Akademik Veri Yönetim Sistemi
SIGNAL IMAGE AND VIDEO PROCESSING, cilt.19, sa.4, 2025 (SCI-Expanded, Scopus)
Epilepsy is a neurological disorder characterized by recurrent seizures caused by abnormal neuronal activity in the brain. It can manifest at any age, from childhood to adulthood, with symptoms varying significantly among individuals. Common symptoms include muscle spasms, loss of consciousness, and fainting, which can profoundly impact daily life and routine activities. While epilepsy may persist throughout life, appropriate treatment and medication can often control seizures, enabling most individuals to lead normal lives. Electroencephalography (EEG) signals are a crucial tool for monitoring epileptic seizures and analyzing their underlying patterns. In this study, a novel feature extraction model based on high-frequency oscillations (HFOs) is proposed for the detection and classification of epileptic seizures using intracranial EEG signals. Unlike conventional seizure detection approaches frequently reported in the literature, this study specifically focuses on identifying distinct seizure phases. Experimental results indicate that the proposed model achieved an accuracy of 77% for HFO-based features and 83% for common features in a multi-class classification task using the Random Forest classifier. Notably, HFO-based features demonstrated robust and consistent performance across classifiers, underscoring their potential as reliable markers for seizure phase detection. Additionally, phase-level analysis revealed that the prodromal phase exhibited the highest detection accuracy, suggesting its utility for early seizure prediction, while the postictal phase showed the lowest accuracy due to structural similarities with other phases. These findings emphasize the feasibility of automatically identifying seizure phases, offering valuable insights and a solid foundation for future research in epilepsy diagnosis and management.