宽频脑电采集系统设计及其初步应用

发布时间：2018-01-30 06:29

本文关键词： 宽频脑电放大器动态直流校正技术 EEG-fMRI 电极材料信噪比眨眼眼电形态学滤波算法表面肌电　出处：《电子科技大学》2016年博士论文　论文类型：学位论文

【摘要】：脑电采集系统是一种对人体无创的电生理信号记录装置,该装置易于操作、实验成本相对较低,其记录到的头表脑电信号(Electroencephalogram,EEG)是大脑神经系统活动时的一种同步信息,因此脑电采集系统是科研工作者研究人类大脑功能最常用的工具之一。由于在EEG信号采集过程中,直流偏移电压对脑电采集系统的影响,以及人们对大脑认识的限制,致使当前市场上多数脑电采集系统带宽较窄,无法获取新发现的一些有生理意义的低频或高频信号。本文采用动态直流校正等技术,解决了当前脑电放大器存在的带宽有限问题。此外,本文也开发了具有兼容性和开放性的软件系统,并将脑电放大器和软件系统结合起来,应用到不同的领域,实现一机多用。本文具体的研究内容为:(1)根据脑电信号的特点,设计开发32通道的宽频脑电放大器。该放大器能够以5000Hz的采样率精确采集DC~1000Hz之间的脑电信号。我们在设计过程中采用动态直流校正技术,实时去除信号采集过程中引入的直流偏移电压,能够采集到传统脑电放大器丢失的低频信号;同时通过提高采样率,优化数据传输速率,拓展了放大器的高频带宽。放大器调试成功之后,我们按照国家相关测试标准,对脑电放大器性能进行测试。并根据测试结果,讨论分析电路中存在的问题,优化放大器设计,最终获得高性能宽频脑电放大器。利用微软基础类库MFC作为开发工具,设计开发一套软件系统。该软件系统与脑电放大器结合起来形成一套完整的脑电采集系统。(2)为了能够将自主设计开发的宽频脑电放大器与功能磁共振(functional Magnetic Resonance Imaging,fMRI)结合起来,获得高时-空分辨率的大脑活动信息。本文分别测试三种典型的脑电电极和三种典型的电极膏及其组合对MRI图像的影响,分析电极材料对MRI图像产生影响的内在原因,并找出适合同步数据采集的电极和电极膏。接着将挑选出的电极和电极膏分别放置在实验水膜和大鼠头部上,在磁共振设备中测试是否存在安全性问题。随后将符合要求的电极、电极膏与脑电放大器结合,在磁共振设备中采集脑电数据。(3)为了解决真实脑-机接口(Brain Computer Interface,BCI)应用中,样本数量较小时,可能引起的在线测试结果变差问题。我们对Z-LDA(Z-score Linear Discriminant Analysis)算法进行改进,然后分别利用仿真数据和自主设计的脑电采集系统采集的真实脑电数据,验证改进的Z-LDA(Enhanced Z-LDA,E Z-LDA)方法训练小样本的效果。(4)利用脑电放大器直接获取眨眼眼电信号,同时以(1)中的软件系统为基础,增加数字形态学滤波算法、动态阈值检测算法和归一化算法,识别主动眨眼信号,并利用眨眼信号在图形界面中实现字符的输入。(5)利用脑电放大器直接获取肌电信号,同时以(1)中的软件系统为基础,加入特征提取、模式识别等技术,识别肌电信号的特征,实现对机械手的控制。(6)分析直接获取拉普拉斯信号中存在的问题,以及间接获取拉普拉斯信号的缺点,探讨通过硬件电路直接获取头皮表面拉普拉斯信号的可行性。综上所述,本文设计了一款宽频脑电采集系统,不仅解决了当前脑电采集系统带宽较窄的问题,并且拓展该脑电采集系统的用途至眼电、肌电信号的采集中,提高了脑电采集系统的使用价值。
[Abstract]:The EEG acquisition system is a non-invasive human electrophysiological signal recording device, the device is easy to operate, the cost is relatively low, scalp EEG signal recorded by (Electroencephalogram, EEG) is a kind of synchronization information brain nervous system activity, so the EEG acquisition system is one of the research workers the study of human brain function. The most commonly used tools in EEG signal acquisition process, effects of DC offset voltage on the EEG acquisition system, and the people of the brain for understanding the limitations of the current market most EEG acquisition system with narrow bandwidth, unable to obtain new found some physiological significance of low or high frequency signal. This paper uses the dynamic correction of DC technology, solves the problems of the current limited bandwidth of EEG amplifier. In addition, this paper also developed the software system has compatibility and openness, and EEG The amplifier and the software system are combined, applied to different fields, to achieve a machine. The main contents of this paper are: (1) according to the characteristics of EEG signal, broadband EEG amplifier design. The development of 32 channel amplifier to the sampling rate of EEG signal acquisition DC~ 1000Hz 5000Hz accurately. We use dynamic DC in the design process of correction technology, DC offset voltage into real-time removal process of signal acquisition, can collect the traditional EEG frequency signal amplifier lost; at the same time by increasing the sampling rate, the optimization of data transmission rate, expand the bandwidth of a high frequency amplifier. The amplifier after successful commissioning, we in accordance with the relevant national test standard testing of EEG amplifier. According to test results, discusses the problems existing in the circuit, optimizing the amplifier design, finally obtains the high performance broadband EEG Amplifier. Using the Microsoft foundation class library MFC as the development tool to develop a software system design. The software system and EEG amplifier combine to form a complete set of EEG acquisition system. (2) in order to be able to independently design and development of the broadband EEG amplifier with functional magnetic resonance (functional Magnetic Resonance Imaging, fMRI) together, get high spatial resolution of brain information. This paper test the effect of three kinds of typical EEG electrodes and three typical electrode paste and its combination of MRI image, and analyze the internal reasons of electrode material affects the MRI image, and find out the electrode for the synchronous data acquisition and then the electrode paste. The selected electrode and electrode paste were placed in the experimental water film and rat head, whether the test in a magnetic resonance device security problems. Then to meet the requirements of the electric pole, electric A cream combined with EEG amplifier, EEG data acquisition in magnetic resonance device (3). In order to solve the real brain computer interface (Brain Computer Interface, BCI) application, the number of samples is small, may cause the deterioration of online test results. The Z-LDA (Z-score Linear Discriminant Analysis) was improved the algorithm, then using simulation data and independently designed EEG acquisition system to collect the real EEG data, verify the improved Z-LDA (Enhanced Z-LDA, E Z-LDA) training method of small sample results. (4) using EEG amplifier direct access to electrical signals at the same time to blink eyes, (1) in the software system based on increasing digital morphological filtering algorithm, dynamic threshold detection algorithm and normalization algorithm, identification and use the blink of an eye blink signal, signal character in a graphical interface input. (5) obtained directly by using EEG amplifier At the same time with the EMG signal, (1) in the software system as the foundation, with feature extraction, pattern recognition, feature recognition of EMG signal, to achieve control of the manipulator. (6) analysis of direct access to the existence of Laplasse signal problems and shortcomings of indirect access to the Laplasse signal, to explore the feasibility of direct access to the surface of the scalp Laplasse signal by hardware circuit. In summary, this paper introduces the design of a broadband EEG acquisition system, not only solve the current acquisition system with narrow bandwidth EEG, and expand the EEG acquisition system to use the EOG signal acquisition, EMG, EEG acquisition system improves the use value.

【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：R318;TN911.7

【相似文献】