Abstract
Epilepsy is a disorder of the brain characterized by repeated seizures. About 1-3% of the people in the world suffer from epilepsy. Electroencephalogran(EEG) is the most useful test in confirming a diagnosis of epilepsy. However, diagnosing epilepsy is a difficult task requiring observation of EGGs and gathering of additional information. In this study, a new approach based on continuous wavelet and artificial neural network techniques was used to analyze EEGs for epileptic rats in order to detect the specific waveform correlated with seizure occurrence automatically. We examined EEGs with the animal model for eplipsy, the Wakayama epileptic rat(WER). WER is a new mutant exhibiting both spontaneuous absence-like behavior and tonic-clonic convulsions. EEGs were recorded from the rat which was in conscious and free to move. For long-term observation, EEGs were recorded on the vibrissal somatic sensory region by a small telemetry transmitter and a receiver system. The recorded EEG waveforms were band-passed between 1.5 and 30Hz. The seizure patterns in the telemetry EEG recording were analyzed to forcus on absence-like seizure. The present study indicates that a new hybrid system comprising continuous wavelet and artificial neural network techniques might be effective for autonomic absence-like seizure detection in long-term epilepsy monitoring. [J Physiol Sci. 2007;57 Suppl:S166]
