HDNN-ArtifactBrainState is a cutting-edge project that integrates Hopfield networks with deep neural networks to enhance brain state decoding, focusing on resilience against artifacts present in anesthesia recordings. This repository houses the code and resources for implementing the HDNN framework described in our recent publication.

OpenReview paper: OpenReview Link

ArXiv preprint: ArXiv Paper

Ensure you have Conda installed on your system for managing packages and environments.

To set up your environment for HDNN-ArtifactBrainState, run the following commands:

$ conda create --name HDNN python=3.9
$ conda activate HDNN
$ conda install tensorflow=2.13.0 numpy=1.24.3 matplotlib=3.5.2 sklearn=1.1.1 seaborn=0.12.2
$ conda install -c conda-forge jupyterlab

(sklearn and tensorflow can be installed via: pip install scikit-learn==1.1.1 and pip install tensorflow==1.24.3 if conda channels can't install these dependencies)

Clone the repository:

$ git clone https://github.com/arnaumarin/HDNN-ArtifactBrainState.git
$ cd HDNN-ArtifactBrainState
$ conda activate HDNN
$ jupyter lab

Open and run the Demo_Hopfield_CNN.ipynb notebook in JupyterLab to explore the functionalities of the HDNN framework.

If you find this paper or implementation useful, please consider citing our Proceedings:

@inproceedings{
marin-llobet2023hopfieldenhanced,
title={Hopfield-Enhanced Deep Neural Networks for Artifact-Resilient Brain State Decoding},
author={Arnau Marin-Llobet and Arnau Manasanch and Maria V. Sanchez Vives},
booktitle={Associative Memory {\&} Hopfield Networks in 2023},
year={2023},
url={https://openreview.net/forum?id=M7yGTXajq5}
}

And/or the Arxiv paper:

@misc{marinllobet2023hopfieldenhanced,
      title={Hopfield-Enhanced Deep Neural Networks for Artifact-Resilient Brain State Decoding}, 
      author={Arnau Marin-Llobet and Arnau Manasanch and Maria V. Sanchez-Vives},
      year={2023},
      eprint={2311.03421},
      archivePrefix={arXiv},
      primaryClass={q-bio.NC}
      }

For any queries or contributions, please open an issue or submit a pull request in this repository.