Related Experiment Video
Updated: Jun 6, 2026
10:14
Motor Imagery Performance Through Embodied Digital Twins in a Virtual Reality-Enabled Brain-Computer Interface Environment
Published on: May 10, 2024
Trial pruning for classification of single-trial EEG data during motor imagery
Boyu Wang1, Chiman Wong, Feng Wan
1Department of Electrical and Electronics Engineering, Faculty of Science and Technology, University of Macau, Av. Padre Tomás Pereira, Taipa, Macau.
Summary
This study introduces a novel genetic algorithm (GA) to detect abnormal electroencephalography (EEG) data in brain-computer interface (BCI) systems. The method significantly improves BCI performance by identifying and excluding misleading trials.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Electroencephalography (EEG) data artifacts degrade brain-computer interface (BCI) performance.
- Motor imagery BCI systems are susceptible to misleading trials from improper movement imagination, further degrading performance.
Purpose of the Study:
- To develop a novel algorithm for detecting abnormal EEG data using a genetic algorithm (GA).
- To improve the performance of motor imagery based BCI systems by addressing data artifacts and misleading trials.
Main Methods:
- A genetic algorithm (GA) was employed to detect and prune abnormal EEG trials.
- Common Spatial Pattern (CSP) and a Gaussian classifier were trained on the selected subset of EEG data.
Main Results:
- The proposed method was tested on Dataset IIa of BCI Competition IV.
- Significant performance improvements were observed for six out of nine subjects.
Conclusions:
- The novel GA-based algorithm effectively detects abnormal EEG data, leading to enhanced BCI performance.
- Trial pruning using the proposed method is a viable strategy for improving motor imagery BCI systems.
