(ICIOT 2024) Research on Automatic Detection and Early Warning of Epilepsy in Electroencephalogram Signals

The complexity of the time series of epileptic electroencephalogram (EEG) signals are characterized and described in different periods. For multi-channel, high-dimensional and heterogeneous EEG signals, considering their uncertain and dynamic characteristics, analyze and discriminate the EEG signals of epileptic patients in multiple periods by analyzing and mining basic data and with the help of deep learning theories and methods, and deeply mine the characteristics of epileptic EEG signals in multiple different periods. On this basis, focusing on the issues of automatic recognition, prediction and decision-making of epilepsy, mine the dynamic change laws of the automatic recognition process of epilepsy, break through the traditional automatic recognition, prediction and decision-making mode of epilepsy, and propose a new method of automatic recognition, prediction and decision-making of epilepsy based on the "multi-channel fusion - association mining - analysis prediction - intelligent decision-making" mode, and transform the collected information into diagnosis, prediction and further control of epileptic seizures. Clarifying the brain mechanism during epileptic seizures through EEG signals can provide more references and inspirations for the research of epilepsy detection.