Mary Beth} and Caffo, {Brian S}",
year = "2014",
month = "11",
day = "1",
doi = "10.1016/j.neuroimage.2014.06.052",
language = "English (US)",
volume = "101",
pages = "531--546",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Evaluating dynamic bivariate correlations in resting-state fMRI

T2 - A comparison study and a new approach

AU - Lindquist, Martin

AU - Xu, Yuting

AU - Nebel, Mary Beth

AU - Caffo, Brian S

PY - 2014/11/1

Y1 - 2014/11/1

N2 - To date, most functional Magnetic Resonance Imaging (fMRI) studies have assumed that the functional connectivity (FC) between time series from distinct brain regions is constant across time. However, recently, there has been an increased interest in quantifying possible dynamic changes in FC during fMRI experiments, as it is thought that this may provide insight into the fundamental workings of brain networks. In this work we focus on the specific problem of estimating the dynamic behavior of pair-wise correlations between time courses extracted from two different regions of the brain. We critique the commonly used sliding-window technique, and discuss some alternative methods used to model volatility in the finance literature that could also prove to be useful in the neuroimaging setting. In particular, we focus on the Dynamic Conditional Correlation (DCC) model, which provides a model-based approach towards estimating dynamic correlations. We investigate the properties of several techniques in a series of simulation studies and find that DCC achieves the best overall balance between sensitivity and specificity in detecting dynamic changes in correlations. We also investigate its scalability beyond the bivariate case to demonstrate its utility for studying dynamic correlations between more than two brain regions. Finally, we illustrate its performance in an application to test-retest resting state fMRI data.

AB - To date, most functional Magnetic Resonance Imaging (fMRI) studies have assumed that the functional connectivity (FC) between time series from distinct brain regions is constant across time. However, recently, there has been an increased interest in quantifying possible dynamic changes in FC during fMRI experiments, as it is thought that this may provide insight into the fundamental workings of brain networks. In this work we focus on the specific problem of estimating the dynamic behavior of pair-wise correlations between time courses extracted from two different regions of the brain. We critique the commonly used sliding-window technique, and discuss some alternative methods used to model volatility in the finance literature that could also prove to be useful in the neuroimaging setting. In particular, we focus on the Dynamic Conditional Correlation (DCC) model, which provides a model-based approach towards estimating dynamic correlations. We investigate the properties of several techniques in a series of simulation studies and find that DCC achieves the best overall balance between sensitivity and specificity in detecting dynamic changes in correlations. We also investigate its scalability beyond the bivariate case to demonstrate its utility for studying dynamic correlations between more than two brain regions. Finally, we illustrate its performance in an application to test-retest resting state fMRI data.

KW - Dynamic conditional correlations

KW - Dynamics

KW - FMRI

KW - Functional connectivity

KW - Resting state

UR - http://www.scopus.com/inward/record.url?scp=84907022675&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907022675&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2014.06.052

DO - 10.1016/j.neuroimage.2014.06.052

M3 - Article

C2 - 24993894

AN - SCOPUS:84907022675

VL - 101

SP - 531

EP - 546

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -