Preprocessing effects on group independent component analysis of fMRI data

Abstract

Functional connectivity networks (FCN) might differ due to tissue loss in brain. Spatially independent components can be gathered with the group independent component analysis, which is one of the methods that can extract FCNs from fMRI data. Comparison of spatial or temporal results of group-wise data is possible for the differences in parameters of the preprocessing or processing steps. In this study, after the fMRI data, which is taken from Alzheimer's disease and mild cognitive impairment patients during an oddball paradigm, is preprocessed by two different methods; a group independent component analysis is done. Both of the preprocessing methods include slice time correction, motion correction, coregistration, normalization and spatial smoothing while they differ in normalization step as the chosen algorithm varies. After the preprocessing, group independent component analysis is applied with the same parameters for both of the methods. As a consequence, the effect of the difference between the two preprocessing methods are investigated. Depending on the results, stability, power spectrums and spatial maps of the components show an alteration by the algorithm used in normalization step.