This repository contains a replication of the EEGNet code, an efficient convolutional neural network architecture for EEG-based brain–computer interface applications. The project is intended to make the original work more accessible and to serve as a basis for further research and experimentation in EEG signal analysis.
EEGNet is a compact CNN architecture originally developed by Lawhern et al. for classifying EEG signals in various BCI tasks. This repository includes:
- Data Preprocessing: Scripts for loading and processing EEG data.
- Model Architecture: Implementation of the EEGNet neural network.
- Training and Evaluation: Pipelines to train the network and evaluate its performance.
If you use this code or its ideas in your research, please cite the original work:
Lawhern, V. J., Solon, A. J., Waytowich, N. R., Gordon, S. M., Hung, C. P., & Lance, B. J. (2018). EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces. Journal of Neural Engineering, 15(5), 056013.
- Python 3.10.14
- Pytorch 2.5.1
- Clone the repository:
git clone https://github.com/yourusername/EEGNet-replication.git
- Navigate to the project directory:
cd EEGNet-replication
To start training the model, run:
python train.py /data_loader.py: you need to have you own data and use the path of the data in the file (e.g. shape:(1,eeg channels, sample pts))./models.py: Definition of the EEGNet architecture. You can change the F1 and D as you want/train.py: Training and evaluation pipelines.README.md: This file.
Contributions are welcome! If you encounter any issues or have suggestions for improvement, please open an issue or submit a pull request.
Many thanks to the original authors for their groundbreaking work on EEGNet, which has inspired this replication. Their contributions are acknowledged through proper citation and reference in this repository.