BIG thank you. Will read the course document and get back if necessary.
nick
Niklas Lenfeldt, PhD, MSc, MAJ SWE A
Clinical Neuroscience, Umeå university
Phone: +46 70 609 66 87
Från: FSL - FMRIB's Software Library [mailto:[log in to unmask]]
För Mark Jenkinson
Skickat: den 15 september 2016 09:41
Till: [log in to unmask]
Ämne: Re: [FSL] registration
Hi,
If you have used topup and corrected for susceptibility distortions then FLIRT to the individual T1 structural should work well with 6 DOF. I recommend using the script epi_reg to do this, as
it gives very accurate results for EPI to structural registration. Just take your distortion-corrected b0 image and register this (without any fieldmaps) to the structural using epi_reg.
All your other images (FA, MD, etc.) will be in the same space as the b0, so there is nothing else you need to do there except invert the transformation matrix to get the structural-to-diffusion
space transformation. Then you can happily transform images between diffusion and structural space (using the transformation matrix from epi_reg or its inverse).
As for using an image in MNI space, for this you should use FNIRT to get a non-linear registration between your individual T1 structural and the MNI space. You can then use invwarp to get the
warp from MNI to structural and from this, combined with the structural to diffusion matrix above, you can transform your ROI from MNI to structural space or from MNI to diffusion space (using applywarp).
I hope this helps.
If you are unsure of some of the details have a look at the second part of the FSL course registration practical, as that goes through most of these steps.
All the best,
Mark
From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of Niklas Lenfeldt <[log in to unmask]>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
Date: Tuesday, 13 September 2016 at 11:22
To: "[log in to unmask]" <[log in to unmask]>
Subject: [FSL] registration
Hi
I intend to perform ROI-analyses of FA, MD, RD and AD images, but I would like to do the outlining in a T1-image of the patient that also was sampled with higher resolution. All DTI-images have been preprocesses (topup, eddy). Should the registration be performed before dtifit or separately to each measure after dtifit? I also wonder if FLIRT is enough to use (it is the same brain) or should FNIRT be applied?
It is old patients with some atrophy, so I would like to keep it on an individual level, hence no registration on group level. However, I am thinking that it would be beneficiary to do the outlining in the MNI-template to which the T1 can been registered, but how do I get the DTI-image into the same space (so I can outline in T1-MNI and transfer ROIs to FA/MD/RD/AD-MNI)? Or is it something I do not get here?
Best,
nick
Niklas Lenfeldt, PhD, MSc, MAJ SWE A
Clinical Neuroscience, Umeå university
Phone: +46 70 609 66 87