AviatorYan / sleep-apnea

Convolutional neural network which uses 30 second single-channel EEG signals to detect apnea events

sleep-apnea

Two models:

• A convolutional neural network (CNN) which takes as input 30-second single-channel EEG signals and produces as output a sleep apnea classification.
• A CNN trained on the same data which classifies sleep stage. Possible classifications are N1, N2, N3, REM, and W (awake).

Both models include implementation, validation, and an example trained model.

Requirements

• Python 3.5-3.8 (you can use pyenv to manage different python versions).
• Training data from Physionet.

Setup

This project is best set up in a virtual environment to isolate package and language versions.

In the terminal, navigate into the project directory and create a new virtual environment using a compatible (3.5-3.8) version of Python:

python3.8 -m venv env

Activate virtual environment:

source env/bin/activate

Install required packages:

python -m pip install -r requirements.txt

You also need wget if you want to automate the downloading of Physionet data:

brew install wget

You're all set! Run python3 -m src.sleepapnea.getdata to retrieve and preprocess physionet data, and python3 -m src.sleepapnea.modelarchitecture to train your first model.