Multifuse multilayer multikernel RVFLN+ of process modes decomposition and approximate entropy data from iEEG/sEEG...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Epileptic seizures are neurological disorders characterized by abnormal brain activity.
• Accurate and timely detection of epileptic seizures from electroencephalogram (EEG) signals is crucial for patient management.
• Existing methods face challenges in effectively analyzing complex, non-stationary EEG data.

Purpose of the Study:

• To develop and evaluate a novel, highly accurate method for the automatic detection of epileptic seizure epochs.
• To combine advanced signal processing techniques with a robust classification model for improved seizure recognition.
• To validate the proposed method on diverse EEG datasets and compare its performance against existing approaches.

Main Methods:

• Extracted features using Variational Mode Decomposition (VMD) and Approximate Entropy (ApEn) from EEG signals.
• Employed a Multilayer Multikernel Random Vector Functional Link Network plus (MMRVFLN+) classifier for seizure epoch recognition.
• Utilized Bonn University iEEG and CHB-MIT sEEG datasets for experimental evaluation.

Main Results:

• The proposed VMD-HT-ApEn-MMRVFLN+ method demonstrated superior classification accuracy and overall performance.
• Achieved a remarkable recognition ability with a negligible false positive rate per hour (FPR/h).
• The method showed simplicity, feasibility, robustness, and practicability for automatic epileptic seizure detection.

Conclusions:

• The developed method effectively recognizes epileptic seizure epochs with high accuracy and reliability.
• The combination of VMD, ApEn, and MMRVFLN+ offers a promising approach for clinical EEG analysis.
• The FPGA implementation validates the practical effectiveness and efficiency of the proposed seizure detection system.