Hello,
Synopsis I have 5 longitudinally acquired high-res volumes at 0.8 mm isotropic resolution. The matrix size is 320x320x240 (240 is long left-right), covering an FOV of 256x256x192 mm3. I've used mri_robust_template to create a template out of these 5 volumes, preferably at the same matrix size and FOV. However, sometimes the matrix size of the created template is 321x321x321, resulting in the FOV of 256.8x256.8x192.8 mm3. This non-standard FOV would lead to error in the subsequent FreeSurfer segmentation analysis.
Details The command I used was: $ mri_robust_template \ -- mov sub1_rep01.nii sub1_rep02.nii sub1_rep03.nii sub1_rep04.nii sub1_rep05.nii \ -- template sub1.nii \ -- lta sub1_rep01.lta sub1_rep02.lta sub1_rep03.lta sub1_rep04.lta sub1_rep05.lta \ -- mapmov rep01_to_mean.nii rep02_to_mean.nii rep03_to_mean.nii rep04_to_mean.nii rep05_to_mean.nii \ -- average 0 \ -- iscale \ -- satit
The size of the output is NOT consistent: Sometimes the matrix becomes 320x320x320 (with the left-right FOV extended), but at other times it becomes 321x321x321. In the latter case, the FreeSurfer recon with -hires flag $ recon-all -i sub1.nii -subject recon_sub1 -all -hires -expert hires.expert-options would fail with the error message (from recon-all.log) ************************************** ERROR! FOV =256.799 > 256 Include the flag -cw256 with recon-all! Inspect orig.mgz to ensure the head is fully visible. ****************************************** (I am reluctant to explore the flag -cw256 for workflow reasons.)
So why the output of mri_robust_template not consistent? Is there anything intrinsic in the data that would affect the output data size?
1) FreeSurfer version: freesurfer-Linux-centos_x86_64-stable-pub-v6.0.0-2bed96c 2) Platform: CentOS Linux release 7.4.1708 (Core) 3) uname -a: Linux host 3.10.0-693.11.1.el7.x86_64 #1 SMP Mon Dec 4 23:52:40 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux