Brain fMRI processing and classification based on combination of PCA and SVM
Songyun Xie, Rong Chun Guo, Ningfei Li, Ge Wang, Haitao Zhao · 2009
fMRI is one of the fundamental tools for functional human brain research. However, fMRI data are often in a high dimensional feature space and suffer greatly from large and complex dataset. To relieve the curse of dimensionality in fMRI image, PCA combines with SVM to form a feature-based classification method in this work. PCA is employed to find a more compact and reasonable representation of the data by extracting features from each fMRI image. Then a linear kernel SVM classifier is trained on the selected features to detect different brain states. The advantage of incorporating PCA with SVM is twofold: Firstly, the computational burden on SVM classifier is reduced significantly. Secondly, a less complex classifier is well established. Experimental results show that the proposed method yields good performance. The correct rate of our hand-movement fMRI study with both healthy subjects and a tumor patient verified the stability and generalization capability of the method.