To run the experiment of labelling ICA components:

Library requirements

# if conda is installed: 
conda activate

# Install virtualenv 
virtualenv env

# Activate virtualenv
source env/bin/activate #for unix 
env/Script/activate #for window

# Install required packages
pip install -r req.txt

Program flow:

• Download an EDF file and put it into edf folder
• Go to config.ini, change path_edf, and path_stage into paths to *.edf file and STAGE.csv file
• Run main.py

# run program to label data:
python main.py

• The program will ask for session id.
• After typing session id, the program will display ICA component of the chunk & topomap of each component.
• After that, it'll promt you to reject a component(s), user will enter list of components-to-reject sperated by space. If there is no component to reject, just leave it bank.
• The program will then display the entire EEG signal of session before and after EOG cleaning.
• If the result is satisfying, press y to save this chunk to training data, otherwise press n.
• Repeat until sufficient amount of samples is collected.
• Upon finish, data will be saved in the csvs folder, where each component record is a csv file.