[Eeglablist] making dataset of timepoints x channels x epochs

[Tarik S Bel-Bahar]

Hello, some notes below for you.
For other questions or unanswered questions, it's recommended that you
follow good guidelines for questions to help lists.
Provide good examples and/or pictures, give details about the settings that
you are using in eeglab, and more details about errors messages you are
getting. Generally speaking, try to break down your questions into more
specific simpler questions.
Cheers!
********************************************************

   1. Make sure you are subscribed correctly to the list:
   https://sccn.ucsd.edu/wiki/EEGLAB_mailing_lists
   2. Google eeglablist archives (for all details about eeglablist and the
   archives of past messages)
   3. Google eeglablist +a specific topic to see/review past messages on a
   specific topic
   4. Responses to queries from users must await initial approval of the
   question and then approval of the response by eeglablist admin. Otherwise
   they don't show up on eeglablist. If a responder choose to CC you on a
   response, then you will get a copy of the response before the admin accepts
   to be posted on eeglablist (as detailed in info about eeglablist at the
   main eeglablist page).
   5. If you are new to eeglab, and you haven't had a chance to yet, first
   review the eeglab wiki, and do the full eeglab tutorial with the eeglab
   tutorial data, then you will get a better sense of what eeglab expects and
   what usual kind of data is.
   6. If you are new to eeg, it's best to start with reading/watching some
   tutorials, google EEGLAB summerschool and ERPLAB for a wide range of
   learning materials (videos, powepoints, papers). Steve Luck's handbook of
   ERP can also help beginners with all the basic details.
   7. Your dataset is highly unusual (much too short) and seems to have too
   few and too short epochs for valid EEG analysis. it's best to find some
   high-quality EEG articles (via Google Scholar) and emulate those (in terms
   of total recording time, and total number of trials/epochs), and then start
   over. Don't give up! But the most important thing is to have proper data
   before starting your data processing
   8. For normalization,in the first instance, emulate a high-quality
   article using a protocol that is similar to yours.
   9. You are unlikely to get an ERP without at least 20 to 50 trials per
   condition. (so one would have 50 trials of your 2-4 second epochs for each
   condition).
   10. For your current data, try the following as some first steps: All
   doable via GUI. Look an on eeglab wiki for basic instructions. Load each of
   your datasets for one person (with the 2-4 second conditions). Preprocess
   the files appropriately (e.g, filtering, cleaning, bad channel removal, and
   other preprocessing steps that may be needed). Then append them so they are
   one file. Then make an epoched .set file for each kind of event. Then
   repeat for each subject all the previous steps.
   11. Your epoching step that you tried should work, but you need to have
   the same event name for each event in the dataset, and if you don't, you
   need to create extra events in the .set file (e.g., at the start of each
   trial). Also, eeglab will expect to cut each epoch into the same length, so
   pay close attention to the settings in the epoching gui. For example if you
   first event is at the start of the file, and one sets the epoch to take -1
   sec to 2 seconds, then there must be data before the event.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20181205/776eeb9d/attachment.html>