Chapter

Independent Component Analysis and Blind Signal Separation

Volume 3195 of the series Lecture Notes in Computer Science pp 782-789

Tree-Dependent and Topographic Independent Component Analysis for fMRI Analysis

  • Anke Meyer-BäseAffiliated withLancaster UniversityDepartment of Electrical and Computer Engineering, Florida State University
  • , Fabian J. TheisAffiliated withLancaster UniversityDepartment of Electrical and Computer Engineering, Florida State UniversityInstitute of Biophysics, University of Regensburg
  • , Oliver LangeAffiliated withLancaster UniversityCarnegie Mellon UniversityDepartment of Electrical and Computer Engineering, Florida State UniversityDepartment of Clinical Radiology, Ludwig–Maximilians University
  • , Carlos G. PuntonetAffiliated withCarnegie Mellon UniversityDepartment of Architecture and Computer Technology, University of Granada

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Recently, a new paradigm in ICA emerged, that of finding “clusters” of dependent components. This striking philosophy found its implementation in two new ICA algorithms: tree–dependent and topographic ICA. Applied to fMRI, this leads to 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 based on (1) correlation and associated time–courses and (2) ROC study. It can be seen that 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 both FastICA and tree–dependent ICA (KGV) using as an approximation of the mutual information the kernel generalized variance.