Hello,

We are having some trouble with autorecon1, specifially the watershed step. Here's the error we get:

> #--------------------------------------------
> #@# Skull Stripping Tue Nov 18 15:19:59 EST 2008
> /autofs/space/madrc_007/users/adeluca/RETRRO_structural/RETRRO_20_old/mri
>
> mri_watershed -T1 -brain_atlas
> /usr/local/freesurfer/stable4/average/RB_all_withskull_2008-03-26.gca
> transforms/talairach_with_skull.lta T1.mgz brainmask.auto.mgz
>
>
> Mode:          T1 normalized volume
> Mode:          Use the information of atlas (default parms, --help for
> details)
>
> *********************************************************
> The input file is T1.mgz
> The output file is brainmask.auto.mgz
> If this is incorrect, please exit with CTL-C
>
> Weighting the input with atlas information before watershed
>
> *************************WATERSHED**************************
> Sorting...
>      first estimation of the COG coord: x=128 y=108 z=113 r=76
>      first estimation of the main basin volume: 1871279 voxels
>
>
>
> *******************************************
> ***********************************************
> ***********************************************
> White matter intensity 0 is lower than CSF intensity 18.
> Please examine input images.  Will terminate ...
> ***********************************************
> ***********************************************
> ***********************************************
>
> **********************************************************
>         MRIstripSkull failed.
> ************************************************************
> Linux aubz 2.6.18-92.1.10.el5 #1 SMP Tue Aug 5 07:42:41 EDT 2008 x86_64
> x86_64 x86_64 GNU/Linux
>
> recon-all exited with ERRORS at Tue Nov 18 15:20:09 EST 2008

Looking at the input file (nu.mgz) it is obvious that the qcquired image isn't as good as it could be. Looks like there is moderate white matter burden, dropout on the ventral side of the scan, and some artifacts possibly made because of dental implants.

Are there any other steps that can be taken besides marking control points and adding the -f flag to mri_normalize?

Thanks,
Jon