Oops, slight edit
import nibabel as nib
file_list = [os.path.join(os.getenv(‘SUBJECTS_DIR’), subject, ‘mri’, ‘brainmask.mgz’)for subject in subject_list] mri_list = [nib.load(fname) for fname in file_list]
On Feb 22, 2022, at 9:53 PM, Fischl, Bruce BFISCHL@mgh.harvard.edu wrote:
In python you should be able to do (assuming you have a list of subject ids in subject_list):
import nibabel as nib
file_list = [os.path.join(os.getenv(‘SUBJECTS_DIR’), subject) for subject in subject_list] mri_list = [nib.load(fname) for fname in file_list]
even for 1000 subjects it won’t take that long, depending of course on your disk speed. You shouldn’t have to convert from .mgz – nibabel can read it
From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu On Behalf Of REEM ABU BAKR BAHATHIQ Sent: Tuesday, February 22, 2022 9:28 PM To: freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] convert .mgz to .nii
External Email - Use Caution
Hello FreeSurfer experts,
I want to run a classification task using deep learning on data processed by Freesurfer recon-all and I have some queries. I will use brainmask .mgz
First of all, based on my simple knowledge, I cannot feed my model with this output directly. .mgz files must be converted to a .nii file for use in analytics. I found that this is done by writing mri_convert any_volume.mgz any_volume.nii.gz
But how can I convert the files of all participants (more than 1000) into .nii files instead of going one by one?
Are there any examples or tutorials provided to demonstrate the use of Freesurfer output in machine learning applications?
My greetings,