Hello experts, I have been trying to use the HCP Diffusion preprocessing pipeline for my own diffusion data. However, I run into a lot of issues at the final stage of the pipeline where the dwi data is registered to the T1w_acpc space (the freesurfer part of the pipeline runs fine on the structural images, the eddy correction also seems to run fine). So far I have not had a single subject where this alignment went smoothly and was accurate. It seems to be a problem with flirt. In the DiffusiontoStructural.sh subscript there are several calls to flirt (or epi_ref_dof) with the T1 as reference and for some reason flirt gives very wrong results like AP inverted or cerebellum to frontal lobe matching etc... I have managed to solve some cases with a call to only flirt and not epi_reg_dof and adding -useqform to the flirt options, but some I only managed by doing a rough manual alignment first. When the initial reg to T1 in ok-ish the bbregister and later parts all work very well. I don't think my dwi data is very special (1.5 mm^3 90 dir 2 shell), so I have no idea what might be the issue here or how to solve it (in a consistent way).
-Maarten _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
