help
513 Subscribers 
help > problem with importing ROI data
Showing 1-2 of 2 posts
Oct 12, 2015 09:10 AM | Till Langhammer - Technical University of Dresden
problem with importing ROI data
follow error occured while importing ROT data...
AND when I add new subjects this step is done for all subjects even though I choose "skip already processed subjects". We habe 239 subjects in total and this step takes more than one day, everytime i try to add new subjects. is this necessary?
ERROR DESCRIPTION:
Error using conn_process (line 557)
No suprathreshold voxels in ROI file D:\C1\rs\18580\functional\c2wmmana.nii after erosion (original file contained 1036 suprathreshold voxels)
Error in conn_process (line 12)
case 'setup', disp(['CONN: RUNNING SETUP STEP']); conn_process([0:5]);
Error in conn (line 2377)
conn_process('setup');
Error in conn_menumanager (line 109)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.15.a
SPM8 + Beamforming DEM FieldMap MEEGtools
Matlab v.2013b
storage: 170.9Gb available
AND when I add new subjects this step is done for all subjects even though I choose "skip already processed subjects". We habe 239 subjects in total and this step takes more than one day, everytime i try to add new subjects. is this necessary?
ERROR DESCRIPTION:
Error using conn_process (line 557)
No suprathreshold voxels in ROI file D:\C1\rs\18580\functional\c2wmmana.nii after erosion (original file contained 1036 suprathreshold voxels)
Error in conn_process (line 12)
case 'setup', disp(['CONN: RUNNING SETUP STEP']); conn_process([0:5]);
Error in conn (line 2377)
conn_process('setup');
Error in conn_menumanager (line 109)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.15.a
SPM8 + Beamforming DEM FieldMap MEEGtools
Matlab v.2013b
storage: 170.9Gb available
Oct 14, 2015 01:10 AM | Alfonso Nieto-Castanon - Boston University
RE: problem with importing ROI data
Hi Till,
That error likely indicates that your white matter masks (e.g. generated during normalization or segmentation, depending on your preprocessing pipeline) does not include as many voxels as one would typically expect. That typically indicates some problem during normalization so I would probably suggest to repeat the normalization step perhaps changing the initial state (e.g. by manually repositioning the structural volume to better match the standard template and/or trying to identify and fix what might have caused an issue in this subject's normalization in the first place).
And regarding the excessively long time when using the "skip already processed subjects" option to process only your new subject(s), you are right that there were a few inefficiencies in how CONN checked the data to see if any of it had changed and needed to be processed or not, and those inefficiencies could quickly accumulate on a relatively large proejct. I am attaching a patch fixing this (this patch is for release 15g, simply unzip this file in the conn distribution folder overwriting the files with the same name there). The patch will:
a) fix some of those inefficiencies (I tested this on ~500 subjects on a machine with somewhat slow I/O access and it should now take no longer than a couple of minutes to go over the entire dataset checking whether some data has changed or not)
and b) added the option to apply CONN processing steps only to a selected set of subjects (for any of the Setup/Denoising/First-level steps). This option was already available in 15g using batch scripts, I simply added the ability to access that same functionality from the GUI. For example, if you have added a few subjects to an already-processed conn project and want to simply process those new subjects you may now in the Setup tab click on the 'Done' button and then uncheck the 'All Subjects' checkbox and select only your new subject(s) (similarly you may select Tools.Run.ConnProcessingSteps and run all Setup/Denoising/First-level steps for only those new subjects in a single step). Also note that you can also mix this functionality with the cluster computing options in 15g to process your new subjects in parallel in a computer cluster environment for even faster performance.
Hope this helps and let me know if you would like to clarify any of the above and/or if you run into any issues
Alfonso
Originally posted by Till Langhammer:
That error likely indicates that your white matter masks (e.g. generated during normalization or segmentation, depending on your preprocessing pipeline) does not include as many voxels as one would typically expect. That typically indicates some problem during normalization so I would probably suggest to repeat the normalization step perhaps changing the initial state (e.g. by manually repositioning the structural volume to better match the standard template and/or trying to identify and fix what might have caused an issue in this subject's normalization in the first place).
And regarding the excessively long time when using the "skip already processed subjects" option to process only your new subject(s), you are right that there were a few inefficiencies in how CONN checked the data to see if any of it had changed and needed to be processed or not, and those inefficiencies could quickly accumulate on a relatively large proejct. I am attaching a patch fixing this (this patch is for release 15g, simply unzip this file in the conn distribution folder overwriting the files with the same name there). The patch will:
a) fix some of those inefficiencies (I tested this on ~500 subjects on a machine with somewhat slow I/O access and it should now take no longer than a couple of minutes to go over the entire dataset checking whether some data has changed or not)
and b) added the option to apply CONN processing steps only to a selected set of subjects (for any of the Setup/Denoising/First-level steps). This option was already available in 15g using batch scripts, I simply added the ability to access that same functionality from the GUI. For example, if you have added a few subjects to an already-processed conn project and want to simply process those new subjects you may now in the Setup tab click on the 'Done' button and then uncheck the 'All Subjects' checkbox and select only your new subject(s) (similarly you may select Tools.Run.ConnProcessingSteps and run all Setup/Denoising/First-level steps for only those new subjects in a single step). Also note that you can also mix this functionality with the cluster computing options in 15g to process your new subjects in parallel in a computer cluster environment for even faster performance.
Hope this helps and let me know if you would like to clarify any of the above and/or if you run into any issues
Alfonso
Originally posted by Till Langhammer:
follow error occured while importing ROT
data...
AND when I add new subjects this step is done for all subjects even though I choose "skip already processed subjects". We habe 239 subjects in total and this step takes more than one day, everytime i try to add new subjects. is this necessary?
ERROR DESCRIPTION:
Error using conn_process (line 557)
No suprathreshold voxels in ROI file D:\C1\rs\18580\functional\c2wmmana.nii after erosion (original file contained 1036 suprathreshold voxels)
Error in conn_process (line 12)
case 'setup', disp(['CONN: RUNNING SETUP STEP']); conn_process([0:5]);
Error in conn (line 2377)
conn_process('setup');
Error in conn_menumanager (line 109)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.15.a
SPM8 + Beamforming DEM FieldMap MEEGtools
Matlab v.2013b
storage: 170.9Gb available
AND when I add new subjects this step is done for all subjects even though I choose "skip already processed subjects". We habe 239 subjects in total and this step takes more than one day, everytime i try to add new subjects. is this necessary?
ERROR DESCRIPTION:
Error using conn_process (line 557)
No suprathreshold voxels in ROI file D:\C1\rs\18580\functional\c2wmmana.nii after erosion (original file contained 1036 suprathreshold voxels)
Error in conn_process (line 12)
case 'setup', disp(['CONN: RUNNING SETUP STEP']); conn_process([0:5]);
Error in conn (line 2377)
conn_process('setup');
Error in conn_menumanager (line 109)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
CONN v.15.a
SPM8 + Beamforming DEM FieldMap MEEGtools
Matlab v.2013b
storage: 170.9Gb available