The effect of preprocessing in dynamic functional network connectivity used to classify mild traumatic brain injury

Victor M. Vergara, Andrew R. Mayer, Eswar Damaraju, Vince Daniel Calhoun

Research output: Contribution to journalArticle

Abstract

Introduction: Dynamic functional network connectivity (dFNC), derived from magnetic resonance imaging (fMRI), is an important technique in the search for biomarkers of brain diseases such as mild traumatic brain injury (mTBI). At the individual level, mTBI can affect cognitive functions and change personality traits. Previous research aimed at detecting significant changes in the dFNC of mTBI subjects. However, one of the main concerns in dFNC analysis is the appropriateness of methods used to correct for subject movement. In this work, we focus on the effect that rearranging movement correction at different points of the processing pipeline has in dFNC analysis utilizing mTBI data. Methods: The sample cohort consists of 50 mTBI patients and matched healthy controls. A 5-min resting-state run was completed by each participant. Data were preprocessed using different pipeline alternatives varying with the place where motion-related variance was removed. In all pipelines, group-independent component analysis (gICA) followed by dFNC analysis was performed. Additional tests were performed varying the detection of temporal spikes, the number of gICA components, and the sliding-window size. A linear support vector machine was used to test how each pipeline affects classification accuracy. Results: Results suggest that correction for motion variance before spatial smoothing, but leaving correction for spiky time courses after gICA produced the best mean classification performance. The number of gICA components and the sliding-window size were also important in determining classification performance. Variance in spikes correction affected some pipelines more than others with fewer significant differences than the other parameters. Conclusion: The sequence of preprocessing steps motion regression, smoothing, gICA, and despiking produced data most suitable for differentiating mTBI from healthy subjects. However, the selection of optimal preprocessing parameters strongly affected the final results.

Original languageEnglish (US)
JournalBrain and Behavior
DOIs
StateAccepted/In press - 2017
Externally publishedYes

Fingerprint

Brain Concussion
Brain Diseases
Cognition
Personality
Healthy Volunteers
Biomarkers
Magnetic Resonance Imaging
Research

Keywords

  • Dynamic functional network connectivity
  • Functional MRI
  • Independent component analysis
  • Traumatic brain injury

ASJC Scopus subject areas

  • Behavioral Neuroscience

Cite this

The effect of preprocessing in dynamic functional network connectivity used to classify mild traumatic brain injury. / Vergara, Victor M.; Mayer, Andrew R.; Damaraju, Eswar; Calhoun, Vince Daniel.

In: Brain and Behavior, 2017.

Research output: Contribution to journalArticle

@article{4ffddd60443040d99e4b7c140cf4a9cd,
title = "The effect of preprocessing in dynamic functional network connectivity used to classify mild traumatic brain injury",
abstract = "Introduction: Dynamic functional network connectivity (dFNC), derived from magnetic resonance imaging (fMRI), is an important technique in the search for biomarkers of brain diseases such as mild traumatic brain injury (mTBI). At the individual level, mTBI can affect cognitive functions and change personality traits. Previous research aimed at detecting significant changes in the dFNC of mTBI subjects. However, one of the main concerns in dFNC analysis is the appropriateness of methods used to correct for subject movement. In this work, we focus on the effect that rearranging movement correction at different points of the processing pipeline has in dFNC analysis utilizing mTBI data. Methods: The sample cohort consists of 50 mTBI patients and matched healthy controls. A 5-min resting-state run was completed by each participant. Data were preprocessed using different pipeline alternatives varying with the place where motion-related variance was removed. In all pipelines, group-independent component analysis (gICA) followed by dFNC analysis was performed. Additional tests were performed varying the detection of temporal spikes, the number of gICA components, and the sliding-window size. A linear support vector machine was used to test how each pipeline affects classification accuracy. Results: Results suggest that correction for motion variance before spatial smoothing, but leaving correction for spiky time courses after gICA produced the best mean classification performance. The number of gICA components and the sliding-window size were also important in determining classification performance. Variance in spikes correction affected some pipelines more than others with fewer significant differences than the other parameters. Conclusion: The sequence of preprocessing steps motion regression, smoothing, gICA, and despiking produced data most suitable for differentiating mTBI from healthy subjects. However, the selection of optimal preprocessing parameters strongly affected the final results.",
keywords = "Dynamic functional network connectivity, Functional MRI, Independent component analysis, Traumatic brain injury",
author = "Vergara, {Victor M.} and Mayer, {Andrew R.} and Eswar Damaraju and Calhoun, {Vince Daniel}",
year = "2017",
doi = "10.1002/brb3.809",
language = "English (US)",
journal = "Brain and Behavior",
issn = "2157-9032",
publisher = "John Wiley and Sons Inc.",

}

TY - JOUR

T1 - The effect of preprocessing in dynamic functional network connectivity used to classify mild traumatic brain injury

AU - Vergara, Victor M.

