EPInorm to replace EPI->T1w->MNI registration

[effigies]

http://onlinelibrary.wiley.com/doi/10.1002/hbm.23737/full

Supersedes #253, and may make ultimately replace SyN-based susceptibility distortion correction (#496/#544).

[chrisgorgo]

When you have read the paper carefully could you list the pros, cons, and risks (things to look out for) of replacing syn based SDC with EPInorm?

[chrisgorgo]

It might be also interesting to think about this in terms of a new optional target template EPIMNI.

[oesteban]

That target would not be the same as MNI but specifying a resolution?

[chrisgorgo]

The original paper used an EPI template instead of a T2w one (which we would have to use if we wanted to stick to 2009c) https://twitter.com/vdcalhoun/status/890217862010355713

[effigies]

Looking at the EPI/T2 templates in SPM12, these at least are very similar; EPI.nii basically looks like a skull-stripped version of T2.nii. I haven't looked at the 2009c ones, but at a first pass, I'd think skull-stripping the T2w template and rescaling to match the EPI (as we do in creating the inverse T1w image) would get us to a similar starting point as EPInorm.

I'd think an EPIMNI target would just be the same as MNI... I'm not sure what you have in mind, unless it's to say don't bother resampling the anatomical to the MNI space. (Which, I'd think we wouldn't bother doing if we use EPInorm, unless we want to have a warp back to subject-native space.)

I feel like a major test of the (EPI -> MNI -> T1w) pipeline replacing SyN-SDC would be surface-sampling.

I'll give the paper a close reading and share my notes.

[chrisgorgo]

I see - thanks for doing this comparison.

[chrisgorgo]

Related: https://neurostars.org/t/severe-epi-distortion-by-using-fmriprep/1185

[chrisgorgo]

Related https://www.frontiersin.org/articles/10.3389/fnins.2018.00064/full