MNEflow: Neural networks for EEG/MEG decoding and interpretation

Abstract

MNEflow is a Python package for applying deep neural networks to multichannel electroencephalograpic (EEG) and magnetoencephalographic (MEG) measurements. This software comprises Tensorflow-based implementations of several popular convolutional neural network (CNN) models for EEG–MEG data and introduces a flexible pipeline enabling easy application of the most common preprocessing, validation, and model interpretation approaches. The software aims to save time and computational resources required for applying neural networks to the analysis of EEG and MEG data.

Keywords

• Neural networks
• Electroencephalography
• Magnetoencephalography
• Tensorflow
• Machine learning