MRIQC for phantom data?

Hi there,

I would like to modify MRIQC to process phantom data (for daily scanner QA).
Before trying to re-invent the wheel, has anybody worked on something similar?

Thanks!

-Pablo

This is done in this paper:

Cheng, C. P.; Halchenko, Y. O. A New Virtue of Phantom MRI Data: Explaining Variance in Human Participant Data. F1000Res 2020, 9, 1131. https://doi.org/10.12688/f1000research.24544.1.

Thanks for your answer, @jsein
That’s a very interesting paper! I bet these findings can be used in the future to model inter-subject variability.

My question, however, was about modifying the MRIQC workflow to make it more suitable for phantom data.
For example, if the phantom is spherical, the rotations that MRIQC will find in the motion correction workflow will be somewhat arbitrary, given the symmetry of the problem (if we leave aside the fact that agar phantoms always have some structure in them). Forcing a motion correction with just translations (and maybe shears) makes more sense to me.
Another example: the template MRIQC is the MNI152NLin2009cAsym. Again, the transformation that MRIQC finds will be meaningless. It might be more meaningful to use a sphere (and limit it to 3D translations), maybe with some pre-defined segmentation for the carpet plot (a.k.a., “Power plot”), so that it can be compared across dates.
The idea is to make the registrations as predictable as possible, so that the results can be compared across dates…