EURASIP Journal on Applied Signal Processing
Volume 2005 (2005), Issue 19, Pages 3089-3102
doi:10.1155/ASP.2005.3089

Clustering of Dependent Components: A New Paradigm for fMRI Signal Detection

Anke Meyer-Bäse,1 Monica K. Hurdal,2 Oliver Lange,1 and Helge Ritter3

1Department of Electrical and Computer Engineering, Florida State University, Tallahassee 32310-6046, FL, USA
2Department of Mathematics, Florida State University, Tallahassee 32306-4510, FL, USA
3Neuroinformatics Group, Faculty of Technology, University of Bielefeld, Bielefeld 33501, Germany

Received 1 February 2004

Copyright © 2005 Anke Meyer-Bäse et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Exploratory data-driven methods such as unsupervised clustering and independent component analysis (ICA) are considered to be hypothesis-generating procedures and are complementary to the hypothesis-led statistical inferential methods in functional magnetic resonance imaging (fMRI). Recently, a new paradigm in ICA emerged, that of finding “clusters” of dependent components. This intriguing idea found its implementation into two new ICA algorithms: tree-dependent and topographic ICA. For fMRI, this represents the unifying paradigm of combining two powerful exploratory data analysis methods, ICA and unsupervised clustering techniques. For the fMRI data, a comparative quantitative evaluation between the two methods, tree-dependent and topographic ICA, was performed. The comparative results were evaluated by (1) task-related activation maps, (2) associated time courses, and (3) ROC study. The most important findings in this paper are that (1) both tree-dependent and topographic ICA are able to identify signal components with high correlation to the fMRI stimulus, and that (2) topographic ICA outperforms all other ICA methods including tree-dependent ICA for 8 and 9 ICs. However for 16 ICs, topographic ICA is outperformed by tree-dependent ICA (KGV) using as an approximation of the mutual information the kernel generalized variance. The applicability of the new algorithm is demonstrated on experimental data.