ABCD - GE data and phase encoding

#1

Hey all,
I was hoping to get some experience from people who have worked on the raw ABCD “dicom” data. I see there have been a couple of attempts of getting ABCD into BIDS format. I ran into a recent problem specifically with the GE data. Using dcm2niix (https://github.com/rordenlab/dcm2niix) I ended up with mislabeled anterior and posterior field maps

. The phase encodings are all labeled “j” in the associated JSON despite some of being A->P and other P->A. Has anyone else noticed this or been able to recover the appropriate information for the GE data

#2

Is there a reason you can’t use the minimally preprocessed data

#3

This is not a limitation of dcm2niix, rather of the ABCD DICOM data. In general, dcm2niix is able to use the methods described here to decode various vendor specific details for GE DICOM images. However, the ABCD DICOM images were not directly created by GE, but rather re-created by the ABCD team. I think this may be because multi-band was not supported in GE DICOMs at the time they started the ABCD data acquisition.

For more details, see here.

#4

I struggled with this same thing about 6 months ago, which is why I suggest you just use the minimally preprocessed data. Otherwise you have to go through all of the mundane data cleaning.

#5

Hey Chris,
I suspected this to be the case but hoped for a easily automatible work-around. Additionally any insights as to whether there is only an Anterior-Posterior switch or may additionally be a left right switch would be appreciated

#6

Thanks Tokazama. It’s nice to know I’m not the only one. I can use the mproc data but was hoping to get it into bids format and use fmriprep for preprocessing. Automatible data cleaning is fine, and understanding that this is a GE specific issue is helpful but I still don’t know where to go from here since the phase encoding is also missing from the GE file names.

#7

The phase encoding is what got me too.