Skip to content

Latest commit

 

History

History
62 lines (36 loc) · 2.68 KB

07-alignments.md

File metadata and controls

62 lines (36 loc) · 2.68 KB
Raw

Alignments

First

  • copy the surfRelax (anatomy) folder into the place that contains the demo dataset (usually this would be in a place determined by the freesurfer install / $SUBJECTS_DIR)

  • cd into our demo data directory 20200721_demo

  • from within the folder start the interactive alignment tool mrAlign

mrAlign()

Plan

  • Destination - the image you want to align to - often the best base/end point is the file produced by freesurfer and converted with mlrImportFreesurfer() - see below

  • Source - the image you want to move / align - usually the anatomy scan in the convention of the data .. in our case the "fake" anatomy file made by averaging the first fMRI scan across time.

  • help mrAlign by moving the source image to roughly the right place (by changing translation params or initialising from header)

  • make sure you check what contrast settings you are using (in our case we are trying to align a T2* weighted image to a T1, so using "reverse" contrast)

  • run a coarse alignment follwed by

  • a fine alignment

  • check that everything worked

  • save and export to mrLoadRet-4.5 to update nifti image and mrLoadRet data structures (this step is important to make sure that the data structures in mrLoadRet folder are updated, not just the qform and sform fields in the nifti file headers)

  • in mrLoadRet, load in the anatomy file and use to get better visualisations

Using freesurfer segmenations

To make things a bit quicker here, I have already run a freesurfer segmentation on the whole head MPRAGE for this participant.

At UoN, the convention is that participants get a unique 5 digit ID (^\d{5}). Here, I will use the made-up ID 55555.

recon-all -subjid 55555 -i dafni_01_FSL_7_1.nii -all

There is good integration of freesurfer segmentations and surfaces into mrTools - for background and detailed explanations see the flatmap & surface tutorial

To get everything into the correct space for mrTools to work there is also a step using mlrImportFreesurfer(), which has already been done here:

% where my freesurfer seg ended up
cd /data/anatomy/freesurfer/subjects/55555
mlrImportFreesurfer() % accept suggestions

% and see resulting folder
ls /data/anatomy/freesurfer/subjects/55555/surfRelax

Important

❗ To make things easier in the long run, it's a good idea to make the _mprage_pp image the default base anatomy image for that participant.

This file is not the same as the file used for input to freesurfer / recon-all (it's padded to a consistent size and the intensities have been normalised, etc).