Related Experiment Video
Updated: Apr 30, 2026
06:34
A Single-Channel and Non-Invasive Wearable Brain-Computer Interface for Industry and Healthcare
Published on: July 7, 2023
3.6K
An automated and fast approach to detect single-trial visual evoked potentials with application to brain-computer
Yiheng Tu1, Yeung Sam Hung1, Li Hu2
1Department of Electrical and Electronic Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong, China.
Summary
This study introduces an automated method for detecting single-trial visual evoked potentials (VEPs), enhancing brain-computer interface (BCI) performance. The approach significantly improves accuracy and enables real-time BCI applications.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Detecting single-trial visual evoked potentials (VEPs) is crucial for brain-computer interfaces (BCIs).
- Existing methods often lack speed and automation for real-time applications.
Purpose of the Study:
- To develop an automated and fast method for single-trial VEP detection.
- To implement this method in a high-performance, real-time BCI system.
Main Methods:
- Utilized common spatial pattern (CSP) for spatial filtering and wavelet filtering (WF) for temporal-spectral filtering.
- Combined CSP and WF to enhance the signal-to-noise ratio (SNR) of single-trial VEPs.
- Assessed the approach in a four-command VEP-based BCI system.
Main Results:
- Offline classification accuracy improved from 67.6% to 97.3% with CSP and WF filtering.
- Successfully implemented in an online BCI system, achieving 90% accuracy with 20 decisions per minute.
Conclusions:
- The proposed approach enables robust and reliable automatic VEP waveform detection.
- Applicable for real-time VEP-based online BCI systems.
- Offers a versatile solution for evoked potentials/event-related potentials (EPs/ERPs) detection in BCI and intraoperative monitoring.
