I got a reply from the author of the longitudinal pipeline!
Hi Robin,
the long-processed images are all in the same space as the base image. This is -kind of- in the middle between your two time points (e.g. if the head is tilted left in TP1 and right in TP2 it will be upright) .
So if you want to use that mask in your fMRI analysis you need to register the fMRI with images from the base dir or one of the longitudinal time points instead of with the original T1 image .
Best, Martin
So yes, it all makes sense! Your fix fixed the normal recon-all. And the output ROI's in the .long dirs are in "tempalte space" and should not overlap with the orig input data! We will use either the T1.nii.gz or SurfVol from the long.TP2 directory as SSWarper input!
Edited 2 time(s). Last edit at 09/03/2021 05:35AM by Robin.