Hi all,
I am trying to do the native surface to template (dHCPSym_32k, 40 week) using the recommend script https://github.com/ecr05/dHCP_template_alignment. But it seems the align_to_template_3rd_release.sh do not suitable for the 3rd release anat pipeline data. The file structure not seems compatible and the nomenclature of some metric files also not identical.
So I slightly modified the script to successfully align the native surfaces to template surface but I found the output template space medial wall label file have some kind of variation across individuals which is not the case in HCP data, so I wondering is this due to the improper running of the script or is normal situation of the current state. And any suggestions to solve this mismatch problem?
Hi @YaoMeng,
There are a few subjects where the distortion around the medial wall is large (like in sub-CC0005000X01) because we reduced the regularisation term of the MSM config file for the third release, to better align the frontal lobe sulcal morphology, and as part of the symmetrisation of the template. This was a choice that we discussed at the time, and felt that the alignment improvements we saw outweighed the subjects where the regularization was too low.
You could run MSM registration to the template using a higher lambda in the config file, although it might impact the symmetry of the left and right hemispheres if that was what you are interested in.
If you want to mask the medial wall with a medial wall ROI that is consistent across subjects, I have left and right medial wall mask for the 40 week dhcpSym space that you can use if you give me your email. I will make sure that these also exist as part of the left-right symmetric template (I havent got around to adding them yet).
Can you also tell me where the naming in the script differed from the dHCP 3rd release names so I can address it?
Does that answer everything @YaoMeng?
Logan
Hi Logan,
Yes, your answer covers all my questions. Huge thanks to your help, I will message you my email.
And the naming problem mainly about the ‘/Native’ folder, the 3rd release did not have this. And the row 193 metric term should be “curv” instead of “curvature”.
Yao
Okay perfect, thanks for flagging that for me. Yes, the directory structure for the third release is slightly different, but must have missed the ‘/Native’ directory part and the curvature naming.
Best wishes,
Logan
Hi,
I am just learning about this procedure and I am trying to use dHCP_template_alignment for structural data of second release of dHCP dataset but I just confused about template volume variable. Can you help me about it ?