Related Experiment Video
Updated: Jun 22, 2025
06:34
A Single-Channel and Non-Invasive Wearable Brain-Computer Interface for Industry and Healthcare
Published on: July 7, 2023
2.3K
A Convolutional Neural Network for SSVEP Identification by Using a Few-Channel EEG.
Xiaodong Li1,2, Shuoheng Yang1,2, Ningbo Fei2
1Orthopedics Center, The University of Hong Kong-Shenzhen Hospital, Shenzhen 518053, China.
Bioengineering (Basel, Switzerland)
|June 27, 2024
Summary
This study introduces an attention-based complex spectrum-convolutional neural network (atten-CCNN) to improve brain-computer interface (BCI) performance using few-channel electroencephalogram (EEG) signals. The novel model enhances steady-state visual evoked potential (SSVEP) identification, making wearable EEG devices more effective.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Computer Science
Background:
- Wearable electroencephalogram (EEG) devices are increasingly used in brain-computer interfaces (BCI) due to their portability and wearability.
- Conventional wearable EEG devices often have fewer channels, potentially decreasing BCI performance.
- Few-channel EEG devices are viable for steady-state visual evoked potential (SSVEP)-based BCI, but performance limitations exist.
Purpose of the Study:
- To address the performance decrease in few-channel EEG-based BCI.
- To propose a novel deep learning model for enhanced SSVEP identification.
- To validate the model's effectiveness with wearable EEG devices.
Main Methods:
- An attention-based complex spectrum-convolutional neural network (atten-CCNN) was developed.
- The model integrates a CNN with a squeeze-and-excitation block.
- The input to the model is the spectrum of the EEG signal.
Main Results:
- The atten-CCNN model was evaluated on wearable 40-class and public 12-class SSVEP datasets.
- Performance was assessed under both subject-independent and subject-dependent conditions.
- The atten-CCNN significantly outperformed baseline models for both three-channel and single-channel EEG SSVEP identification.
Conclusions:
- The atten-CCNN model effectively enhances SSVEP-BCI performance with few-channel EEG signals.
- This algorithm is well-suited for SSVEP identification in wearable EEG devices.
- The proposed method offers a viable solution for improving BCI applications with limited EEG channels.
