Re: [Freesurfer] Artefactual skull stripping for some subjects

Hi Tudor

the T1.mgz is intensity corrected and conformed (the orig.mgz is just conformed). Have you read the documentation on fixing skull strip problems on the wiki?

cheers Bruce On Mon, 8 Jul 2013, Tudor Popescu wrote:

Dear FSers,

Still haven't got to the bottom of this, any suggestions would be much appreciated, thanks in advance!

Tudor

On 5 July 2013 17:26, Tudor Popescu tudor3@gmail.com wrote: Hi everyone

  The skull-stripping step of recon-all must have gone bad for my sample, because I noticed that there are artefacts
  (chips) in the brain-extracted structurals of several of my subjects. Since I ran recon-all on all my sample without
  having noticed this initially, this of course also lead to the surfaces being created wrongly, i.e. following the
  wrong volumetric information.

  I attach screenshots showing some of the more obvious artefactual skull extractions, corresponding to subjects 04,
  06 and 17, and also the brain.mgz and T1.mgz for subject 17.

  FSL's BET extracted the skull with no problems from the same original structurals, so I'm guessing it's maybe the
  case that I need to override some parameters in recon-all in order for the initial recon-all to produce good brain
  extractions?

  And by the way, I noticed that, in any one subject's folder, the images /mri/T1.mgz, /mri/orig.mgz and
  /mri/orig/001.mgz all seem to contain the subject's original structural scan. Why does this exist in 3 separate
  copies, which also seem to be of different file sizes? The only difference I can see between them is in terms of
  brightness correction...

  Many thanks!

  Tudor