TY - GEN

T1 - Semi-blind ICA of FMRI

T2 - Machine Learning for Signal Processing XIV - Proceedings of the 2004 IEEE Signal Processing Society Workshop

AU - Calhoun, V.

AU - Adali, T.

N1 - Funding Information:
We would like to thank the research staff at the Olin Neuropsychiatry Research Center who helped collect and process the data. Supported by the National Institutes of Health under grant 1 R01 EB 000840-01 (to VC).

PY - 2004

Y1 - 2004

N2 - Independent component analysis (ICA), a data-driven approach utilizing high-order statistical moments to find maximally independent sources, has found fruitful application in functional magnetic resonance imaging (fMRI). ICA, being a blind source separation technique, does not require any explicit constraints upon the fMRI time courses. In some cases, such as for the analysis of a rapid eventrelated paradigm, it would be useful to incorporate paradigm information into the ICA analysis in a flexible way. In this paper, we present an approach for constrained or semi-blind ICA (sbICA) analysis of fMRI data. We demonstrate the performance of our approach using simulations and fMRI data of an auditory oddball paradigm. Simulation results suggest that 1) a regression approach slightly outperforms ICA when prior information is accurate and ICA outperforms the general linear modeling (GLM) approach when prior information is not completely accurate, 2) prior information improves the robustness of ICA in the presence of noise, and 3) and ICA analysis using prior information with weak constraints can outperform a regression approach when the prior information is not completely accurate.

AB - Independent component analysis (ICA), a data-driven approach utilizing high-order statistical moments to find maximally independent sources, has found fruitful application in functional magnetic resonance imaging (fMRI). ICA, being a blind source separation technique, does not require any explicit constraints upon the fMRI time courses. In some cases, such as for the analysis of a rapid eventrelated paradigm, it would be useful to incorporate paradigm information into the ICA analysis in a flexible way. In this paper, we present an approach for constrained or semi-blind ICA (sbICA) analysis of fMRI data. We demonstrate the performance of our approach using simulations and fMRI data of an auditory oddball paradigm. Simulation results suggest that 1) a regression approach slightly outperforms ICA when prior information is accurate and ICA outperforms the general linear modeling (GLM) approach when prior information is not completely accurate, 2) prior information improves the robustness of ICA in the presence of noise, and 3) and ICA analysis using prior information with weak constraints can outperform a regression approach when the prior information is not completely accurate.

UR - http://www.scopus.com/inward/record.url?scp=17644372786&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=17644372786&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:17644372786

SN - 0780386086

SN - 9780780386082

T3 - Machine Learning for Signal Processing XIV - Proceedings of the 2004 IEEE Signal Processing Society Workshop

SP - 443

EP - 452

BT - Machine Learning for Signal Processing XIV - Proceedings of 2004 IEEE Signal Processing Society Workshop

A2 - Barros, A.

A2 - Principe, J.

A2 - Larsen, J.

A2 - Adali, T.

A2 - Douglas, S.

Y2 - 29 September 2004 through 1 October 2004

ER -