Can Nipype generate a registration outline/report for FLIRT and FNIRT?

1

Hi all,

I’m trying to find out if I can generate registration reports from coregistration and spatial normalization using FSL (FLIRT and FNIRT) with Nipype (for example, shown here) I can’t seem to find any information on that, but I may have missed something. Is it possible? Or, if it’s not possible/available, are there other alternatives that I can use to check my registrations in my Nipype pipeline?

Thanks!

2

This as far as I know is not build into nipype directly, but the niworkflows package provides FLIRTRPT Interface that provides animated coregistration reports (see https://github.com/poldracklab/niworkflows/blob/master/niworkflows/interfaces/registration.py#L160). Those have been used in MRIQC and FMRIPREP. See http://fmriprep.readthedocs.io/en/stable/workflows.html for example images.

3

Thanks for point that out Chris. Currently, is there any way to quickly integrate FLIRTRPT into an existing nipype workflow already? Or would this still require swapping out nipype to niworkflow’s nipype (based on what I read here)?

Regardless, I did end up finding an alternative solution using Nilearn’s plotting tools (and posting here for future reference). Using nilearn.plotting.plot_anat, and overlaying edges from an MNI template (or whatever the target image may be), it replicates the FLIRT report quite nicely. Some sample code:

nilearn.plotting.plot_anat(anatomical_image, draw_cross=False).add_edges(mni_template)
4

Currently you would need to swap nipype with niworkflows.nipype.

FLIRTRPT also uses nilearn internally.