External Email - Use Caution
Hello,
I am attempting to run hippocampal segmentation with freesurfer 7.3.1, using the segmentHA_T2.sh script with a high-resolution hippocampus T2 sequence as additional scan input. The code runs fine, but oddly it places the outputs into the mri subfolder rather than into stats (output file names shown below).
mri/
rh.amygNucVolumes-T2HighResHippo.v22.txt rh.hippoSfVolumes-T2HighResHippo.v22.txt
rh.hippoAmygLabels-T2HighResHippo.v22.CA.FSvoxelSpace.mgz rh.hippoAmygLabels-T2HighResHippo.v22.CA.mgz rh.hippoAmygLabels-T2HighResHippo.v22.FS60.FSvoxelSpace.mgz rh.hippoAmygLabels-T2HighResHippo.v22.FS60.mgz rh.hippoAmygLabels-T2HighResHippo.v22.FSvoxelSpace.mgz rh.hippoAmygLabels-T2HighResHippo.v22.HBT.FSvoxelSpace.mgz rh.hippoAmygLabels-T2HighResHippo.v22.HBT.mgz rh.hippoAmygLabels-T2HighResHippo.v22.mgz
The stats folder also contains output .stats files for hippocampus and amygdala, but they only have one sentence in them: " # Amygdala nuclei volumes as created by segmentHA_T2.sh " These are the names of the files in the stats directory:
stats/ amygdalar-nuclei.lh.T2.v22.T2HighResHippo.stats amygdalar-nuclei.rh.T2.v22.T2HighResHippo.stats hipposubfields.lh.T2.v22.T2HighResHippo.stats hipposubfields.rh.T2.v22.T2HighResHippo.stats
It seems to me that the output files are initialized in the stats directory, but the values themselves get written to files in the mri directory.
When I go to compile the data across subjects, the ConcatenateSubregionsResults script is missing from the installation (I have confirmed this both for 7.3.1 and 7.4.1 installations). I found the following thread in archive with a similar issue of this file missing: https://secure-web.cisco.com/1D6pNu3LKBDnQ_g29D1J9DrkaQ4h3xi6d3NHKg1r8YUImCq...
In this thread Antoine shared a direct download for the ConcatenateSubregionsResults, then Jackson shares a batched version of the script to resolve an issue with it looking for files in the /mri directory (which is where those files are written in my case anyways).
Either by using the unbatched script or using the batched script and manually moving the files from the mri to the stats, I am able to run the script but run into another issue. The output file contains in the first row the values for the first subject and then for all subsequent subject rows it has region labels. I looked at each subject's output and it looks like this:
Hippocampal_tail 547.477135 subiculum-body 198.862009 CA1-body 108.611407 subiculum-head 162.383267 hippocampal-fissure 132.385149 presubiculum-head 100.360234 CA1-head 477.119333 presubiculum-body 85.674081 parasubiculum 58.297345 molecular_layer_HP-head 180.316933 molecular_layer_HP-body 231.129446 GC-ML-DG-head 132.306221 CA3-body 66.096743 GC-ML-DG-body 115.076547 CA4-head 108.334300 CA4-body 105.624384 fimbria 86.764321 CA3-head 89.481195 HATA 52.204101 Whole_hippocampal_body 997.838938 Whole_hippocampal_head 1360.802930 Whole_hippocampus 2906.119003
My guess is that the labels are being incorrectly pulled from the second column and the values from the first column.
Doe anyone have any insight on how to resolve this? I could always pull what I need manually, but if this is an issue with the code it would be nice to see it resolved.
Many thanks! Jenya C. Indiana University School of Medicine