【摘要】 癫痫是一种常见的神经系统疾病,其长期反复发作,使患者身心不断遭受伤害。若能提前预测到癫痫的发作,则可以使患者及时采取措施进行预防和保护。在癫痫发作预测研究中,动物模型可以提供丰富的样本,验证算法可行性,揭示癫痫发作的规律。本文制作了癫痫大鼠模型并建立了数据采集系统。利用同步性和复杂度两种非线性特征,对12次癫痫大发作进行分析,发现癫痫发作前同步性特征呈下降或上升趋势,复杂度特征则呈与同步性相反的变化趋势。这种变化在发作前几分钟出现,分析结果验证了癫痫发作的可预测性。更多还原

【Abstract】 Epilepsy is a common chronic neurological disorder.The reiterative seizure onsets bring severe hurts to the body and mind of patients.If seizures can be predicted even a few minutes before their onsets,this might offer sufficient time to take preventive measures to protect the patient from sudden injury.In the research of epileptic seizure prediction,animal models can provide sufficient data so as to validate the proposed prediction algorithms and discover the rules of epilepsy.In this paper,an epileptic rat model and the data acquisition system were set up.Nonlinear measures, amplitude synehrony and second-order complexity,were utilized to predict epileptic seizures through analyzing the EEG of epileptic rats.The data set contains 12 seizure onsets.The changes of synchrony and complexity measures could be detected from EEG during pre-seizure periods-decrease or increase,and the variation trends of the two measures were always opposite.The trends all appeared several minutes before seizures,which indicated the predictability of the epileptic seizures.更多还原