FlagEEG_v1

Repo of the winter EEG project by the Club d'Intelligence Artificielle of Laval University. The goal of this project is to interpret EEG signals in order to transform them into realtime game inputs.

Configuration 16 canaux (Cyton + Daisy)

Le projet supporte 16 canaux EEG avec Cyton + Daisy :

• Enregistrement : main_annotation.py utilise CYTON_DAISY_BOARD
• Training : data_handler.py charge 16 canaux (compatibilité 8ch avec padding)
• Inference : model_inference.py et main_inference.py utilisent 16 canaux
• CSV : Noms d'électrodes 10-20 (FP1, FP2, F7, F8, FZ, CZ, O1, O2, F3, F4, T3, T4, T5, T6, P3, P4)

Prérequis : Module Daisy branché sur Cyton, 8 électrodes supplémentaires, réentraîner le modèle avec données 16 canaux.

Table of Contents

• FlagEEG_v1
• Table of Contents
  • Python version
  • How to get started with the project
  • How to get started with the environment
  • How to get started with the EEG
    • 1. Install the OpenBCI GUI : download
    • 2. Connect your computer to the helmet
    • 3. Read the OpenBCI docs
  • How to run the baseline
  • How to run the annotation pipeline

Python version

Repo is tested for python 3.12.8

How to get started with the project

1. Go read the project-chart.md here
2. Go read the project-organization-manual.md here

How to get started with the environment

1. Install the Git CLI
2. Clone the FLapEEG_v1 repo with this command : git clone git@github.com:cia-ulaval/FlagEEG_v1.git (If it doesn't work, you need to setup an SSH key on your GitHub account)
3. Install conda
4. Inside the repo you cloned, run this command to build the required environment :

   conda env create -f env.yml
   

   This env.yml environment was tailored to be compatible with all systems including Windows, Linux and MacOs (but was only tested with Windows and Linux)
5. Activate the environement to have all the required dependencies:

   conda activate cia_flapeeg
   

   You should now be good to go to launch the wanted scripts!

How to get started with the EEG

1. Install the OpenBCI GUI : download

This is the official OpenBCI support user interface to allow streaming and seeing data from the EEG. Please download it and follow the installation instructions.

• Download
• Installation instructions

Important for Linux users : For Linux users only, there is a necessary hardware step, to make in order to allow the necessary permissions for your USB ports. Go look at the tutorial of the fix. The fix is in the section : Linux Users: Serial Port Permissions

For lazy bones, essentially you need to open permissions to usb ports for uses. In my case, on my Ubuntu machine this was the command that fixed it :

sudo usermod -aG dialout $USER

Then reboot.

2. Connect your computer to the helmet

1. Connect the Daisy dongle to your USB port
2. Run the OpenBCI GUI on your computer and start a streaming session : CYTON -> Serial (from Dongle)

3. Read the OpenBCI docs

To help connect your python scripts and the EEG helmet, OpenBCI provides us with a SDK : Brainflow. Go read the docs to learn how python and the EEG interacts.

How to run the baseline

0. Plug in the helmet

0.1 Plug the Dongle in the right most USB port of the front of CASTOR

Important : the Dongle's switch should be on the GPIO_6 and NOT the RESET switch

0.2 Put the battery in the helmet and open the switch

1. Open the OpenBCI GUI:

1.1 Two options to open it :

• Run this command in a terminal : ./OpenBCI_GUI
• Or you could look inside the OpenBCI_GUI folder, right click and run the OpenBCI_GUI script as a program :

1.2 Start the stream session

• In the OpenBCI GUI, you need to the session : CYTON (live) -> Serial (from Dongle) -> AUTO-CONNECT :
• You should have this inferface :

1.3 Change the Accelerometer Window to a Networking window

• 

1.4 Change the UDP protocol to the LSL protocol

• 

1.5 Select the Stream 1 Datatype to be TimeSeriesFilt

• 

1.6 Start the LSL Data Stream

• 

1.7 Start the global DataStream

• 

2. Start the FlapEEG game:

• Second option : open a console or a code editor like Visual Studio Code and open this project : EEG_flappy_bird
  • Then open a terminal
  • Then activate the base conda environement : conda activate
  • Then acivate the cia_flapeeg environement : conda activate cia_flapeeg
  • Then launc the script : python main.py

3. Start the detector - two options :

• Second option : open a console or a code editor like Visual Studio Code and open this project : FlapEEG_interface_v1
  • Then open a terminal
  • Then activate the base conda environement : conda activate
  • Then acivate the cia_flapeeg environement : conda activate cia_flapeeg
  • Then cd inside the good directory to run the script : cd src/data_streamer
  • Then launc the script : python lsl_streamer.py

4. Focus back on the game window

• To register the keyboard presses, you need to focus back the mouse on the game window. So click back on the window.

How to run the annotation pipeline

1. Setup the environment : How to get started with the environment

2. Connect the EEG with the dongle

3. Start the pipeline :

   python main_annotation.py

   An interface like this should show up :

4. Select the target folder where all the .csv files are exported :

   

5. Write the name that your exported .csv file will take :

   ⚠️ Important:

    - Do NOT write `.csv` in the name, the extension is automatically   added.
   
   
    - The new data recorded is appended to the end of the document  inputed. So if you want to have separate files for different recording session, change the name of the input.
   

   

6. Start recording with the Record button. A window is supposed to show up to help you annotate your EEG data. Please blink when the FlappyBrain shows up.

7. Stop recording with the red Stop button.