Low Frequency Analysis of Resting-State Fmri

Lianghua He,Hongfei Ji,Meng Wan,Shuang Liu
DOI: https://doi.org/10.1109/csip.2012.6308859
2012-01-01
Abstract:Resting state functional magnetic resonance imaging (rsfMRI) is a relatively new and an powerful method for evaluating regional interactions that occur when a participant is not performing an explicit task. Because of sensitiveness to time delay and length of time courses, respectively, traditional methods of correlation and coherence based on region of interest are not very suitable for rsfMRI. In this paper, we analyzed low frequency property of rsfMRI data. Firstly, we calculated the low frequency of resting-state Blood-Oxygenation-Level-Dependent(BOLD) signal on every voxel using Welch method, based on which we analyzed the statistical distribution cross subjects and voxels. The low frequency feature is proved by a series of experiments to be more robust and consistent comparing with coorelation in time domain and coherency in frequency domain. We found that the property of low frequency statistical distribution is dominant on narrow band, which is around 0.0137HZ.
What problem does this paper attempt to address?