Hi guys,
The problem was garbage in - garbage out:
1. The issue manifested in the skull stripping step but was caused by bad input 2. I am a data science guy (not neurology), new to freesurfer, and followed the instructions at https://surfer.nmr.mgh.harvard.edu/fswiki/FreeSurferBeginnersGuide
recon-all \ -i <one slice in the anatomical dicom series> \
3. When I issued recon-all, I gave it a -i of an image file from my mri data. I assumed that it would find the T1- weighted data from my collection automatically as in dcmunpack.
4. It seems that the -i requires specifically the T1-weighted data, because when I looked at the tkmedit brainmask images they were inverted (maybe from T2 or something else).
5. I re-ran the recon-all on the proper T1 data and it ran to completion with no errors.
I received several offline requests from other users, so it seems like this may be a common mistake.
I recommend making the requirement for input of T1-weighted data more explicit in the wiki page.
As a feature request, it seems pretty easy to use dcmunpack to find the correct data from a collection and make it part of initial recon-all.
I had uploaded my original data to the ftp site per reply below, but feel free to delete and close the issue as solved.
Cheers,
-Matt
------------
Hi Matt
no idea what that would happen. If you upload the dataset we will take a look
cheers Bruce On Mon, 5 Mar 2018, Matthew Grecsek wrote:
Following your advice below, it was clearly apparent that the output from the skull stripping step was missing a lot. I tried adjusting the -wsthresh flag from 50 -> 200 and it still left major parts of the brain out.
I have a good quality mri and subject with no known anatomical abnormalities.
The log file indicates "surface validation has detected possible Error"
Does this suggest a more meaningful defect in the base image or something that can be corrected via other parameters or manual editing?
Log file attached.
Thanks,
-Matt
Bruce Fischl Fri, 02 Mar 2018 08:00:11 -0800
Hi Matt
check the output of the skull stripping. This error usually means that too much (everything?) has been erased and no plausible cc could be found
cheers Bruce
On Thu, 1 Mar 2018, Matthew Grecsek wrote:
My recon-all ended prematurely with the following message:
final transformation (x=136.0, yr=-10.054, zr=-0.114): 0.98464 0.00200 -0.17457 15.51372; -0.00197 1.00000 0.00035 17.22442; 0.17457 0.00000 0.98464 -17.83761; 0.00000 0.00000 0.00000 1.00000; mri_cc: no WM voxels found with norm > 40 -- check skull stripping
No such file or directory Linux centos-7-1 3.10.0-693.17.1.el7.x86_64 #1 SMP Thu Jan 25 20:13:58 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
recon-all -s matt2 exited with ERRORS at Thu Mar 1 04:35:09 UTC 2018
I am running: freesurfer-Linux-centos6_x86_64-stable-pub-v6.0.0-2beb96c
My full log file is attached.
Please let me know suggestions for correcting and continuing the process.
Thanks,
-Matt