search

nimh-dsst / abcd-fasttrack2bids

0 stars 0 forks source link

In pipeline.py, log when a "corrupt volume" is removed from any DICOM series #8

Open ericearl opened 2 weeks ago

ericearl commented 2 weeks ago

It could be something like: sub-*/ses-*/func/sub-*_ses-*_*_bold_warning.txt

ericearl commented 1 week ago

Maybe this message for inside the text file's text:

The first 3D volume (60 slices) in this 4D timeseries was removed prior to Dcm2Bids/dcm2niix DICOM to NIfTI conversion due to the presence of "Raw Data Storage" instead of "MR Image Storage" in the first slice's Media Storage SOP Class DICOM field (0002,0002). This 4D volume now has one less 3D volume than expected and this missing timepoint/frame/repetition should be accounted for during analysis.

If you would like more information, you can read the report made to dcm2niix: https://github.com/rordenlab/dcm2niix/issues/830

ericearl commented 1 week ago

There is also the argument to not put it in the directory alongside the EPI image, but instead in the scans.tsv at the root level as a new column whose data dictionary description is "An integer count of how many corrupt 3D volumes were removed to create the 4D timeseries". Maybe corrupt_volume_removal? It would typically be zero, but sometimes one, and hopefully never more than one.

ericearl commented 5 days ago

Text here to be inserted into the BIDS README.

ericearl commented 4 days ago

This issue is undergoing testing since commit d972d96 and the README was updated in commit 40bb14b.