I am a bit stumped. Can you check that cache.th20.pos.sig.ocn.mgh and concat_output.nii.gz? I'm sure they most be. If so, send both to me on our file drop. doug

On 11/04/2013 07:00 PM, Michael Waskom wrote:

Hi Doug,

I'm running into a strange seg fault when trying to use mri_segstats (it's actually ocurring within mri_glmfit-sim, but I can reproduce it manually)

This is the line that fails (stripping the leading paths for readability):

mri_segstats --seg cache.th20.pos.sig.ocn.mgh --exclude 0 --i concat_output.nii.gz --sum mri_glmfit-sim.junk.84357 --avgwf cache.th20.pos.y.ocn.dat

With the log:

$Id: mri_segstats.c,v 1.75.2.9 2013/02/16 00:09:33 greve Exp $ cwd cmdline mri_segstats --seg /Users/mwaskom/Studies/PUnCH/cache/punch-switch_control/fsaverage_group/_l1_contrast_decision_neg/_hemi_lh/glmfit/_glm_results/osgm/cache.th20.pos.sig.ocn.mgh --exclude 0 --i /Users/mwaskom/Studies/PUnCH/cache/punch-switch_control/fsaverage_group/_l1_contrast_decision_neg/_hemi_lh/glmfit/_glm_results/concat_output.nii.gz --sum mri_glmfit-sim.junk.84357 --avgwf cache.th20.pos.y.ocn.dat sysname Darwin hostname ben-octocore.stanford.edu http://ben-octocore.stanford.edu machine i386 user mwaskom UseRobust 0 Loading /Users/mwaskom/Studies/PUnCH/cache/punch-switch_control/fsaverage_group/_l1_contrast_decision_neg/_hemi_lh/glmfit/_glm_results/osgm/cache.th20.pos.sig.ocn.mgh Loading /Users/mwaskom/Studies/PUnCH/cache/punch-switch_control/fsaverage_group/_l1_contrast_decision_neg/_hemi_lh/glmfit/_glm_results/concat_output.nii.gz Voxel Volume is 1 mm^3 Generating list of segmentation ids Found 5 segmentations Computing statistics for each segmentation 1 1 10934 10934.000 2 2 2409 2409.000 3 3 1620 1620.000 4 4 2044 2044.000

Reporting on 4 segmentations Computing spatial average of each frame 0 1 2 3Segmentation fault

Intriguingly, it does not segfault if I remove the --avgwf parameter OR the --exclude 0 parameter. It does segfault if I use --avgwfvol instead of --avgwf

I'm not too sure what's going on, I can otherwise read the segmentation file and --y file and do statistical operations on them, and I can write other files to the directory I'm testing this in.

Here's my system info:

FREESURFER_HOME: /usr/local/freesurfer/5.3.0

Build stamp: freesurfer-Darwin-snowleopard-i686-stable-pub-v5.3.0

Kernel info: Darwin 10.8.0 i386

Any thoughts?

Michael

---

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