Intelligence computing approach for seizure detection based on intracranial electroencephalogram (IEEG)

Epilepsy is a neurological disorder which causes two million people in the United States for suffering. In this research, we proposed a seizure detection method based on intracranial electroencephalogram (IEEG) by using signal processing and intelligence computing approach, which combines filtering, back propagation neural network (BPNN), multi-resolution Teager energy operator (MTEO), smooth window, and data fusion. The system overall performance for detecting seizures onset is achieved 74.9% accuracy, and its specificity and sensitivity reaches 87.7% and 47.3%, respectively.