AU - Mayer, Andrew R.

AU - Damaraju, Eswar

AU - Calhoun, Vince Daniel

PY - 2017

Y1 - 2017

N2 - Introduction: Dynamic functional network connectivity (dFNC), derived from magnetic resonance imaging (fMRI), is an important technique in the search for biomarkers of brain diseases such as mild traumatic brain injury (mTBI). At the individual level, mTBI can affect cognitive functions and change personality traits. Previous research aimed at detecting significant changes in the dFNC of mTBI subjects. However, one of the main concerns in dFNC analysis is the appropriateness of methods used to correct for subject movement. In this work, we focus on the effect that rearranging movement correction at different points of the processing pipeline has in dFNC analysis utilizing mTBI data. Methods: The sample cohort consists of 50 mTBI patients and matched healthy controls. A 5-min resting-state run was completed by each participant. Data were preprocessed using different pipeline alternatives varying with the place where motion-related variance was removed. In all pipelines, group-independent component analysis (gICA) followed by dFNC analysis was performed. Additional tests were performed varying the detection of temporal spikes, the number of gICA components, and the sliding-window size. A linear support vector machine was used to test how each pipeline affects classification accuracy. Results: Results suggest that correction for motion variance before spatial smoothing, but leaving correction for spiky time courses after gICA produced the best mean classification performance. The number of gICA components and the sliding-window size were also important in determining classification performance. Variance in spikes correction affected some pipelines more than others with fewer significant differences than the other parameters. Conclusion: The sequence of preprocessing steps motion regression, smoothing, gICA, and despiking produced data most suitable for differentiating mTBI from healthy subjects. However, the selection of optimal preprocessing parameters strongly affected the final results.

AB - Introduction: Dynamic functional network connectivity (dFNC), derived from magnetic resonance imaging (fMRI), is an important technique in the search for biomarkers of brain diseases such as mild traumatic brain injury (mTBI). At the individual level, mTBI can affect cognitive functions and change personality traits. Previous research aimed at detecting significant changes in the dFNC of mTBI subjects. However, one of the main concerns in dFNC analysis is the appropriateness of methods used to correct for subject movement. In this work, we focus on the effect that rearranging movement correction at different points of the processing pipeline has in dFNC analysis utilizing mTBI data. Methods: The sample cohort consists of 50 mTBI patients and matched healthy controls. A 5-min resting-state run was completed by each participant. Data were preprocessed using different pipeline alternatives varying with the place where motion-related variance was removed. In all pipelines, group-independent component analysis (gICA) followed by dFNC analysis was performed. Additional tests were performed varying the detection of temporal spikes, the number of gICA components, and the sliding-window size. A linear support vector machine was used to test how each pipeline affects classification accuracy. Results: Results suggest that correction for motion variance before spatial smoothing, but leaving correction for spiky time courses after gICA produced the best mean classification performance. The number of gICA components and the sliding-window size were also important in determining classification performance. Variance in spikes correction affected some pipelines more than others with fewer significant differences than the other parameters. Conclusion: The sequence of preprocessing steps motion regression, smoothing, gICA, and despiking produced data most suitable for differentiating mTBI from healthy subjects. However, the selection of optimal preprocessing parameters strongly affected the final results.

KW - Dynamic functional network connectivity

KW - Functional MRI

KW - Independent component analysis

KW - Traumatic brain injury

UR - http://www.scopus.com/inward/record.url?scp=85029428934&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029428934&partnerID=8YFLogxK

U2 - 10.1002/brb3.809

DO - 10.1002/brb3.809

M3 - Article

C2 - 29075569

AN - SCOPUS:85029428934

JO - Brain and Behavior

JF - Brain and Behavior

SN - 2157-9032

ER -