Matching HBCD processing pipeline on own data

[nbustos]

Hi, I am trying to replicate HBCD processing on my own fmri/diffusion data.

Using parameters/flags based off this site - https://hbcd-cbrain-processing.readthedocs.io/latest/tools/nibabies.html

1st Question: Should I use v25.0.1 or 25.0.0rc1 as used above ?
(why was an rc used for an official data release...?)

2nd Question: Getting MCRIBS error for not having precomputed T2 seg ?
(isn't FSL-FAST ran regaurdless?)
(If I am supposed to have a precomputed T2seg before nibabies when using MCRIBS, then how do I match HBCD processing, there is no documentation of that procedure (ie - iBEAT or DrawEM) in above docs)

I am getting an error right away:

Traceback (most recent call last):
File "/opt/conda/envs/nibabies/bin/nibabies", line 8, in
sys.exit(main())
^^^^^^
File "/opt/conda/envs/nibabies/lib/python3.11/site-packages/nibabies/cli/run.py", line 62, in main
retval = build_workflow(config_file)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/nibabies/lib/python3.11/site-packages/nibabies/cli/workflow.py", line 76, in build_workflow
retval['workflow'] = init_nibabies_wf(config.execution.unique_labels)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/nibabies/lib/python3.11/site-packages/nibabies/workflows/base.py", line 152, in init_nibabies_wf
single_subject_wf = init_single_subject_wf(
^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/nibabies/lib/python3.11/site-packages/nibabies/workflows/base.py", line 458, in init_single_subject_wf
anat_fit_wf = fit_wf(**wf_args)
^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/nibabies/lib/python3.11/site-packages/nibabies/workflows/anatomical/fit.py", line 1034, in init_infant_anat_fit_wf
surface_recon_wf = init_mcribs_surface_recon_wf(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/conda/envs/nibabies/lib/python3.11/site-packages/nibabies/workflows/anatomical/surfaces.py", line 63, in init_mcribs_surface_recon_wf
raise NotImplementedError(

Thanks
NB

[mgxd]

Hi Nico,

Should I use v25.0.1 or 25.0.0rc1 as used above ?

I'm surprised the RC was used, though not much was changed from the official release. I would recommend using the latest in the release series (currently 25.0.2) - these will be bug-fix releases

Getting MCRIBS error for not having precomputed T2 seg

Our implementation of MCRIBS requires finely accurate subcortical segmentations, rather than just tissue classification. I believe HBCD achieves this by first running BIBSnet, potentially followed by manual intervention if needed. Pinging @erikglee as he knows better. To feed these in, you can use the --derivatives flag - see the documentation for more detail.

[erikglee]

Hi Both,

Like @mgxd mentioned, for the first release we gave Nibabies access to manually corrected BIBSNet segmentations.

The full Nibabies settings we used for the release can be seen here: https://hbcd-cbrain-processing.readthedocs.io/release_1.0/tools/nibabies.html. Some of the details cover how we choose which files to include during processing, and other details are more specific to Boutiques Descriptors+CBRAIN (both of which may not be of interest). But you may find the "Other Processing Settings" section helpful, since it describes the flags that we used.

[nbustos]

Ah ok - thank you very much! - I didnt catch it on the "external requirements".

Is there any recommended procedure (HBCD style) for the manual correction on BIBSNet ?

I suppose the manual intervention is part of why we cant add BIBSNet into the pipeline for hbcd nibabies.

[erikglee]

That is a good question! @LuciMoore @ericfeczko do either of you have documentation/references for folks wanting to do 'HBCD'/'BIBSNet' style manual segmentation corrections?

[LuciMoore]

yup the SOP for manual corrections can be found in the supplementary material for this paper : https://www.biorxiv.org/content/10.1101/2024.10.02.616147v2.supplementary-material

[nbustos]

uu awesome thanks!