Hello,

Ive had some issues re-running my .sh scripts for conn batch processes. I end up with coc0c0T1w_restore.1.60.nii prepended to the original structural image T1w_restore.1.60.nii. 

How can I re-run a node without these issues? Where is it pulling this from when everything points to T1w_restore.1.60.nii? I try deleting the coc0c0T1w_restore.1.60.nii file that was made so it doesn't try to pull from something already in the folder but it just keeps adding c0 to the file.

Thank you!
Sarah

CONN: RUNNING SETUP.PREPROCESSING STEP
Preparing structural Segmentation (Grey/White/CSF tissue estimation). Please wait...
Performing structural Segmentation (Grey/White/CSF tissue estimation). Please wait...

SPM preprocessing job
(1).spm.spatial.preproc.channel.vols = /tmp/yassamri2/HCP_yassa_rsFMRI/HCP_download/sub001/MNINonLinear/c0c0T1w_restore.1.60.nii

Hi Sarah,

I imagine adding to your scripts explicitly the definition of your structural data should help fix this, e.g.

  batch.Setup.structurals={'/tmp/yassamri2/HCP_yassa_rsFMRI/HCP_download/sub001/MNINonLinear/T1w_restore.1.60.nii'};

In general, batch commands, just like the GUI, will act on your current data, and the c0c0c0* prefixes indicate that you are running segmentation sequentially many times (i.e. you are segmenting the already-segmented data). In particular the c0 prefix simply indicates that your structural data (after segmentation) is skull-stripped. 

Hope this helps
Alfonso
Originally posted by Sarah Kark:

Thank you Alfonso. 

I'd like to see if there's anything I can change in the nose .sh or  node .mat files that were created for each participant in parallel. Or do I need to use the script to create a new batch? 132 participants completed successfully, just 3 didn't work so I tried to re-run them and then ran into this problem. I thought the script would start back at the original submitted batch.

I have attached the script I used that made 135 .sh files (1 for each participant). If I re-run this for just the 3 that need it, I'm afraid it will make a separate project for just these 3 participants and not integrate them with the other 132 participants.

I have been searching in the node .mat files and the project file for anywhere to swap out c0T1 for just the original file.

Thanks!
Sarah
Originally posted by Alfonso Nieto-Castanon:

Hi Sarah,

I see what you mean now and that is a very good point, thanks for all the additional information. I believe simply deleting the .dmat file associated with those 3 participants (e.g. if node.123456 failed, delete the file CONN_*.123456.dmat) and re-running once again their .sh files will work, making sure that preprocessing re-starts from the original input files. After that finishes, re-loading your original conn project should then merge all of the (now finished) 135 participant files. 

Please let me know if that seems to fix this issue and I will see if I can create a patch to avoiding similar issues from re-occurring again

Thanks
Alfonso

Originally posted by Sarah Kark: