Hello,
I am running Freesurfer version 5.3.0 with the FLAIR surface refinement bbregister patch installed. I have recently encountered a problem with applying pial edits concurrently with the option of surface adjustment using T2 FLAIR scans during recon-all. After initially running recon-all with the FLAIR option flags:
'recon-all -subjid <subj> -all -FLAIR <flair_scan.mgz> -FLAIRpial'
I had tried to apply pial edits without the -FLAIRpial flag on the 2nd recon assuming the adjustments would carry over.:
'recon-all -subjid <subj> -autorecon-pial'
This did not work out and the surfaces were worse than before as T2 adjustments were not applied. I then reran the same command adding the -FLAIRpial option:
'recon-all -subjid <subj> -autorecon-pial -FLAIRpial'
Although the surfaces were better due to the T2 adjustment and similar to the initial recon, recon-all did not apply any of the pial edits. This was odd because other scans that required control points (with '-autorecon2-cp -autorecon3') or white matter edits (with 'autorecon2-wm -autorecon3') did apply all manual edits, including adjusting the pial surfaces for deleted voxels and refining by the T2 scans.
After some detective work looking into the log file and recon-all code, I think I may have found the problem. Turns out, just in the case of pial edits, the ?h.woFLAIR.pial files are not being re-written to reflect the new surfaces correcting for edited voxels following the "Make Pial Surf" step. This is noted in the date stamp of the ?h.woFLAIR.pial files being older than all other surface files following the last recon-all procedure. This is likely because the ?h.pial output files from the "Make Pial Surf" are being overwritten by the "Refine Pial Surfs w/ T2/FLAIR" procedure, instead of being copied over to the ?h.woFLAIR.pial files which are the input to the FLAIR adjustment, and would thus carry on the pial edits. Here is the recon-all code:
set nv1=(`mris_info $sdir/${hemi}.pial |& grep nvertices`) set nv2=(`mris_info $sdir/${hemi}.wo${t2flair}.pial |& grep nvertices`) if ( (! -e $sdir/${hemi}.wo${t2flair}.pial) || \ ("$nv1" != "$nv2") ) then set cmd = (cp -v $sdir/${hemi}.pial $sdir/${hemi}.wo${t2flair}.pial)
I have attached the log file for one of my subjids that had this issue (Note, the 10/20/15 recon-all is after installing the patch due to the bbregister issues, and the last recon-all on 12/2/15 I tried starting the stream earlier with the -autorecon2-wm flag just to see if that would help).
I think it would help to just copy the ?h.pial to ?h.woFLAIR.pial regardless of nvertices being the same. Please let me know if I can provide any more information!
Thank you for your time and support!
Victoria J. Williams, PhD Postdoctoral Fellow
VA Boston Healthcare System Jamaica Plain Neuroimaging Research for Veterans Center (NeRVe) Memory Disorders Research Center 150 South Huntington Ave D11-102 Boston, MA 02130