Performance of blind source separation algorithms for fMRI analysis using a group ICA method

Nicolle Correa, Tülay Adali, Vince Daniel Calhoun

Research output: Contribution to journalArticle

Abstract

Independent component analysis (ICA) is a popular blind source separation technique that has proven to be promising for the analysis of functional magnetic resonance imaging (fMRI) data. A number of ICA approaches have been used for fMRI data analysis, and even more ICA algorithms exist; however, the impact of using different algorithms on the results is largely unexplored. In this paper, we study the performance of four major classes of algorithms for spatial ICA, namely, information maximization, maximization of non-Gaussianity, joint diagonalization of cross-cumulant matrices and second-order correlation-based methods, when they are applied to fMRI data from subjects performing a visuo-motor task. We use a group ICA method to study variability among different ICA algorithms, and we propose several analysis techniques to evaluate their performance. We compare how different ICA algorithms estimate activations in expected neuronal areas. The results demonstrate that the ICA algorithms using higher-order statistical information prove to be quite consistent for fMRI data analysis. Infomax, FastICA and joint approximate diagonalization of eigenmatrices (JADE) all yield reliable results, with each having its strengths in specific areas. Eigenvalue decomposition (EVD), an algorithm using second-order statistics, does not perform reliably for fMRI data. Additionally, for iterative ICA algorithms, it is important to investigate the variability of estimates from different runs. We test the consistency of the iterative algorithms Infomax and FastICA by running the algorithm a number of times with different initializations, and we note that they yield consistent results over these multiple runs. Our results greatly improve our confidence in the consistency of ICA for fMRI data analysis.

Original languageEnglish (US)
Pages (from-to)684-694
Number of pages11
JournalMagnetic Resonance Imaging
Volume25
Issue number5
DOIs
StatePublished - Jun 2007
Externally publishedYes

Fingerprint

Blind source separation
Independent component analysis
magnetic resonance
Magnetic Resonance Imaging
Joints
information analysis
estimates
confidence
eigenvalues
Chemical activation
Statistics
statistics
activation
Decomposition
decomposition

Keywords

  • fMRI
  • Functional
  • ICA
  • Independent component analysis
  • Visuomotor

ASJC Scopus subject areas

  • Biophysics
  • Clinical Biochemistry
  • Structural Biology
  • Radiology Nuclear Medicine and imaging
  • Condensed Matter Physics

Cite this

Performance of blind source separation algorithms for fMRI analysis using a group ICA method. / Correa, Nicolle; Adali, Tülay; Calhoun, Vince Daniel.

In: Magnetic Resonance Imaging, Vol. 25, No. 5, 06.2007, p. 684-694.

Research output: Contribution to journalArticle

@article{d1a341638bb14468897393ef3113e45c,
title = "Performance of blind source separation algorithms for fMRI analysis using a group ICA method",
abstract = "Independent component analysis (ICA) is a popular blind source separation technique that has proven to be promising for the analysis of functional magnetic resonance imaging (fMRI) data. A number of ICA approaches have been used for fMRI data analysis, and even more ICA algorithms exist; however, the impact of using different algorithms on the results is largely unexplored. In this paper, we study the performance of four major classes of algorithms for spatial ICA, namely, information maximization, maximization of non-Gaussianity, joint diagonalization of cross-cumulant matrices and second-order correlation-based methods, when they are applied to fMRI data from subjects performing a visuo-motor task. We use a group ICA method to study variability among different ICA algorithms, and we propose several analysis techniques to evaluate their performance. We compare how different ICA algorithms estimate activations in expected neuronal areas. The results demonstrate that the ICA algorithms using higher-order statistical information prove to be quite consistent for fMRI data analysis. Infomax, FastICA and joint approximate diagonalization of eigenmatrices (JADE) all yield reliable results, with each having its strengths in specific areas. Eigenvalue decomposition (EVD), an algorithm using second-order statistics, does not perform reliably for fMRI data. Additionally, for iterative ICA algorithms, it is important to investigate the variability of estimates from different runs. We test the consistency of the iterative algorithms Infomax and FastICA by running the algorithm a number of times with different initializations, and we note that they yield consistent results over these multiple runs. Our results greatly improve our confidence in the consistency of ICA for fMRI data analysis.",
keywords = "fMRI, Functional, ICA, Independent component analysis, Visuomotor",
author = "Nicolle Correa and T{\"u}lay Adali and Calhoun, {Vince Daniel}",
year = "2007",
month = "6",
doi = "10.1016/j.mri.2006.10.017",
language = "English (US)",
volume = "25",
pages = "684--694",
journal = "Magnetic Resonance Imaging",
issn = "0730-725X",
publisher = "Elsevier Inc.",
number = "5",

