Freesurfer Import MRI Dimension Errors

Bugs
, ,
1

Hello,

I am having some issues with the imported anatomies of my subjects in brainstorm. To run recon-all, I take the raw DICOM images and convert to a T1.mgz. I then take this input volume and run -recon-all -autorecon1 to produce a brainmask.mgz, and given there is little error with the brainmask I continue the rest of the reconstruction.

When I upload the anatomy folder into brainstorm, the MRI dimensions are not aligned in the way brainstorm wants, so I permute the dimensions to match the correct orientations ie. sagittal, axial, and coronal. Upon doing so, the ASEG and DKT volumes become unrecognizable since the dimensions don't match. Other times, when I have applied the transformation to the MRI and then ran recon-all, my subcortical surface becomes shifted along the y axis.

I was wondering if there was a standard pipeline with respect to preparing the subject anatomy prior to importing so these issues don't keep occurring? I noticed that this issue doesn't occur with .nii input files, but recon-all automatically converts the files to .mgz.

2

Hi @ethanberi,

Have you look at this tutorial?
https://neuroimage.usc.edu/brainstorm/Tutorials/LabelFreeSurfer

You may want to process your Anatomy with other too CAT12
in this list you can find links to other segmentation tools available through Brainstorm:
https://neuroimage.usc.edu/brainstorm/Tutorials#Advanced_tutorials

3

Hi,

Thanks for your response! As is turns out, this was a clerical error on my part. When running mri_convert on the DICOM files, I was naming the ouput "T1.mgz", thus, brainstorm was utilizing a T1 file that did not contain the proper transformation, which resulted from -autorecon1. To correct this error, I save the mri_convert output as "001.mgz" as this is the naming convention desired by freesurfer for the initial reconstruction steps.

Ethan