First, do you have \ at the end of each line for the mrconvert call (as described here)? Just want to make sure that it’s actually detecting the -fslgrad, -json_import, and other options because without the \, it wouldn’t know that it’s part of the same line.
Then, if you verified the above, did you look inside the *.json file? Is there a field for “PhaseEncodingDirection”? If not, then it may be that:
- the data has been anonymized in such a way that’s stripped the phase encoding direction info from the DICOMs
- the scanner itself (if it’s a Philips scanner) does not give PhaseEncodingDirection, but only PhaseEncodingAxis
If you have access to the original DICOMs, maybe you can try using that as direct input into your pipeline to also rule out that the conversion process didn’t fail to retain the phase encoding direction info.
You can read about phase encoding here: