An Intelligent Epileptic Prediction System Based on Synchrosqueezed Wavelet Transform and Multi-Level Feature CNN for...

Area of Science:

• Neurology
• Artificial Intelligence
• Signal Processing

Background:

• Epilepsy is a prevalent neurological disorder characterized by recurrent seizures, impacting millions globally.
• Current treatments manage seizures in approximately 70% of patients, but a cure remains elusive.
• Accurate seizure prediction is crucial for preventing injuries and enhancing patient quality of life.

Purpose of the Study:

• To develop an intelligent system for accurate and timely epileptic seizure prediction.
• To leverage advanced signal processing and machine learning for enhanced epilepsy management.
• To integrate this system into smart healthcare Internet of Things (IoT) networks.

Main Methods:

• Synchrosqueezed Wavelet Transform (SWT) was employed to analyze electroencephalogram (EEG) signals in the time-frequency domain.
• Multi-Level Feature Convolutional Neural Network (MLF-CNN) was utilized for extracting and classifying seizure-related features from processed EEG data.
• The system's efficacy was validated using the CHB-MIT and ZJU4H epilepsy datasets.

Main Results:

• The proposed system demonstrated high performance, achieving an accuracy of 96.99% on the CHB-MIT dataset and 94.25% on the ZJU4H dataset.
• Exceptional sensitivity (up to 97.76%) and specificity (up to 97.46%) were recorded across both datasets.
• A low false prediction rate (FPR/h) of 0.031 and 0.049 was achieved, indicating reliable seizure forecasting.

Conclusions:

• The developed intelligent system effectively predicts epileptic seizures using SWT and MLF-CNN.
• This approach offers a promising tool for proactive epilepsy management within smart healthcare environments.
• The system's high accuracy and low false prediction rate support its potential for clinical application.