基于动态模式识别方法的BCI康复系统

发布时间：2018-06-02 03:32

本文选题：脑机接口系统 + 脑卒中后遗症康复　；参考：《上海交通大学》2014年硕士论文

【摘要】：脑卒中症是近年来人类的高发病症，其后遗症所导致的脑神经损伤，会严重影响患者日常生活。基于运动想象的功能性电刺激康复系统能将患者的运动想象与实际的肢体运动建立起神经回路，是近年来脑机接口系统领域研究的一个重要课题。在实际应用中，我们发现脑卒中患者的运动想象信号与正常人的信号有着许多不同，因此我们在正常人的脑电信号基础上，根据脑卒中患者的信号特点，试图采用动态模式识别的方法来识别脑卒中患者的脑电信号。本文设计了一个BCI康复系统，，并通过临床试验采集BCI康复数据。我们比较了脑卒中患者与正常成年人的运动想象信号，找出其异同点，并针对患者脑电特点，提取出符合其特点的动态特征。我们将这些动态特征应用于几种流行的动态模式识别方法，通过比较找出适合患者脑电信号的模式，验证患者脑电特点,并对这些动态识别方法进行比较。
[Abstract]:Stroke is a high incidence of human disease in recent years, its sequelae caused by the brain nerve damage, will seriously affect the daily life of patients. The functional electrical stimulation rehabilitation system based on motor imagination can establish neural circuits between patients' movement imagination and actual limb movement, which is an important subject in the field of brain-computer interface system in recent years. In practical application, we found that there are many differences between the signal of motion imagination of stroke patients and that of normal people, so we based on the EEG signals of normal people, according to the characteristics of the signals of stroke patients. This paper attempts to identify the EEG of stroke patients by dynamic pattern recognition. A BCI rehabilitation system is designed, and BCI rehabilitation data are collected through clinical trials. We compared the motor imagination signals between stroke patients and normal adults, found out the similarities and differences, and extracted the dynamic features according to the characteristics of EEG. We apply these dynamic features to several popular dynamic pattern recognition methods. By comparing and finding out suitable patterns for patients' EEG signals, we verify the characteristics of patients' EEG, and compare these dynamic recognition methods.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.7;R743.3

【参考文献】