Support Vector Machines for neuroimage analysis: Interpretation from discrimination

Authors

Luping Zhou, University of WollongongFollow
Lei Wang, University of WollongongFollow
Lingqiao Liu, Australian National University, University of Adelaide
Philip O. Ogunbona, University of WollongongFollow
Dinggang Shen, University of North CarolinaFollow

RIS ID

86792

Publication Details

Zhou, L., Wang, L., Liu, L., Ogunbona, P. O. & Shen, D. (2014). Support Vector Machines for neuroimage analysis: Interpretation from discrimination. Support Vector Machines Applications (pp. 191-220). Germany: Springer.

Abstract

Support vector machines (SVMs) have been widely used in neuroimage analysis as an effective multivariate analysis tool for group comparison. As neuroimage analysis is often an exploratory research, it is an important issue to characterize the group difference captured by SVM with anatomically interpretable patterns, which provides insights into the unknown mechanism of the brain. In this chapter, SVM-based methods and pplications are introduced for neuroimage analyis from this point of view. The discriminative patterns are decoded from SVMs through distinctive feature selection, SVM decision boundary interpretation, and discriminative learning of generative models.