Hi CONN community,

I'm running Matlab 2025b, Conn 25.b, and SPM25. I'm trying to run functional outlier detection on a project with 21 subjects, 19 of whom have 8 sessions and 2 of whom have 6 sessions (so 164 sessions total). 

I ran it with distributed processing (run on Background process (Windows)) and set the number of parallel processes to 7, as my PC has 24 cores and is pretty fast.

When I returned to the project, it looked like everything ran as expected (subject folders now contained the expected art_outlier...mat files, etc.), but the data were not merged back into the project, 1st and 2nd level covariates were not updated with the outliers or QC metrics, etc.

After some digging, I saw this error in the logs:

Error using conn (line 1171)
invalid filename, project NOT saved

Error in conn_jobmanager (line 933)
conn save; % creates new dmat

As a test, I ran outlier detection on a smaller project with 2 subjects with 2 sessions each with two parallel processes. It ran as expected, with updated covariates added to the project and no errors.

I tried running my large project again, this time with 4 parallel processes (and I have since tried it with just 2). Again, the logs for all of the nodes show same errors as above, and if I let it run for longer, that error shows up several times.

Otherwise, as far as I can tell, the module runs as it should, with the log stating that "Mapping files for session X" is "done" across mutliple sessions, and reporting the global statistics and number of outliers identified across sessions, and stating that the SPM regressor files were saved to expected locations.

I can't figure out where the error is coming from but it seems to be preventing the larger project from successfully merging the separate nodes and saving the results into the project. I'm also not sure why the smaller 2-subject project worked without this error.

Has anyone else encountered this?

Thank you!

Jeff

Hi Jeff,

Not sure what this may be about. To help me figure out / debug this could you please zip and send me the contest of the last subdirectory within the [yourprojectname].qlog directory (assuming that the last time you run something in parallel was one of these failed attempts; those subdirectories will be named&sorted by date)

Best

Alfonso

Originally posted by johnsojp13:

Hi CONN community,

I'm running Matlab 2025b, Conn 25.b, and SPM25. I'm trying to run functional outlier detection on a project with 21 subjects, 19 of whom have 8 sessions and 2 of whom have 6 sessions (so 164 sessions total). 

I ran it with distributed processing (run on Background process (Windows)) and set the number of parallel processes to 7, as my PC has 24 cores and is pretty fast.

When I returned to the project, it looked like everything ran as expected (subject folders now contained the expected art_outlier...mat files, etc.), but the data were not merged back into the project, 1st and 2nd level covariates were not updated with the outliers or QC metrics, etc.

After some digging, I saw this error in the logs:

Error using conn (line 1171)
invalid filename, project NOT saved

Error in conn_jobmanager (line 933)
conn save; % creates new dmat

As a test, I ran outlier detection on a smaller project with 2 subjects with 2 sessions each with two parallel processes. It ran as expected, with updated covariates added to the project and no errors.

I tried running my large project again, this time with 4 parallel processes (and I have since tried it with just 2). Again, the logs for all of the nodes show same errors as above, and if I let it run for longer, that error shows up several times.

Otherwise, as far as I can tell, the module runs as it should, with the log stating that "Mapping files for session X" is "done" across mutliple sessions, and reporting the global statistics and number of outliers identified across sessions, and stating that the SPM regressor files were saved to expected locations.

I can't figure out where the error is coming from but it seems to be preventing the larger project from successfully merging the separate nodes and saving the results into the project. I'm also not sure why the smaller 2-subject project worked without this error.

Has anyone else encountered this?

Thank you!

Jeff

Hi Alfonso, 

Hi Jeff,

for some reason I am not able to attach the patch here, in any case the current development version of CONN in github (github.com/alfnie/conn) now includes this fix (and this specific patch is in the file conn.m)

Thanks again, Alfonso. I updated my CONN with the patch and tried running in parallel again and it worked perfectly without any errors. To answer your question: I didn't use a batch script to run the parallel process -- I simply selected "distributed processing (run on Background process (Windows))" in the GUI, specified the number of parallel processes (I tried it several times with the number ranging from 2-10 and got the error every time), and which outlier detection thresholds to use (I always chose default intermediate). That was it. I hope this is helpful.