I took a look at the xhemireg script and I feel like editing this to serve my purposes might be too tall of an order.

However, I think I have a processing stream that will work — I want to make sure there’s nothing problematic about it.

1. Run recon-all on the properly oriented T1 images through -segmentation 2. Take the old hand-edited wm.mgz file [containing extensive hand-edits] and left/right flip this file (using mri_convert to change LIA to RIA) 3. Visually confirm alignment of this hand-edited, flipped wm.mgz file. It turns out this brain was shifted by 1 voxel in the left/right direction; I manually adjusted this after converting the file to COR-format, and then converted back to .mgz and visually confirmed proper alignment. 4. Drop this hand-edited, L/R flipped and 1-voxel-shifted wm.mgz file into the subject/mri directory from running recon-all in (#1), then proceed with recon-all -autorecon2-wm -autorecon3

Is there anything inherently flawed about this approach? Are there any other files that I also need to flip (i.e., wm.seg.mgz or anything else?)

thanks, Dan

On 4/17/17 6:48 PM, Douglas Greve wrote:

you can try xhemireg. it is kind of meant to do this kind of thing, but not to the extent that you need it. If you know a little programming, you can try to edit the script so that it does everything.

On 4/17/17 6:43 PM, Dan Grupe wrote:

Hi all,

After running a bunch of subjects through FreeSurfer and making extensive manual edits to the white matter masks, we discovered that the data were left/right flipped during the dicom-to-nifti conversion process. I hope to maintain these manual edits while processing data in the proper orientation and running them through the rest of the pipeline. Is this possible to do?

An old thread from 10 years ago directed me to this page: https://surfer.nmr.mgh.harvard.edu/fswiki/LeftRightReversal that offered some tips on how to go about this. However, after confirming that the volume was reversed running recon-all -autorecon1 in a new directory, I encountered an error message:

...

ERROR: talairach_afd: Talairach Transform: transforms/talairach.xfm

***FAILED*** (p=0.0000, pval=0.0000 < threshold=0.0050)

This problem does not arise when running the same command on a copy of this subject’s non-flipped data.

Are there different steps that I should be following to rectify the problem? Or can you help me troubleshoot this error?

Best, Dan