Dear experts,

I was wondering what is your opinion on the fact, that, according to the mindboggle paper

http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.100535...

it is beneficial to use ANTS volume-based gray matter segmentation to refine gray matter segmentation in FreeSurfer.

I am not expert in ANTS but looking at the ANTS code it seems to me that the ANTS routine does not do anything more than the volume-based nonlinear registration to the template and intensity and probabilistic atlas-based voxel labeling. Therefore, the method is in principle identical to GCA volume-based labeling (mri_ca_register and mri_ca_label, part of recon-all) in FreeSurfer which also produces cortical gray matter mask, far not so precise as the segmentation based on pial and white surface estimation.

I think that FreeSurfer pipeline for estimation of white and pial surfaces is much more sophisticated and precise than any method using volume-based registration and volume-based labeling by intensity and anatomical priors.

The FreeSurfer gray matter underestimation showed in figure 2 of the paper http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.100535... seems to me like either partial-volume effect or skull-strip error (which cuts out part of gray matter) which can be corrected by proper inspection and manual correction of brainmask.mgz. In contrast, the errors in presented ANTS gray matter segmentation in figure 2 are much more severe, spanning not only the blue encircled region.

I would very appreciate your expert opinion on this.

Regards,

Antonin Skoch

Announcing the official release of Mindboggle (http://mindboggle.info), open source software http://mindboggle.info/software.html and data https://osf.io/ydyxu/ for analyzing the shapes of brain structures from human MRI data (processed through FreeSurfer and optionally through ANTs). The release coincides with a publication in PLoS Computational Biology that documents and evaluates the software:Klein A, Ghosh SS, Bao FS, Giard J, Hame Y, Stavsky E, Lee N, Rossa B, Reuter M, Neto EC, Keshavan A. (2017) Mindboggling morphometry of human brains. PLoS Computational Biology 13(3): e1005350. doi:10.1371/journal.pcbi.1005350 http://doi.org/10.1371/journal.pcbi.1005350

Cheers, @rno