This study addresses the goal of reducing
site-variance in the correlation values of multisite rs-fMRI data by correcting
for site-specific EPI-related geometric distortions. Three geometric distortion
correction methods (B0 fieldmap and image-based unwarping) were applied separately
to a calibration dataset consisting of six subjects traveling to three sites.
The change of the correlation values and their site-homogeneity was evaluated
in dependency of the quantified geometric distortion correction values for six
networks and all subjects. It could be shown that the distortion correction had
no significant impact on the correlation values and no correlation between the
two parameters was found.