Multisubject fuzzy cluster analysis of functional MRI
Jahanian H, Hossein-Zadeh G, Soltanian-Zadeh H, Shams S
Image Analysis Lab. Radiology Dept., Henry Ford Health System, Control and Intelligent Processing Center of Excellence, University of Tehran
In fMRI multisubject analysis, some methods suppose specific assumptions on the signal and noise structure which may not hold for fMRI data. A novel method, based on fuzzy cluster analysis is introduced to address this issue. In the proposed method a vector is obtained for each voxel using the cross-correlation analysis of its corresponding time-series. Then, a feature space is derived from the obtained vectors. Fuzzy cluster analysis of the proposed feature space generates the activation map. The proposed method is evaluated using an experimental visual task dataset (consisting of 11 subjects). The results show that the proposed method improves the activation detection sensitivity.