search

LIMO-EEG-Toolbox / limo_tools

Hierarchical Linear Modelling for MEEG data
https://limo-eeg-toolbox.github.io/limo_meeg/
Other
58 stars 27 forks source link

ENH - updates to accommodate FT data and code improvement suggestions #88

Closed schoffelen closed 2 years ago

schoffelen commented 2 years ago

I have changed some of the code logic, to accommodate FT data in the form of mat-files, but along the way I also encountered a few issue with the saving out of the Y, Yhat, Res etc. The code was doing a cd into the directory into which the results were to be saved, and subsequently it kind of relies (in other functions, e.g. limo_tf_4d_reshape) that the 'pwd' corresponds to the directory in which the LIMO file is saved. I think this is tricky, and might have worked overall, but I think it would be better to save out files using the full filepath (i.e. rather than doing cd(directory); save file variable, do save(fullfile(directory,file),variable) However, when I was thinking a bit about this I had trouble in the limo_design_matrix_tf function, calling limo_tf_4d_reshape. The latter function is originally called with a single input argument, which means that limo_tf_4d_reshape tries to load the LIMO file, assuming that it is the pwd... (are you still following?) Alternatively, I don't see a reason why the force_dim can not be entered as a second input argument. Let me know if you want to quickly chat about my suggested changes in a zoom session or so.

CPernet commented 2 years ago

old code, totally agree, all new code also uses fullfile(path, name) one should not 'move' from current directory - I'll double check this, thx.

CPernet commented 2 years ago

a lot of broken parts :-( working on it but better looking code 👍🏻

schoffelen commented 2 years ago

a lot of broken parts :-( working on it but better looking code 👍🏻

sorry about that: do you have an error report, or a pointer to the unit tests? Let me know if I can help.

Also, feel free to revert to an older version of the branch and only keep the minimally necessary changes.

CPernet commented 2 years ago

this now works fine for basics using https://github.com/LIMO-EEG-Toolbox/limo_test_stats/blob/master/limo_test_integration.m still some errors for contrasts

this very same script is using preprocessed data from Wakeman and Henson 2015, I use in my tutorial - this is what @robertoostenveld would like to use