Sleep Well Baby: sleep stage prediction

Author: Richard Bartels Email: r.t.bartels-6@umcutrecht.nl

Introduction

This repository contains code to run the sleep-well baby model that can classify sleep stages in preterm infants based on vital signs. It only contains code for model inference, not for training.

Important: this software is for research use only and not intended for clinical or diagnostic purposes. Use at your own risk.

Getting started

Installation

⚠️ Warning:
This software only runs on Python 3.8 because the model file was produced using that version.

The easiest way to install the sleepwellbaby package is to to use the package manager uv, a modern Python package manager that simplifies dependency management and ensures reproducibility:

uv sync --frozen

Starterkit

An example of how to run this code can be found in notebooks/example.ipynb

Documentation

Dataset and model information can be found in the dataset card and model card, respectively.

Citation instructions

Reference should be made to the sleep-well baby paper when using this software.

@article{10.1093/sleep/zsac143,
    author = {Sentner, Thom and Wang, Xiaowan and de Groot, Eline R and van Schaijk, Lieke and Tataranno, Maria Luisa and Vijlbrief, Daniel C and Benders, Manon J N L and Bartels, Richard and Dudink, Jeroen},
    title = {The Sleep Well Baby project: an automated real-time sleep–wake state prediction algorithm in preterm infants},
    journal = {Sleep},
    volume = {45},
    number = {10},
    pages = {zsac143},
    year = {2022},
    month = {06},
    issn = {0161-8105},
    doi = {10.1093/sleep/zsac143},
    url = {https://doi.org/10.1093/sleep/zsac143},
    eprint = {https://academic.oup.com/sleep/article-pdf/45/10/zsac143/45986688/zsac143.pdf},
}