There isn’t currently a data-field that directly indicates whether a T1 structural scan passed or failed QC- however, it should still be fairly straightforward to identify the failed T1 scans!
One way to do this is by looking at what is inside the zip file. For T1 scans that failed QC, the zip archive contains only the file T1_orig_defaced.nii.gz, and this is stored inside a subfolder called “unusable”. When unzipped, the structure looks like this:
T1/unusable/T1_orig_defaced.nii.gz
In contrast, T1 scans that passed QC contain many additional files produced by the processing pipeline, such as a “T1_fast” folder, a “T1_first folder”, a “T1_brain_mask.nii.gz” file, and various other outputs.
Because of this difference, the size of the zip file is also a useful indicator you could use to systematically search for failed T1s. Passed T1 scans are usually around 40–50 MiB, while failed scans tend to be much smaller (below 30 MiB).
For example, if you want an automated way to identify the failed T1 scans, you can use the dx toolkit to list all file IDs for data-field 20252 (using dx find data), and then use dx describe to obtain and filter for the file size for each archive. From there, you can get a list of the files that are smaller than about 30 MiB (i.e. failed QC T1 files) that ready for you to analyse.
Comments
1 comment
Hi Yumiko,
There isn’t currently a data-field that directly indicates whether a T1 structural scan passed or failed QC- however, it should still be fairly straightforward to identify the failed T1 scans!
One way to do this is by looking at what is inside the zip file. For T1 scans that failed QC, the zip archive contains only the file T1_orig_defaced.nii.gz, and this is stored inside a subfolder called “unusable”. When unzipped, the structure looks like this:
T1/unusable/T1_orig_defaced.nii.gz
In contrast, T1 scans that passed QC contain many additional files produced by the processing pipeline, such as a “T1_fast” folder, a “T1_first folder”, a “T1_brain_mask.nii.gz” file, and various other outputs.
Because of this difference, the size of the zip file is also a useful indicator you could use to systematically search for failed T1s. Passed T1 scans are usually around 40–50 MiB, while failed scans tend to be much smaller (below 30 MiB).
For example, if you want an automated way to identify the failed T1 scans, you can use the dx toolkit to list all file IDs for data-field 20252 (using dx find data), and then use dx describe to obtain and filter for the file size for each archive. From there, you can get a list of the files that are smaller than about 30 MiB (i.e. failed QC T1 files) that ready for you to analyse.
Please sign in to leave a comment.