Independent Component Analysis Of Resting State FMRI Data

Resting-state functional magnetic resonance imaging (RS-fMRI), is an emerging technology in the research of brain function. In variety of methods in RS-fMRI analysis, independent component analysis (ICA) occupies an important position because of its unique characteristics of Blind Source Separation (BSS).This paper studied the application of the ICA in the RS-fMRI analysis from three aspects as follows.Firstly, the application of ICA in the noise extraction of resting state fMRI was analyzed. Experimental results showed that the ICA could effectively extract and remove the noise which related to the physical activity such as eye movements, breathing, heartbeat, head movement, and so on.Secondly, the influences of different ICA strategies upon individual brain function network spatial maps were evaluated. When comparing the spatial maps acquired by individual-based ICA to those acquired by group ICA, it was found that more satisfying spatial maps could be acquired using group ICA. The number of independent components affects the obtained spatial maps, especially when this number was set to a small value relative to the sample size of the subject group.Finally, the brain basis of individual differences in handedness was analyzed based on resting state fMRI. It was found that handedness scores are significantly correlated with coactivations within such networks as the primary motor network and attention network.To summarize, ICA exhibited its unique virtue in resting state fMRI data analysis. ICA can not only extract multiple networks simultaneously, but also effectively extract components corresponding to noises and thus facilitate de-noising. In addition, attention should be paid upon the influences of ICA analysis strategies when performing fMRI data analysis based on ICA. The present study also demonstrated that ICA analysis of resting state fMRI data can be effectively used in the brain function basis of typically developing human subjects.