SSVEP-Neural-Generative-Models

Code to accompany our International Joint Conference on Neural Networks (IJCNN) paper entitled - Simulating Brain Signals: Creating Synthetic EEG Data via Neural-Based Generative Models for Improved SSVEP Classification.

The code is structured as follows:

• Data_pred.py contains functions to pre-proposes EEG data;
• EEG_Gen.py A sample script showing how all models can be run to generate data;
• EEG_DCGAN.py Our DCGAN based model for generating SSVEP-based EEG data;
• EEG_WGAN.py Our Wasserstein GAN based model for generating SSVEP-based EEG data;
• EEG_VAE.py Our Variational Autoencoder based model for generating SSVEP-based EEG data;

The Sampledata directory contains some sample SSVEP EEG data on which the models can be trained.

Dependencies and Requirements

The code has been designed to support python 3.6+ only. The project has the following dependencies and version requirements:

• torch=1.1.0+
• numpy=1.16++
• python=3.6.5+
• scipy=1.1.0+

Cite

Please cite the associated papers for this work if you use this code:

@article{aznan2019simulating,
  title={Simulating Brain Signals: Creating Synthetic EEG Data via Neural-Based Generative Models for Improved SSVEP Classification},
  author={Aznan, Nik Khadijah Nik and Atapour-Abarghouei, Amir and Bonner, Stephen and Connolly, Jason and Moubayed, Noura Al and Breckon, Toby},
  journal={arXiv preprint arXiv:1901.07429},
  year={2019}
}