As a sanity check in a pipeline I'm running I tried converting an atlas file containing visual ROIs from .mgz to .nii.gz and then immediately converting the resulting .nii.gz file back to .mgz:
mri_convert lh.atlas.mgz lh.atlas.nii.gz --reslice_like subject_t1.nii.gz --resample_type nearest --out_data_type int --no_scale 1
mri_convert lh.atlas.nii.gz lh.atlas_reconverted.mgz --reslice_like subject_t1.nii.gz --resample_type nearest --out_data_type int --no_scale 1
Something is getting lost in this conversion and reconversion, because when I run the following two lines (to generate a label for V1) on the two different .mgz files, the original file correctly results in the identification of 1591 label voxels, while the converted/reconverted file incorrectly results in the identification of 111 label voxels, none of which overlap with the 1591 correct voxels:
mri_vol2surf --src mri/lh.atlas.mgz \
--src_type FLOAT \
--srcreg mri/register.dat \
--hemi lh \
--surf white \
--o mri/lh.atlas.mgh
mri_cor2label --c mri/lh.atlas.mgh \
--id 1 \
--surf SubID lh \
--l label/lh.V1_ventral.label
Any feedback on where I'm going wrong here would be greatly appreciated!
Thanks for your time,
JD Knotts
1) FreeSurfer version: freesurfer-Darwin-lion-stable-pub-v5.3.0
2) Platform: Mac OSX 10.11.6
3) Darwin Jeffreys-MacBook-Pro.local 15.6.0 Darwin Kernel Version 15.6.0: Thu Jun 23 18:25:34 PDT 2016; root:xnu-3248.60.10~1/RELEASE_X86_64 x86_64
--
Doctoral Candidate, Cognitive Neuroscience
Department of Psychology
University of California, Los Angeles