There is a bug in the FSFAST https://surfer.nmr.mgh.harvard.edu/fswiki/FsFast functional connectivity seed-based analysis (current version and all previous versions). The amplitude of the seed waveforms are not rescaled in any way. This can affect the amplitude of the regression coefficient with the rest of the image (which is rescaled). This will not have any effect under several circumstances: (1) it will have no impact on the sig.nii.gz (p-value) or the Pearson correlation (pcc) at the individual level, (2) the seed uses a pca/svd, (3) if the intensity scale in each image/subject is the same (or close to it), or (4) if you are passing up the Pearson correlation (pcc) to the higher level. Note that the typical FSFAST analysis passes up the contrast effect size (ces) to the higher level which will be affected by this bug. The problem will manifest itself as extra variability at the group level due to the variability of the mean intensity at the individual level (so your group results should probably improve when fixed). This has been fixed in the dev version of FS and will be in the next release. The change will require that you specify the type of rescaling you want. If you run the new fcseed-sess on an old configuration that does not have a scaling method specified, it will exit with a error message (this can be overridden by adding -no-require-rescale-type).
Please reach out with any questions and sorry about the inconvenience.
doug