Dear CONN users,

as far as I've read in this forum, the CONN toolbox works fine for multi-band data. I used a script to set up a CONN batch to then load in the GUI and continue further processing. The setup works well until functional data are imported. Those are multi-band accelerated EPI with a TR of 0.376 s for 2 sessions (10 min & 20 min) and 55 subjects. The GUI just doesn't start importing functional data, even after quite some time.

It might be an issue of processing power. However, I ran the setup on a more powerful machine as well and ran into the same issue. Previously, doing the same procedure for functional data with a TR of 2.41 s worked well, on both computers. Does anybody have experience with loading such large functional datasets in CONN? Is it rather a problem concerning CONN (or just the CONN GUI) or do I need a more well equipped computer?

Or is there a problem in the way I set up conditions for those two sessions (session 2 is split up into 2 conditions with condition 3 for ~ the first 10 min and condition 2 for ~ the second 10 minutes)?

for subj = 1:length(para.subjs)
batch.Setup.conditions.onsets{1}{subj}{1} = [0];
batch.Setup.conditions.durations{1}{subj}{1} = [inf];
batch.Setup.conditions.onsets{1}{subj}{2} = [];
batch.Setup.conditions.durations{1}{subj}{2} = [];
end

for subj = 1:length(para.subjs)
batch.Setup.conditions.onsets{2}{subj}{2} = [599.73];
batch.Setup.conditions.durations{2}{subj}{2} = [599.72];
batch.Setup.conditions.onsets{2}{subj}{1} = [];
batch.Setup.conditions.durations{2}{subj}{1} = [];
end

for subj = 1:length(para.subjs)
batch.Setup.conditions.onsets{3}{subj}{2} = [0];
batch.Setup.conditions.durations{3}{subj}{2} = [599.72];
batch.Setup.conditions.onsets{3}{subj}{1} = [];
batch.Setup.conditions.durations{3}{subj}{1} = [];
end

Thank you,
Best,
Philipp

I ran the CONN setup on a high-performance computer. The COON toolbox again did not start loading the functional images. When trying to open the NIfTI files in MRIcron I received the error message "Error in reading NIfTI header: the headers need to be at least 348 bytes.". I checked the header with load_untouch_nii.m and it looked fine. Opening the functional images with Mango viewer worked well.

So, could it be the case that CONN has the same issue with the functional images as MRIcron viewer? Is there any chance to fix this?

Best,
Philipp

I split up the 20 min session into 2 chunks of 10 min sessions and set up a conn batch with 3 instead of 2 conditions. This way, the setup works. However, this does not really solve the problem mentioned in 'Update'. And, I am asking myself, if it is legitimate to split up one session with 2 conditions and process them in CONN as they were separate sessions (data are already preprocessed out of CONN, but I am having regression of confounders & temporal filtering in mind here). Any help is very much appreciated.

Hi Philipp,

Could you please provide a bit more information about what "did not start loading the functional images" mean? Do you get any particular error messages or does CONN just hangs/takes-forever? at what step in the processing stream exactly? If you are running things remotely in your cluster then look at the cluster log for any error messages (e.g. in conn_*.qlog/[date]/*.stderr or *.stdout), or, if you are running things locally on your computer, then simply look in Matlab command-line or in the error message GUI that pops up (if any) for a detailed description of the error message. 

Best
Alfonso
Originally posted by Philipp Riedel:

Dear Alfonso,

thank you for your response. There is no error message in the GUI that pops up. The Matlab command-line is as follows:

"Checking if data files have been edited or moved. Please wait...
Error with file [path]/*.nii"

The path is correctly displayed.

Best,
Philipp