الفهرس | Only 14 pages are availabe for public view |
Abstract Mental disorders, especially schizophrenia, are still challenging to diagnose in early phases. Nowadays, computer-aided diagnosis techniques based on Resting-state functional Magnetic Resonance Imaging (Rs-fMRI) have been growingly developed to tackle this challenge. In this study, we investigate different combinations of features computed for each brain region in order to discriminate the schizophrenic from normal subjects.The set of features include the Regional Homogeneity (ReHo), Voxel-Mirrored Homotopic Connectivity (VMHC), fractional Amplitude of Low-Frequency Fluctuations (fALFF) and Amplitude of Low-Frequency Fluctuations (ALFF). Data denoising and preprocessing were first applied, followed by the feature extraction module.The extracted features were then reduced using the Principal Component Analysis (PCA) transformation, and the best discriminative features were selected using different feature selection algorithms such as the Fisher score and t-test methods. A Support Vector Machine (SVM) classifier was trained and tested on the COBRE dataset formed of 70 schizophrenic and 70 healthy subjects.The highest average classification accuracy of 98.57% has been achieved |