}

TY - JOUR

T1 - Performance of blind source separation algorithms for fMRI analysis using a group ICA method

AU - Correa, Nicolle

AU - Adali, Tülay

AU - Calhoun, Vince Daniel

PY - 2007/6

Y1 - 2007/6

N2 - Independent component analysis (ICA) is a popular blind source separation technique that has proven to be promising for the analysis of functional magnetic resonance imaging (fMRI) data. A number of ICA approaches have been used for fMRI data analysis, and even more ICA algorithms exist; however, the impact of using different algorithms on the results is largely unexplored. In this paper, we study the performance of four major classes of algorithms for spatial ICA, namely, information maximization, maximization of non-Gaussianity, joint diagonalization of cross-cumulant matrices and second-order correlation-based methods, when they are applied to fMRI data from subjects performing a visuo-motor task. We use a group ICA method to study variability among different ICA algorithms, and we propose several analysis techniques to evaluate their performance. We compare how different ICA algorithms estimate activations in expected neuronal areas. The results demonstrate that the ICA algorithms using higher-order statistical information prove to be quite consistent for fMRI data analysis. Infomax, FastICA and joint approximate diagonalization of eigenmatrices (JADE) all yield reliable results, with each having its strengths in specific areas. Eigenvalue decomposition (EVD), an algorithm using second-order statistics, does not perform reliably for fMRI data. Additionally, for iterative ICA algorithms, it is important to investigate the variability of estimates from different runs. We test the consistency of the iterative algorithms Infomax and FastICA by running the algorithm a number of times with different initializations, and we note that they yield consistent results over these multiple runs. Our results greatly improve our confidence in the consistency of ICA for fMRI data analysis.

AB - Independent component analysis (ICA) is a popular blind source separation technique that has proven to be promising for the analysis of functional magnetic resonance imaging (fMRI) data. A number of ICA approaches have been used for fMRI data analysis, and even more ICA algorithms exist; however, the impact of using different algorithms on the results is largely unexplored. In this paper, we study the performance of four major classes of algorithms for spatial ICA, namely, information maximization, maximization of non-Gaussianity, joint diagonalization of cross-cumulant matrices and second-order correlation-based methods, when they are applied to fMRI data from subjects performing a visuo-motor task. We use a group ICA method to study variability among different ICA algorithms, and we propose several analysis techniques to evaluate their performance. We compare how different ICA algorithms estimate activations in expected neuronal areas. The results demonstrate that the ICA algorithms using higher-order statistical information prove to be quite consistent for fMRI data analysis. Infomax, FastICA and joint approximate diagonalization of eigenmatrices (JADE) all yield reliable results, with each having its strengths in specific areas. Eigenvalue decomposition (EVD), an algorithm using second-order statistics, does not perform reliably for fMRI data. Additionally, for iterative ICA algorithms, it is important to investigate the variability of estimates from different runs. We test the consistency of the iterative algorithms Infomax and FastICA by running the algorithm a number of times with different initializations, and we note that they yield consistent results over these multiple runs. Our results greatly improve our confidence in the consistency of ICA for fMRI data analysis.

KW - fMRI

KW - Functional

KW - ICA

KW - Independent component analysis

KW - Visuomotor

UR - http://www.scopus.com/inward/record.url?scp=34249321979&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34249321979&partnerID=8YFLogxK

U2 - 10.1016/j.mri.2006.10.017

DO - 10.1016/j.mri.2006.10.017

M3 - Article

C2 - 17540281

AN - SCOPUS:34249321979

VL - 25

SP - 684

EP - 694

JO - Magnetic Resonance Imaging

JF - Magnetic Resonance Imaging

SN - 0730-725X

IS - 5

ER -