PreFreeSurfer fails with LegacyStyleData and hcp_t2=NONE in QuNEx 1.3.0

Hi,

I’m trying to preprocess a longitudinal dataset using QuNEx 1.3.0. Most subjects have both T1w and T2w images, but a small number of sessions are missing the T2w image. I would therefore like to process those sessions using LegacyStyleData.

For the sessions without T2w, I set the following in my batch parameter file:

_hcp_processing_mode : LegacyStyleData
_hcp_t2              : NONE

However, hcp_pre_freesurfer fails immediately with:

ERROR: missing the following mandatory parameter(s):
--t2template --t2templatebrain --t2template2mm
ABORTING

I checked the generated PreFreeSurferPipeline.sh call, and when hcp_t2=NONE, QuNEx does not pass any of the three T2 template arguments.

I also inspected process_hcp.py in the QuNEx 1.3.0 container and found that the T2 template variables appear to be explicitly set to empty strings when hcp_t2 == “NONE”:

if options["hcp_t2"] == "NONE":
    t2template = ""
elif options["hcp_prefs_t2template"] is None:
    t2template = os.path.join(
        hcp["hcp_Templates"],
        "MNI152_T2_%smm.nii.gz" % (options["hcp_prefs_template_res"]),
    )
else:
    t2template = options["hcp_prefs_t2template"]

The same logic is used for t2templatebrain and t2template2mm.

I confirmed that the corresponding template files do exist inside the container:


/opt/HCP/HCPpipelines/global/templates/MNI152_T2_1mm.nii.gz
/opt/HCP/HCPpipelines/global/templates/MNI152_T2_1mm_brain.nii.gz
/opt/HCP/HCPpipelines/global/templates/MNI152_T2_2mm.nii.gz

As a control, when I process a subject that does have T2w, using HCPStyleData and _hcp_t2 : t2, QuNEx correctly generates a call containing the T2 image and all three T2 template arguments, and PreFreeSurfer proceeds normally.

Is there a recommended way to run T1-only data with LegacyStyleData in QuNEx 1.3.0? In particular, is there another parameter I should be setting when using _hcp_t2 : NONE, or is the omission of the T2 template arguments in this case a bug?

I would prefer to keep the same QuNEx/container version across the entire longitudinal cohort, so I would especially appreciate a solution that allows these few T1-only sessions to be processed within QuNEx 1.3.0.

Thank you!

Hi!

I believe we already fixed this. Can you try running this with the latest QuNex - 1.5.1 and then report back your results.

In terms of merging 1.3.0 resuilts with 1.5.1 you are fine. We are super careful when it comes to backwards compatibiilty.

Best, Jure

Hi,

Any updates here. Did updating your QuNex version resolve your issue?

Best, Jure

still working on it… will update soon