omarjnb / dl-eeg-playground

Deep Learning EEG Playground

dl-eeg-playground

You have found the Deep Learning EEG Playground, put together by the Montreal Hacknight.

The repo is a bit messy, but what you should find in here:

• examples on how to usual stuff with colab
• a pyRiemann comparative example
• brain-decode based experimentations:
  • tutorial from their website
  • x86 execution
  • colab based code
  • sklearn wrapper

We are currently working toward integrating braindecode into MOABB, feel free to join us every other Fridays @ District 3 Innovation Center

SETUP

• We assume you are using Anaconda, python 3.5

• Install Brain Decode: https://github.com/robintibor/braindecode

  • If you are on Windows, You can install PyTorch using these instructions. You only need to go up to step 4.A.

• Go through the TrialWise Tutorial: https://robintibor.github.io/braindecode/notebooks/TrialWise_Decoding.html to make sure everything is setup properly

• Download this dataset : Two class motor imagery (002-2014) at http://bnci-horizon-2020.eu/database/data-sets

• put everything in an new folder (here)/BBCIData/

The remaining of the project is described in our jupyter notebook(s) 1 - Two-Classes Classification (BNCI)

SETUP (Google Colab)

1. Download 2 - Two-Classes Classification (BNCI) Colab.ipynb and upload it on your Google Drive.
2. Open Google Colab
3. Instructions for dataset download and python library installation are described in the Jupyter notebook.

Resources

For more papers on DL applications on EEG, you could refer to this repo. Feel free to contribute.