Analyzing MEG data with MNE-python, and their extensions

^ This is the dataflow of overall process. Please note.

1. Place all your raw files in the same directory where python codes are.

   • Naming convention: xx_RAW/ where xx is your initial. (e.g. mk_RAW/)
   • Raw file directory should have sub-directories for each subject (e.g. mk_RAW/ABCDEFGH)
   • Each sub-directory should have one bevioral data file (.csv) and one or more MEG data files (.ds).

2. Change your configuration at mk_config.py

   • Set shift_onset as True if you want movement onset results.
   • Set split_cond as True if you want to decode by each condition.
   • Set decode_next as True if you want to decode (n+1)th trial.
   • Set filter parameters and epoch parameters.
   • Choose which variables to decode in the list analyses.
   • Set directory names for epoched, results, and plots directories.
   • Naming convention: xx_yyyyyy_zzdecim/

3. Run mk_save_epochs.py

   • This will create xx_epoched_zzdecim/ and sub-directories.
   • Epoched data (-epo.fif) will be generated per subject.
   • Behavioral data (.csv) will be copied per subject.

4. Run mk_decode_motor_task.py

   • This will create xx_results_zzdecim/ and sub-directories.
   • Evoked data (-ave.fif) of filters and patterns will be generated per analysis per subject.
   • Data-only numpy array (.npy) of filters, patterns and scores will be generated per analysis per subject.
   • xx_results_zzdecim/_average/ will be created also and the average data across subjects are stored.

   (Note: mk_decode_motor_task_csp.py is under construction!)

5. Run mk_make_topomap.py

   • Topomaps of patterns and filters will be created per analysis per subject.
   • Topomaps of average across subjects will be created in xx_results_zzdecim/_average/.

6. Run mk_plot_decoding.py

   • This will create xx_plots_zzdecim/ with no sub-directory.
   • Four plots will be generated from the decoding results per analysis.

1. Create .swarm files.
2. Change these lines:

      # subject = sys.argv[1]     # Use these lines when using swarm!
      # if True:                  # Use these lines when using swarm!
      for subject in subjects:    # Delete this line when using swarm!

   to this:

      subject = sys.argv[1]       # Use these lines when using swarm!
      if True:                    # Use these lines when using swarm!
      # for subject in subjects:  # Delete this line when using swarm!

3. Do NOT use swarm for mk_make_topomap.py.
4. For mk_plot_decoding.py, argv[1] should be the name of analyses, not subjects.