Testing the effects of pre-processing on voxel based morphometry analysis

Voxel based morphometry (VBM) is an automated analysis technique which allows voxel-wise comparison of mainly grey-matter volumes between two magnetic resonance images (MRI). Two main analysis processes in VBM are possible. One is cross-sectional data analysis, where one group is compared with another to depict see the regions in the brain, which show changes in their grey-matter volume. Second is longitudinal data analysis, where MRIs, taken at different time points, are compared to see the regions in the brain that show changes in their grey matter volume for one time point with respect to another time point. Both types of analyses require pre-processing steps before performing the statistical analysis. In this study, we examined grey matter differences for patients with blepharospasmus (BFS) before and after treatment, at two different time points. The main evidence base therapy for this condition is the “botulinum toxin” injection in the respective muscles. The main aim of this study was to look at the effects of different pre-processing steps, namely, normalization and smoothing on the results of the longitudinal data analysis. A second aim was to analyze structural grey-matter differences before and after the treatment. Our results showed that the DARTEL normalization and the lower width for smoothing as preprocessing steps delivered pathophysiological plausible results. The longitudinal analysis revealed significant temporal differences after the injection of the botulinum toxin injection mainly in patients with BFS.