Improving reliability of subject-level resting-state fMRI parcellation with shrinkage estimators

Amanda F. Mejia, Mary Beth Nebel, Haochang Shou, Ciprian M Crainiceanu, James J Pekar, Stewart H Mostofsky, Brian S Caffo, Martin Lindquist

Research output: Contribution to journalArticle

Abstract

A recent interest in resting state functional magnetic resonance imaging (rsfMRI) lies in subdividing the human brain into anatomically and functionally distinct regions of interest. For example, brain parcellation is often a necessary step for defining the network nodes used in connectivity studies. While inference has traditionally been performed on group-level data, there is a growing interest in parcellating single subject data. However, this is difficult due to the inherent low signal-to-noise ratio of rsfMRI data, combined with typically short scan lengths. A large number of brain parcellation approaches employ clustering, which begins with a measure of similarity or distance between voxels. The goal of this work is to improve the reproducibility of single-subject parcellation using shrinkage-based estimators of such measures, allowing the noisy subject-specific estimator to "borrow strength" in a principled manner from a larger population of subjects. We present several empirical Bayes shrinkage estimators and outline methods for shrinkage when multiple scans are not available for each subject. We perform shrinkage on raw inter-voxel correlation estimates and use both raw and shrinkage estimates to produce parcellations by performing clustering on the voxels. While we employ a standard spectral clustering approach, our proposed method is agnostic to the choice of clustering method and can be used as a pre-processing step for any clustering algorithm. Using two datasets - a simulated dataset where the true parcellation is known and is subject-specific and a test-retest dataset consisting of two 7-minute resting-state fMRI scans from 20 subjects - we show that parcellations produced from shrinkage correlation estimates have higher reliability and validity than those produced from raw correlation estimates. Application to test-retest data shows that using shrinkage estimators increases the reproducibility of subject-specific parcellations of the motor cortex by up to 30%.

Original languageEnglish (US)
Pages (from-to)14-29
Number of pages16
JournalNeuroImage
Volume112
DOIs
StatePublished - May 5 2015

Fingerprint

Cluster Analysis
Magnetic Resonance Imaging
Brain
Motor Cortex
Signal-To-Noise Ratio
Reproducibility of Results
Population
Datasets

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

Improving reliability of subject-level resting-state fMRI parcellation with shrinkage estimators. / Mejia, Amanda F.; Nebel, Mary Beth; Shou, Haochang; Crainiceanu, Ciprian M; Pekar, James J; Mostofsky, Stewart H; Caffo, Brian S; Lindquist, Martin.

In: NeuroImage, Vol. 112, 05.05.2015, p. 14-29.

Research output: Contribution to journalArticle

@article{01d6538aca6e47bfb8b3606a5feb48f5,
title = "Improving reliability of subject-level resting-state fMRI parcellation with shrinkage estimators",
abstract = "A recent interest in resting state functional magnetic resonance imaging (rsfMRI) lies in subdividing the human brain into anatomically and functionally distinct regions of interest. For example, brain parcellation is often a necessary step for defining the network nodes used in connectivity studies. While inference has traditionally been performed on group-level data, there is a growing interest in parcellating single subject data. However, this is difficult due to the inherent low signal-to-noise ratio of rsfMRI data, combined with typically short scan lengths. A large number of brain parcellation approaches employ clustering, which begins with a measure of similarity or distance between voxels. The goal of this work is to improve the reproducibility of single-subject parcellation using shrinkage-based estimators of such measures, allowing the noisy subject-specific estimator to {"}borrow strength{"} in a principled manner from a larger population of subjects. We present several empirical Bayes shrinkage estimators and outline methods for shrinkage when multiple scans are not available for each subject. We perform shrinkage on raw inter-voxel correlation estimates and use both raw and shrinkage estimates to produce parcellations by performing clustering on the voxels. While we employ a standard spectral clustering approach, our proposed method is agnostic to the choice of clustering method and can be used as a pre-processing step for any clustering algorithm. Using two datasets - a simulated dataset where the true parcellation is known and is subject-specific and a test-retest dataset consisting of two 7-minute resting-state fMRI scans from 20 subjects - we show that parcellations produced from shrinkage correlation estimates have higher reliability and validity than those produced from raw correlation estimates. Application to test-retest data shows that using shrinkage estimators increases the reproducibility of subject-specific parcellations of the motor cortex by up to 30{\%}.",
author = "Mejia, {Amanda F.} and Nebel, {Mary Beth} and Haochang Shou and Crainiceanu, {Ciprian M} and Pekar, {James J} and Mostofsky, {Stewart H} and Caffo, {Brian S} and Martin Lindquist",
year = "2015",
month = "5",
day = "5",
doi = "10.1016/j.neuroimage.2015.02.042",
language = "English (US)",
volume = "112",
pages = "14--29",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Improving reliability of subject-level resting-state fMRI parcellation with shrinkage estimators

AU - Mejia, Amanda F.

