Hi all!
I’m noticing two issues when running the hcp_fmri_volume processing step. In some subjects, including HCP example subject 100307, I’m seeing the following error in the log file:
START: OneStepResampling
(standard_in) 1: syntax error
(standard_in) 1: syntax error
WARNING: sampling 0 less than 0.00001 mm
Setting to 1.0 sampling instead
Any idea what this refers to? Could this be the result of some parameter being set incorrectly?
Not clear if causal, but in the same subjects I’m also noticing that processing aborts with exit code: 137, which indicates a memory error. This seems paradoxical to me (running this on 64G RAM, Docker RAM is set to 64G, and doesn’t seem to max out during the run). Have you seen this before?
Thanks again for all the support!
batch.txt (5.5 KB)
error_hcp_fmri_volume_1_Garbarini_Logan_2025-04-03_19.59.20.103392.txt (159.1 KB)
All the best,
Dustin
Hi Dustin,
Sorry for a late reply, somehow I did not see this topic. Are there any updates here? I have never seen this, I imagine this is 100307 from HCYA? We and several other users successfully processed HCYA data and had no issues. Can you maybe paste here all the commands you ran before hcp_fmri_volume. Thanks!
Best, Jure
Hi Jure,
No problem, I’m just revisiting the issue. I’ve now re-run the pipeline on one of my subjects, and while I see the same error in hcp_fmri_volume, it does finish successfully. Attaching all log files below. Seems challenging to diagnose because I can’t see any other errors in the preceding processing steps.
batch2.txt (5.3 KB)
done_hcp_fmri_volume_1_LG_2025-04-10_17.21.51.103642.log (164.0 KB)
done_hcp_freesurfer_LG_2025-04-03_23.23.39.839798.log (592.1 KB)
done_hcp_post_freesurfer_LG_2025-04-07_18.39.53.878806.log (1.8 MB)
done_hcp_pre_freesurfer_LG_2025-04-03_21.32.35.615533.log (25.5 KB)
Thanks!
Dustin
Hi Dustin,
I see the issue, it seems like your fMRI resolution parameter is not set correctly: --fmrires="True". This should be by default and according to the batch file set to 2. The relevant parameter here is hcp_bold_res. Can you provide the full command call, the parameter is set correctly in batch, it seems like you are overwriting it somehow. It finishes cine I assume that True gets converted to 1. The errors are because the code is trying to do some math with True.
Best, Jure
Thanks Jure, good catch. I think I may have found the issue, or at least part of it. This was my call:
qunex_container hcp_fmri_volume --batchfile=“${STUDY_FOLDER}/processing/batch2.txt” --sessionsfolder=“${STUDY_FOLDER}/sessions” --parsessions=“1” --parelements=“1” --overwrite=“yes” --container=“${QUNEX_CONTAINER}” --hcp_processing_mode=‘LegacyStyleData’ --hcp_bold_sbref=“NONE” --hcp_bold_res =2 --containeropt=“–rm --name ${RUNNER} ${QUNEX_CONTAINER}”
It seems hcp_bold_res may have been incorrectly applied due to the extra space. Unsure why that would set the parameter to “True” but I’ll try to re-run it with --hcp_bold_res=“2”.
Thanks again, really appreciate the support!
If there is no value after the parameter, the parameter is treated as a flag (set/unset, true/false). For example --overwrite will set overwrite to True.
Best, Jure