Voxel Selection Framework with Feature Extraction for Classification of Brain Activity in fMRI
Abstract
This paper presents fMRI (functional Magnetic Resonance Imaging) signal analysis methodology using Principal Component Analysis (PCA) and Mutual Information (MI) based voxel selection framework. Previously, the fMRI signal analysis has been carried out either using Principal Component Analysis (PCA) model or voxel selection on raw fMRI signal. The first methodology does feature extraction that makes voxel selection process easy while the latter methodology does selection of relevant voxels (or features). Both these advantages are added by our methodology in which Principal Component Analysis (PCA) is used for feature extraction to decrease the dimension of fMRI data. The proposed methodologies are adopted for classification of brain activity. Experimentations are carried out in the publicly available fMRI dataset of six subjects and comparisons are made with existing PCA model and voxel selection framework. The superiority of the proposed methodology gets validated by the comparative results.
Keywords— fMRI, voxel, decomposition, PCA, MI.