ADNI - MPRAGEMETA.csv mock generator script

[Andrew Hannum]

Hi all, I thought this might help some of you depending on your situation.

    https://gist.github.com/kwertyops/1673ff8a2193530aa3fc6fc643e0af80

It is not a long-term solution, but by downloading the "MRI Images" table from the IDA ARC Builder, I've written a script that attempts to mock the data that should be in MPRAGEMETA.csv and creates a modified MRIQUALITY.csv to match the mocked data.

I'm supplying this with no guarantees of efficacy or accuracy, but my experience was that the MPRAGEMETA generated by this script was able to get me what looks so far like a valid BIDs directory with T1 images and fMRI images, using a small subset of ADNI3 data, with a few caveats:

    1. I had to remove one subject for whom clinica could not find a preferred T1 scan (subject number ending in 0031) for my sessions.

    2. Clinica errors out after setting up directories and converting images, seemingly while trying to create or update participants.tsv due to other missing files. After running clinica with "-m T1" and "-m fMRI" I still end up with participants.tsv that looks okay to me, so this isn't an issue for me.

    3. I don't care about most of the metadata for my purposes, and I can't verify how correct it is, especially in terms of "visit" descriptions or scan quality.

If you're going to run clinica with fMRI images you'll also need to download the "fMRI Images" table and update it with the following to create MRILIST.csv, which is even worse of a hack than the MPRAGEMETA file in terms of field values, but again seems to work for my use case where the much of the metadata is unimportant:

    https://gist.github.com/kwertyops/1673ff8a2193530aa3fc6fc643e0af80?permalink_comment_id=5571669#gistcomment-5571669