神经网络算法在癫痫预测模型中的应用研究综述

Review of Application of Neural Networks in Epileptic Seizure Prediction

癫痫作为一种大脑神经元异常放电导致的中枢神经系统疾病,给患者的正常生活带来了极大影响,提前预测癫痫发作并及时采取防范措施可以有效提高患者的生活质量。随着数据科学和大数据技术的发展,神经网络算法越来越多地应用于癫痫预测领域,并展现出了巨大的应用潜力。对神经网络算法在癫痫预测领域的应用情况和不足之处进行了综述,按照癫痫预测模型的搭建流程依次从数据集、数据预处理、特征提取、神经网络算法模型几个模块进行论述。在介绍了脑电信号特点和常用数据集类别、常见数据预处理手段、常见的特征提取方法特别是手工设计特征的提取方法后,重点对多层人工神经网络和脉冲神经网络算法原理及其在癫痫预测领域的应用进行分析梳理和归纳总结,系统性地对神经网络算法的缺点进行剖析,并对神经网络算法在癫痫预测领域的进一步应用发展进行了讨论和展望。

Epilepsy,a central nervous system disease caused by abnormal discharge of brain neurons,has a significant impact on patients’normal life.Early prediction of epileptic seizures and timely preventive measures can effectively improve the quality of life of patients.With the development of data science and big data technology,neural networks are increasingly being applied in the field of epilepsy prediction and have shown great potential for application.This paper provides a review of the application and deficiencies of neural networks in the field of epilepsy prediction,discussing the construction process of epilepsy prediction models in the following order:datasets,data preprocessing,feature extraction,and neural networks.After introducing the characteristics of EEG signals,common types of datasets,common data preprocessing methods,and common feature extraction methods,especially manual feature extraction methods,this paper focuses on analyzing and summarizing the principles and applications of multi-layer artificial neural networks and spiking neural networks in the field of epilepsy prediction.The disadvantages of neural networks are systematically analyzed,and further application of neural networks in the field of epilepsy prediction is prospected.