External Email - Use Caution
Hi,
I am trying to use the longitudinal pipeline of the “hippocampal subfields and nuclei of amygdala”. Keep getting error messages when running it longitudinal but with no error message running it cross-sectional. I want to know if there is any way to fix the error and to run a longitudinal pipeline successfully.
I am using Freesurfer/dev and matlab_runtime/2014b.
I got two timepoints for each subject and finished recon-all longitudinal with no errors.
1) First Try
segmentHA_T1_long.sh SUBJ_base
run for about 15 mins and exited with error message:
“Error using kvlGEMSMatlab
/autofs/space/panamint_005/users/iglesias/software/freesurfer/GEMS2/kvlAtlasMeshDeformationOptimizer.cxx:327:
itk::ERROR: AtlasMeshDeformationLBFGSOptimizer(0x2ac1fead7740): search direction is not a descent direction!
Error in kvlStepOptimizer (line 11)
Error in SegmentSubfieldsT1Longitudinal”
no mri/lh.amygNucVolumes-T1.v21 and mri/lh.hippoSfVolumes-T1.v21
generated.
2) Second Try
segmentHA_T1.sh SUBJ1_timepoint1.long.SUBJ1_base
segmentHA_T1.sh SUBJ1_timepoint2.long.SUBJ1_base
run for about 16 mins, most successful but a few with errors:
“Error in segmentSubjectT1_autoEstimateAlveusML (line 1539)”
Or
“Open file failed because file '/tmp/MCR_2841219388/.mcrCache8.4/segmen0/autofs/space/panamint_005/users/iglesias/software/freesurfer.GEMS2.terrier/bin/kvlSetMeshCollectionPositions.m' has invalid content. Details: 'file is closed.failbit is set.'
The file
"/tmp/MCR_2841219388/.mcrCache8.4/segmen0/autofs/space/panamint_005/users/iglesias/software/freesurfer.GEMS2.terrier/bin/kvlSetMeshCollectionPositions.m"
cannot be executed.
Error in segmentSubjectT1_autoEstimateAlveusML (line 538)”
successful done subject or hemisphere generated
mri/lh.amygNucVolumes-T1.v21 and mri/lh.hippoSfVolumes-T1.v21.
3) Third Try cross-sectional
segmentHA_T1.sh SUBJ1_timepoint1
segmentHA_T1.sh SUBJ1_timepoint2
run for about 15 mins, all successful with no errors:
all subjects have mri/lh.amygNucVolumes-T1.v21 and
mri/lh.hippoSfVolumes-T1.v21 generated.
Manual inspection with freeview, it seems fine with all successful longitudinal and cross sectional subfield segmentation outputs mgz.
I wonder if I can fix first try and second try errors so that longitudinal output can be generated.
Thank you for your help!
Best,
Qi