[Eeglablist] running AMICA with multiple models for subject data

Hauge,Theresa C theresa.hauge at ufl.edu
Tue Jun 18 11:09:17 PDT 2024


Hello Seyed,

Thank you for your help and clarification, once I switched to our HPC things became much easier to navigate. I successfully prepared and ran AMICA1.7 with multiple models within our university's HPC and also submitted batches directly to the computing cluster. I am able to load the AMICA components and weights for each model and view the extended component properties, but not all of the plotting functions in the postAmicaUtility are running successfully (I am using the updated version of the postAmicaUtility plugin).

For example, when attempting to "Plot Model Probability Time series" with the pop_modprobplot() function (either in the GUI or in the command window), I receive an error message when attempting to plot the second time-window figure:
In pop_modprobplot (line 142)
A dot name structure assignment is illegal when the structure is empty.  Use a subscript on the structure.

Error in scrollplot_fast (line 65)
    jScrollBar.Minimum = 0;

Error in modprobplot (line 101)
scrollplot_fast(y,L,models2plot,T,EEG.srate, events, [EEG.xmin EEG.xmax], events2plot,smoothlength);

Error in pop_modprobplot (line 142)
[v2plot, llt2plot] = modprobplot(EEG,models2plot,smoothlength,events2plot);

(I also receive an error message for another function in the GUI, specifically using the function ozeegplot() when testing on multiple subjects).

Is there any insight into what my next steps should be? I'm not entirely sure that I know how to send data or images to the cloud service, but I am happy to send what is needed in case this should be replicated or submitted to the postAmicaUtility GitHub issues section.


Best,
        Theresa

Theresa C. Hauge, M.A.
She/Her/Hers
PhD Candidate, Applied Physiology & Kinesiology
Neuromotor Behavior Laboratory
University of Florida
theresa.hauge at ufl.edu
@TheresaCHauge

-----Original Message-----
From: eeglablist <eeglablist-bounces at sccn.ucsd.edu> On Behalf Of Seyed Yahya Shirazi via eeglablist
Sent: Wednesday, June 12, 2024 09:22
To: EEGLAB List <eeglablist at sccn.ucsd.edu>
Subject: Re: [Eeglablist] running AMICA with multiple models for subject data

[External Email]

Good day, Theresa,

Good first submission.

It seems that you ran amica17_nsg.m on your local PC, and it failed.

This version will not run on a PC or a local computer (especially Windows PCs; the "export" command is a Unix-only command, and there are also Slurm-only commands in this version that require a high-performance computing cluster, HPC). You need to run it on a Matlab instance on a supercomputer with Slurm.


Overall, you need to do the following:

Start a Matlab Session on your HPC, this can be by having a graphical or command-line interactive sessions, based on waht is supported on your HPC.
Complete preprocessing of the dataset how ever you aim.
Make sure that the latest AMICA respository is on your path (you can cloine it from GitHub https://urldefense.com/v3/__https://github.com/sccn/amica__;!!Mih3wA!F6H2FTKA5_J9NLO1uK211GuDh8tYhRhVFGRj8xYWbrnz9xK87qsTqOcbqN3p0zMuV9y0olvmtsS5aMtRpoaO$ ) Call runamica17_nsg.m <https://urldefense.com/v3/__https://github.com/sccn/amica/blob/master/runamica17_nsg.m__;!!Mih3wA!F6H2FTKA5_J9NLO1uK211GuDh8tYhRhVFGRj8xYWbrnz9xK87qsTqOcbqN3p0zMuV9y0olvmtsS5aE28x0-S$ > with your EEG structure from the command line. You can't invoke this version from GUI.  Your current parameters seem to be ok.
Collect the results from the output folder, load the wieghts, or use the postAmicaUtility plug-in to analyze the results.


Please let us know how it goes with this setup on your HPC.

BTW, I think attachments are not carried through this service. If you want to share one, you should upload it to a cloud service and then sahre the link.

Best,
Yahya

______________________
Seyed (Yahya) Shirazi, Ph.D., Asst. Project Scientist, Swartz Center for Computational Neuroscience, Institute for Neural Computation, UCSD, neuromechanist.github.io <https://urldefense.com/v3/__http://neuromechanist.github.io/__;!!Mih3wA!F6H2FTKA5_J9NLO1uK211GuDh8tYhRhVFGRj8xYWbrnz9xK87qsTqOcbqN3p0zMuV9y0olvmtsS5aIpLOzV8$ > _______________________________________________
Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
For digest mode, send an email with the subject "set digest mime" to eeglablist-request at sccn.ucsd.edu


More information about the eeglablist mailing list