AU - Nebel, Mary Beth

AU - Shou, Haochang

AU - Crainiceanu, Ciprian M

AU - Pekar, James J

AU - Mostofsky, Stewart H

AU - Caffo, Brian S

AU - Lindquist, Martin

PY - 2015/5/5

Y1 - 2015/5/5

N2 - A recent interest in resting state functional magnetic resonance imaging (rsfMRI) lies in subdividing the human brain into anatomically and functionally distinct regions of interest. For example, brain parcellation is often a necessary step for defining the network nodes used in connectivity studies. While inference has traditionally been performed on group-level data, there is a growing interest in parcellating single subject data. However, this is difficult due to the inherent low signal-to-noise ratio of rsfMRI data, combined with typically short scan lengths. A large number of brain parcellation approaches employ clustering, which begins with a measure of similarity or distance between voxels. The goal of this work is to improve the reproducibility of single-subject parcellation using shrinkage-based estimators of such measures, allowing the noisy subject-specific estimator to "borrow strength" in a principled manner from a larger population of subjects. We present several empirical Bayes shrinkage estimators and outline methods for shrinkage when multiple scans are not available for each subject. We perform shrinkage on raw inter-voxel correlation estimates and use both raw and shrinkage estimates to produce parcellations by performing clustering on the voxels. While we employ a standard spectral clustering approach, our proposed method is agnostic to the choice of clustering method and can be used as a pre-processing step for any clustering algorithm. Using two datasets - a simulated dataset where the true parcellation is known and is subject-specific and a test-retest dataset consisting of two 7-minute resting-state fMRI scans from 20 subjects - we show that parcellations produced from shrinkage correlation estimates have higher reliability and validity than those produced from raw correlation estimates. Application to test-retest data shows that using shrinkage estimators increases the reproducibility of subject-specific parcellations of the motor cortex by up to 30%.

AB - A recent interest in resting state functional magnetic resonance imaging (rsfMRI) lies in subdividing the human brain into anatomically and functionally distinct regions of interest. For example, brain parcellation is often a necessary step for defining the network nodes used in connectivity studies. While inference has traditionally been performed on group-level data, there is a growing interest in parcellating single subject data. However, this is difficult due to the inherent low signal-to-noise ratio of rsfMRI data, combined with typically short scan lengths. A large number of brain parcellation approaches employ clustering, which begins with a measure of similarity or distance between voxels. The goal of this work is to improve the reproducibility of single-subject parcellation using shrinkage-based estimators of such measures, allowing the noisy subject-specific estimator to "borrow strength" in a principled manner from a larger population of subjects. We present several empirical Bayes shrinkage estimators and outline methods for shrinkage when multiple scans are not available for each subject. We perform shrinkage on raw inter-voxel correlation estimates and use both raw and shrinkage estimates to produce parcellations by performing clustering on the voxels. While we employ a standard spectral clustering approach, our proposed method is agnostic to the choice of clustering method and can be used as a pre-processing step for any clustering algorithm. Using two datasets - a simulated dataset where the true parcellation is known and is subject-specific and a test-retest dataset consisting of two 7-minute resting-state fMRI scans from 20 subjects - we show that parcellations produced from shrinkage correlation estimates have higher reliability and validity than those produced from raw correlation estimates. Application to test-retest data shows that using shrinkage estimators increases the reproducibility of subject-specific parcellations of the motor cortex by up to 30%.

UR - http://www.scopus.com/inward/record.url?scp=84924347949&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84924347949&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2015.02.042

DO - 10.1016/j.neuroimage.2015.02.042

M3 - Article

C2 - 25731998

AN - SCOPUS:84924347949

VL - 112

SP - 14

EP - 29

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -