[Eeglablist] EEGLAB2022 study precomputed files directory and parameters

[Dr Cyril, Pernet]

Hi Agatha

> Q1. I have different subsets of trials extracted as epoch files. It looks like if I put them in different directories - the code will do what I want. It is a little unfortunate that the specification of a directory is lost - the produced files look a little out of place in the BIDS structure.
if you DM me some examples of directory I'd happily check what is the 
code made the folder structure not BIDS compliant (derivatives, no 
derivatives, something else?)
> Q2. If this is documented anywhere the link would be welcome - I couldn’t find it in the web docs.
good point I'll document that better - temporary files are created with 
the trials corresponding to your design and this is what Yr is within 
LIMO directory ; every command executed on which file is stored in the 
pipeline.mat (the closest to an explanation for that is 
https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_LIMO-2DEEG-2DToolbox_limo-5Ftools_wiki_debugging-2D1st-2Dlevel-2Dstats&d=DwIDaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=i5T9hZ_fUhWB6t0Wr6d7ywgH4VchDJAM0s_ZBMYmwk8_3jkjoOnVy5ODsX5qYCH6&s=4kN8PGl_fnf5v5kl8BnCbWGmemkOBE-H4g8RWH2Nz5c&e= )
> It seems like right now if I run a pre-computation of the single trial data, I need to choose parameters that will fit any analysis I may want. But what about if I want to run one analysis on ERSPs with one set of time bins, and then again with another set of time bins.
LIMO EEG doesn't create folder with this info since it is design for 
whole space (channel, ica, source) data analysis
> Or if I want to repeat an analysis with ERSP calculated with different baselines.
then it the previous situation - it seen as different data (but good 
point because of where this applies in the pipeline that a little 
annoying to have to copy data just for that, I'll look into chaing this 
but it will take a while)
> It seems like right now - the precomputation will overwrite existing precomputed files. I used to be able to handle that by creating new designs. As it stands - I don’t see how multiple precomputation parameters can be tested.

well it can, i just ran 10 different pipelines, but you need to copy 
files as new sets of parameters are getting applied

Cyril