Detecting fMRI activation by meanshift clustering method based on voxel neighborhood information
(1. Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences, Suzhou, Jiangsu 215163, China; 2.Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin 130000, China; 3.University of Chinese Academy of Sciences, Beijing 100049, China; 4.Department of Radiology, University of Iowa, Iowa City, 52240, USA; 5.Department of Neurology, The Second Affiliated Hospital of Soochow University, Suzhou, Jiangsu 215004, China)
Abstract：To improve the antinoise capacity and precision of fMRI activation area detecting method, a meanshift clustering method was proposed based on voxel neighborhood (VNMSC). A twodimensional feature space was constructed for VNMSC based on temporal properties of a voxel and its neighbors. The correlation coefficient between MRI timeseries and stimulation response function of each voxel was calculated by crosscorrelation analysis. The correlation coefficient between timeseries of a voxel and the neighboring voxels was also calculated by the same method. Based on the feature space, a meanshift clustering was adopted to detect active region of fMRI to obtain simulated and real fMRI data. The VNMSC method was tested by simulation data and practical fMRI data. The results show that the sensitivity and specificity of MSCVN technique are better than those of the traditional crosscorrelation analysis (CCA) and CCA plus cluster analysis in any activation area size when the kernel size is appropriate. The results of real fMRI data demonstrate that MSCVN and other two methods have good consistency in accuracy, while the detection region of the proposed method is more complete.
2016, Vol. 37 
