G-3433

Epileptic Seizure Prediction Using EEG Signal Analysis and Simultaneous EEG Image Processing

 Sogand Abbasi Azizi ¹ © ℗, Reza Erfani Far ², Kaveh Samimi ³   

 
Abstract

Introduction: Epilepsy is one of the most common neurological disorders, characterized by uncontrolled seizures. Predicting seizures, especially prior to their occurrence, can play a crucial role in disease management and preventing injuries. One of the most important tools for seizure prediction is Electroencephalography (EEG) signals. EEG signal analysis combined with image processing techniques, particularly through machine learning and image processing methods, can significantly enhance seizure prediction accuracy. Methods and Materials: In this study, EEG data from 72 patients with epilepsy, collected from Rasoul Akram Hospital (2025), were analyzed. Initially, EEG signals were processed using filtering techniques and time-frequency feature extraction methods. Simultaneous EEG image processing was then performed by applying wavelet transforms and image processing techniques such as PCA (Principal Component Analysis) and CNN (Convolutional Neural Networks) to extract visual features from the EEG signals. For seizure prediction, machine learning models such as SVM (Support Vector Machine) and LSTM (Long Short-Term Memory networks) were employed. Results: The results indicated the superior performance of machine learning models in predicting epileptic seizures. The LSTM model achieved an accuracy of 93.5%, with a sensitivity of 91.2% and specificity of 94.1% (p 0.001), outperforming other models such as SVM and CNNs. Additionally, simultaneous EEG image processing using wavelet transforms and PCA significantly enhanced the feature extraction process, enabling the identification of abnormal brain activity patterns prior to seizure onset. Conclusion and Discussion: This study demonstrates that EEG signal analysis, combined with simultaneous EEG image processing, can be a powerful tool for predicting epileptic seizures. Machine learning models, particularly LSTM, were able to predict seizures with high accuracy, and these results could aid in the development of improved diagnostic and therapeutic tools for epilepsy patients. The application of these techniques in clinical settings could lead to more accurate seizure prediction and a reduction in the associated risks.

Keywords: Epilepsy, Machine Learning, EEG,CNN, Seizure