Closed nimzodisaster closed 1 year ago
It occurred to me this morning that the 1D file: denoised_func_data_nonaggr_0_trans_tcat_flirt_roistat.1D contains lines without any data (probably)
for example:
#/data/proj/app/pipeline/rs/outputs/working/cpac_103430-100_3_ses-1/nuisance_regressors_default_271/_scan_func-1/Functional_2mm_flirt/denoised_func_data_nonaggr_0_trans_tcat_flirt.nii.gz 299[?]
so the error probably occurred earlier, perhaps when denoised_func_data_nonaggr_0_trans_tcat_flirt_roistat.1D was created.
This may be related to bad image registration to template space, resulting in very little overlap with priors.
closed: user error.
After fixing my own error that prevented good registration, I thought perhaps this bug went away. However, the bug still appears, though not as frequently.
My best guess right now is this: Functional dropout or bad registration in a subject causes some ROI masks to have no timeseries data crashing the node.
Hi, thanks for reaching out! We'd be glad to help. As a first quick check, could you take a look at your CSF and WM masks from segmentation? These are both being used in your pipeline config as nuisance regressors and it is possible that the masks are too stringent, which could lead to the timeseries data not making it through.
Describe the bug
Build_nuisance_regressor crash: number of time points is inconsistent
looking at the the two files, there is indeed the same number of data points in each.
last line of the 1D file:
#/data/proj/app/pipeline/rs/outputs/working/cpac_103430-100_3_ses-1/nuisance_regressors_default_271/_scan_func-1/Functional_2mm_flirt/denoised_func_data_nonaggr_0_trans_tcat_flirt.nii.gz 299[?]
which is indexed from 0, indicating 300.and the relevant image dimensions: fslinfo denoised_func_data_nonaggr_0_trans_tcat.nii.gz
This crashed on every participant I ran. Here is a list of missing outputs:
To reproduce
No response
Preconfig
Custom pipeline configuration
Run command
run /data/proj/app/pipeline/rs/dat/SITES/PRIMARYSITES/MIND /data/proj/app/pipeline/rs/outputs participant --pipeline_file /data/proj/app/pipeline/rs/Pipeline_M.yml --data_config_file /data/proj/app/pipeline/rs/subject_configs/data_config_T123.yml
Expected behavior
process the nuisance regression and continue
Acceptance criteria
no crash
Screenshots
No response
C-PAC version
v1.8.4
Container platform
No response
Docker and/or Singularity version(s)
No response
Additional context
There is always a possibility that our subject id names are problematic, which follow the format 123456-100 123456-100_1