The method used to correct for the volume fraction effect is detailed in our paper:

Greve, Douglas N., et al. "Different partial volume correction methods lead to different conclusions: An 18 F-FDG-PET study of aging." /Neuroimage/ 132 (2016): 334-343.

There is not a way to include a non-linear registration or B0/voxel shift map in mri_gtmpvc -- at the time I programmed it, I had not envisioned anyone using it in this way! mri_vol2surf and bbregister will take a voxel shift map which can be generated from a B0 map; see epidewarp.fsl

Not sure what you are asking about in those steps

On 9/8/17 10:23 AM, Victor Montal Blancafort wrote:

Dear Freesurfer experts,

I have been working with the PETsurf pipeline. I was wondering if there is any chance to use a non-linear registration (i.e ANTS) in order to compute the %tissue per voxel in native space.

Currently, I am using your pipeline to process DWI data. Due to some EPI artifacts during the acquisition, I would like to use a deformation algorithm to get rid off those artifacts that would generate false values when projecting to surface (i.e misalignment between DTI and T1 in (shrinked) frontal areas, will cause erroneous (outliers) values.

If this is not possible, could you please explain in detail how the mri_gtmpvc algorithm works to compute the % tissue fraction of each voxel (to implement it)? Correct me if I'm wrong:

1. Downsample the DWI to a 3x3x3 grip (any command for that)?
2. Normalize to the high-res MRI.
3. Using the aparc+aseg?, compute the % of each ROI in that voxel

(again, command?). 4) Assign in the native space, the % of ROI at each voxel. 5) Compute the tissue prob as the sum of all ROIs for a certain tissue type in that voxel.

BONUS: mri_vol2surf doesn't allow either non-linear registration?

Any insight will be really appreciate!! Victor Montal

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer