LINEAR AND KERNEL FISHER DISCRIMINANT ANALYSIS FOR STUDYING DIFFUSION TENSOR IMAGES IN SCHIZOPHRENIA

A new method is explored to study schizophrenia using diffusion tensor imaging (DTI). Both linear discriminant analysis (LDA) and kernel fisher discriminant analysis (KFDA) are combined with principal components analysis (PCA). Thus, a linear and non-linear combination of voxels is sought that separates patients from controls. PCA/KFDA does not show an improvement over PCA/LDA in classification. Because the PCA/LDA-mapping can be visualized, which enables localisation of differences, this method is preferred for analysis.