Deep Learning Based EEG forecasting toolbox

What is does?

1. Predict EEG from EEG.

2. Predict EEG from Stimului.

3. Predict Stimuli from EEG.

Install

1. Clone the repo

git clone https://github.com/vasudev-sharma/DeepEEG-Forecast
cd EEG-Forecast

2. Set up Virtual Environment

virtualenv eeg_env --python=3.6
source eeg_env/bin/activate

3. Now, in the terminal run

bash install_files.sh

Data

5 subjects EEG data downsampled at 160Hz

Each subject has

• 64 channels
• 192 trials
• 840 time points (5.25s)

Data can be found here - https://drive.google.com/drive/folders/1_gV6t5f2FDWo8OYTcAxsDxz4yhRcYKU0?usp=sharing

Contributing

We ❤️ contributions. Feel free to send us a PR or raise an issue.

1. Create an issue if there is one.
2. Fork the repo.
3. Create your feature branch (git checkout -b your-feature).
4. Add and commit your changes (git commit -am 'message').
5. Push the branch (git push origin your-feature).
6. Create a new Pull Request.

Tracking the performance metrics and logging the models

Weights and Biases

TODO

• Add instructions in readme to perform all the 3 tasks
• Implement Attention model
• Implement ESRNN model
• Deploy the model
• Refactor Code