Abstract:OBJECTIVE: Brain-computer interfaces (BCI) enable direct communication with a computer, using neural activity as the control signal. This neural signal is generally chosen from a variety of well-studied electroencephalogram (EEG) signals. For a given BCI paradigm, feature extractors and classifiers are tailored to the distinct characteristics of its expected EEG control signal, limiting its application to that specific signal. Convolutional neural networks (CNNs), which have been used in computer vision and speech recognition to perform automatic feature extraction and classification, have successfully been applied to EEG-based BCIs; however, they have mainly been applied to single BCI paradigms and thus it remains unclear how these architectures generalize to other paradigms. Here, we ask if we can design a single CNN architecture to accurately classify EEG signals from different BCI paradigms, while simultaneously being as compact as possible.APPROACH: In this work we introduce EEGNet, a compact convolutional neural network for EEG-based BCIs. We introduce the use of depthwise and separable convolutions to construct an EEG-specific model which encapsulates well-known EEG feature extraction concepts for BCI. We compare EEGNet, both for within-subject and cross-subject classification, to current state-of-the-art approaches across four BCI paradigms: P300 visual-evoked potentials, error-related negativity responses (ERN), movement-related cortical potentials (MRCP), and sensory motor rhythms (SMR).MAIN RESULTS: We show that EEGNet generalizes across paradigms better than, and achieves comparably high performance to, the reference algorithms when only limited training data is available across all tested paradigms. In addition, we demonstrate three different approaches to visualize the contents of a trained EEGNet model to enable interpretation of the learned features.SIGNIFICANCE: Our results suggest that EEGNet is robust enough to learn a wide variety of interpretable features over a range of BCI tasks. Our models can be found at: https://github.com/vlawhern/arl-eegmodels.

CBGT-Net: A Neuromimetic Architecture for Robust Classification of Streaming Data

Hierarchical Gate Network for Fine-Grained Visual Recognition.

CSNN: an Augmented Spiking Based Framework with Perceptron-Inception

EEGGAN-Net: enhancing EEG signal classification through data augmentation

RB-GAT: A Text Classification Model Based on RoBERTa-BiGRU with Graph ATtention Network

EEG_RL-Net: Enhancing EEG MI Classification through Reinforcement Learning-Optimised Graph Neural Networks

Brain-inspired Dual-Pathway Neural Network Architecture and Its Generalization Analysis

Brain Cognition-Inspired Dual-Pathway CNN Architecture for Image Classification

EEG-NeXt: A Modernized ConvNet for The Classification of Cognitive Activity from EEG

EBT Deep Net: Ensemble brain tumor Deep Net for multi-classification of brain tumor in MR images

GT-Net: global transformer network for multiclass brain tumor classification using MR images

EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces

EEGNet: A Compact Convolutional Network for EEG-based Brain-Computer Interfaces

EEG-ITNet: An Explainable Inception Temporal Convolutional Network for Motor Imagery Classification

Cog-Net: A Cognitive Network for Fine-Grained Ship Classification and Retrieval in Remote Sensing Images

Multi-Glimpse Network: A Robust and Efficient Classification Architecture based on Recurrent Downsampled Attention.

MBGA-Net: A multi-branch graph adaptive network for individualized motor imagery EEG classification

BGaitR-Net: An effective neural model for occlusion reconstruction in gait sequences by exploiting the key pose information

CSA-Net: An Adaptive Binary Neural Network and Application on Remote Sensing Image Classification

EEG classification for visual brain decoding with spatio-temporal and transformer based paradigms

A Spiking Neural Network With Adaptive Graph Convolution and LSTM for EEG-Based Brain-Computer Interfaces