Convolutional neural network based on recurrence plot for EEG recognition.

Area of Science:

• Neuroscience
• Signal Processing
• Machine Learning

Background:

• Electroencephalogram (EEG) signals are crucial physiological indicators.
• Accurate classification of EEG signals has significant implications for health and safety.
• Existing methods for EEG signal classification face challenges in accuracy and efficiency.

Purpose of the Study:

• To develop a novel and effective method for classifying EEG signals.
• To utilize recurrence plots and convolutional neural networks (CNNs) for enhanced EEG signal analysis.
• To validate the proposed method on epileptic and fatigue driving EEG signals.

Main Methods:

• EEG signals were transformed into recurrence plots.
• A convolutional neural network (CNN) framework was constructed for signal classification.
• The method was evaluated on epileptic EEG data (normal vs. seizure) and fatigue driving EEG data.

Main Results:

• The proposed framework achieved high accuracy in distinguishing normal and seizure states in epileptic EEG signals.
• The method demonstrated good classification accuracy for fatigue driving EEG signals.
• Comparative analysis showed significant improvements over existing EEG detection methods.

Conclusions:

• The combination of recurrence plots and CNNs offers a powerful approach for EEG signal classification.
• This novel method significantly enhances the accuracy of detecting critical brain states like epilepsy and fatigue.
• The findings suggest a promising direction for developing advanced EEG-based diagnostic and monitoring systems.