Automatic Epileptic Seizure Detection in EEGs Using Time-Frequency Analysis and Probabilistic Neural Network

The Electroencephalogram(EEG) has become one of the most important diagnostic tools in clinical neurophysiology since its first recording in humans performed in 1929. The traditional method of EEG analysis relies mostly on its visual inspection. Due to the fact that visual inspection is very subjective, several methods were proposed in order to quantify the information of the EEG. These methods include time domain methods, frequency domain methods, and time-frequency methods. This work discusses the appropriateness of Time Frequency Analysis(TFA) to classify EEG signals for epileptic seizures and compares several methods for time-frequency analysis. The methods are evaluated using three classification problems. The Probabilistic Neural Network (PNN) is used for classification of EEG segments. To the knowledge of authors, there exists no similar work in literature.