S4Sleep: Elucidating the design space of deep-learning-based sleep stage classification models

Welcome to the official GitHub repository for the paper "S4Sleep: Elucidating the design space of deep-learning-based sleep stage classification models". If you consider this repository useful for you research, we would appreciate a citation of our preprint.

@article{wang2023s4sleep,
    title={S4Sleep: Elucidating the design space of deep-learning-based sleep stage classification models}, 
    author={Tiezhi Wang and Nils Strodthoff},
    year={2023},
    eprint={2310.06715},
    archivePrefix={arXiv},
    primaryClass={cs.LG},
    journal={arXiv preprint 2310.06715}
} 

Table of Contents

• Setup & Environment
• Datasets
• Preprocessing
• Classification
• Contribution
• Acknowledgments

Setup & Environment

1. Environment: The environment details and requirements for Linux systems are provided in the environment.yml folder. Please ensure to set up the provided environment to avoid compatibility issues.

2. Additional Package: You'll need to install the cauchy or other related packages for the S4 model. See the official repository for more details:

   https://github.com/HazyResearch/state-spaces/

Datasets

We utilize two primary datasets for our research:

• Sleep-EDF: https://www.physionet.org/content/sleep-edfx/1.0.0/
• SHHS visit1: https://sleepdata.org/datasets/shhs

Both datasets should be appropriately placed in the main directory or as instructed by specific scripts.

Preprocessing

To preprocess the raw data:

Run the preprocess.py script: python preprocess.py

This will process the raw data from the aforementioned datasets and prepare them for classification.

Classification

To run the classification:

1. Configuration: Modify the config files located at ./conf/data to specify the directory containing the preprocessed data.

2. Training a sleep staging model: Set the desired config file using the --config-name flag. Example for a time-series-based model on SEDF: python main.py --config-name=sedf_ts.yaml

You can use other config settings by specifying different .yaml files located in the ./classification/code/conf directory. We provide config files for the best-performing model architectures for time series and spectrograms as input representations.

Contribution

Feel free to submit pull requests, raise issues, or suggest enhancements to improve the project. We appreciate community contributions!

Acknowledgments

This work partly builds on the S4 layer kindly provided by https://github.com/HazyResearch/state-spaces/

For any further queries or concerns, please refer to the official publication or contact the project maintainers.