Related Experiment Video
Updated: Aug 15, 2025
Assessment of Audio-Tactile Sensory Substitution Training in Participants with Profound Deafness Using the Event-Related Potential Technique
Published on: September 7, 2022
An efficient deep learning framework for P300 evoked related potential detection in EEG signal
Pedram Havaei1, Maryam Zekri1, Elham Mahmoudzadeh2
1Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran; Medical Image and Signal Processing Research Center, Isfahan University of Medical Sciences, Isfahan, Iran.
This study introduces a novel deep learning framework, TGT-MHOG-CNN, for detecting P300 evoked related potentials (ERP) in EEG signals. The method achieves high accuracy and precision with improved efficiency compared to existing approaches.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Machine Learning
Background:
- P300 evoked related potential (ERP) detection is crucial for brain-computer interfaces (BCIs).
- Existing deep learning methods for P300 detection can be computationally intensive and time-consuming.
- There is a need for efficient and accurate P300 detection frameworks.
Purpose of the Study:
- To propose a novel deep learning framework, TGT-MHOG-CNN, for enhanced P300 ERP detection.
- To combine the strengths of Gabor Transform (GT), modified Histogram of Oriented Gradients (MHOG), and Convolutional Neural Network (CNN).
- To achieve high accuracy and efficiency in P300 detection from EEG signals.
Main Methods:
- Tuning Gabor Transform (GT) with a triangular function to capture time-frequency information.
- Modifying Histogram of Oriented Gradients (MHOG) to extract gradient patterns from 2D EEG signals.
- Fusing GT and MHOG features and feeding them into a CNN for precise P300 detection.
Main Results:
- The TGT-MHOG-CNN framework demonstrated superior performance on BCI Competition II and III datasets.
- Achieved classification accuracy exceeding 98.7% and precision over 98.7% for BCI Competition II.
- Reached 99% accuracy and 100% precision for BCI Competition III, with significantly reduced execution time.
Conclusions:
- The proposed TGT-MHOG-CNN framework offers an effective and efficient solution for P300 ERP detection.
- The fusion of GT and MHOG features enhances CNN's ability to accurately identify P300 components.
- This novel approach provides a promising advancement for BCI applications.
