AI inspired EEG-based spatial feature selection method using multivariate empirical mode decomposition for emotion...

Area of Science:

• Neuroscience
• Artificial Intelligence
• Biomedical Engineering

Background:

• Electroencephalography (EEG) based emotion classification is crucial for healthcare and well-being.
• Noisy and non-linear EEG signals pose challenges for accurate emotion recognition and feature extraction.
• Existing methods often generate large feature vectors, increasing computational complexity.

Purpose of the Study:

• To propose an efficient spatial feature extraction and selection method for EEG-based emotion recognition.
• To reduce processing time and computational cost while maintaining high classification accuracy.
• To enable real-time emotion analysis for applications in healthcare and stress monitoring.

Main Methods:

• Intensive Multivariate Empirical Mode Decomposition (iMEMD) to decompose EEG signals into Intrinsic Mode Functions (IMFs).
• Complex Continuous Wavelet Transform (CCWT) for spatio-temporal analysis.
• Deep Neural Networks (DNNs) for feature extraction, combined with differential entropy and mutual information for feature selection.
• K-means clustering for dimensionality reduction.

Main Results:

• The proposed method achieves efficient feature extraction and selection with reduced dimensionality.
• Validated on SEED and DEAP datasets, demonstrating good classification performance.
• The system offers fast real-time sentiment analysis with lower computational expense compared to modern methods.

Conclusions:

• The developed method provides an efficient and cost-effective solution for real-time EEG-based emotion recognition.
• It addresses the challenges of noisy EEG data and large feature vectors.
• The approach holds significant potential for improving patient monitoring and mental healthcare applications.