Skip to content

A Deep Learning Library for State-Based EEG Analysis

License

Notifications You must be signed in to change notification settings

konspatl/DL-EEG

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Introduction

A Deep Learning-based EEG (DL-EEG) decoding library for state-based tasks/analyses, using MNE and Tensorflow/Keras.

This library is intended for EEG researchers and includes dataset processing tools, implementations of several cNN models, and scripts for model training, testing and visualization.

The aim of this project is to provide:

  • a standard framework for EEG preprocessing and dataset extraction, suitable for machine learning analyses
  • methods for creation and training of deep learning models that employ EEG-driven network designs and methodologies
  • methods for evaluation and testing of deep learning models over a variety of EEG systems and learning tasks
  • a consistent EEG-processing pipeline, by exploiting the spatio-temporal structure of the EEG
  • support for multi-study integration and cross-subject/cross-study validation schemes, aiming at improved research reproducibility

Requirements

  • Python >= 3.6
  • mne == 0.19
  • tensorflow >= 2.6
  • numpy
  • scipy
  • natsort
  • matplotlib
  • pandas
  • scikit-learn

Deep Learning-based EEG Models

The following deep learning models are implemented in the library (utils.NN):

  • cNN_3D [1]
  • cNN_topomap [1]
  • BrainDecode_Deep4 [2]
  • BrainDecode_Shallow [2]
  • EEGNet_v1 [3]
  • EEGNet_v2 [3]

Usage

To use this library, place the contents of the DL-EEG folder in your PYTHONPATH environment variable

Citation

If you use this library for your research, please cite the following work:

@phdthesis{kar97272,
           month = {September},
           title = {Deep Learning for Electrophysiological Investigation and Estimation of Anesthetic-Induced Unconsciousness},
          school = {University of Kent,},
          author = {Konstantinos Patlatzoglou},
            year = {2022},
        keywords = {deep learning EEG anesthesia consciousness},
             url = {https://kar.kent.ac.uk/97272/}
}