【作者】 张晓娜；

【导师】 万红；

【作者基本信息】 郑州大学 ， 检测技术与自动化装置， 2013， 硕士

【摘要】 初级视觉皮层(V1区)是大脑皮层处理视觉信息的第一站,深入分析V1区信号的响应特性对视觉信息处理机制的研究具有深远意义。在该领域的研究大都基于动作电位(spike)和局部场电位(LFP)两种信号。然而spike在检测时易受噪声干扰而造成漏检和误检等问题,严重影响后续分析。因此基于LFP响应特性的研究逐步受到学者们的亲睐。已有研究表明LFP中携带视觉刺激的基本特征等信息,然而对于LFP中具有特征调谐作用的频带范围却没有达成共识。因此准确提取LFP响应频带对进一步从LFP分析视觉皮层信息处理机制奠定了基础。据此,本文针对LFP信号的非平稳性,基于希尔伯特黄变换(HHT)提取其响应特征频带,研究了其对空间频率和朝向刺激特征的响应调谐特性,并与小波变换进行了对比。主要研究内容如下：1.信号采集与信号分析。采集并记录了麻醉LE大鼠V1区的spike和LFP信号,研究了LFP信号的特征及常用分析方法,为研究LFP特征调谐性能奠定了先验知识基础。2.基于小波变换研究LFP的响应调谐特性。运用小波分解提取LFP的响应频带Gamma频带,研究了Gamma频带对空间频率和朝向视觉特征的响应特性,并且与同一电极记录到的spike信号进行了对比。结果表明Gamma频带能量对不同视觉刺激特征具有不同的响应,与spike信号的调谐特性具有较好一致性。但是,小波变换结果依赖于小波基函数的选择,这是小波变换分析信号的难点和缺点所在。3.基于HHT方法研究LFP的响应调谐特性。该方法对LFP的整个频带的调谐性能都进行了研究,而不局限于对LFP局部频带的研究。首先将LFP分解为若干固有模态分量(IMF),据此分析了各阶固有模态分量对空间频率和朝向视觉刺激的响应调谐特性。结果表明：相比其他IMF,第二阶固有模态分量对视觉刺激特征的调谐特性最强,并且与spike信号的调谐特性具有较好一致性。该方法避免了传统小波算法寻找合适小波基的麻烦,更加准确的给出了响应特征,为LFP信号的特征提取及后续研究提供了有效手段,具有较高的推广价值。4.对基于小波变换和HHT两种方法得到的结果进行对比,分析了两种方法对空间频率和朝向光栅刺激的调谐性能。结果表明基于HHT方法得到的调谐指数的均值高于小波变换。因此,基于HHT方法在LFP响应频带特征提取上更有优势。更多还原

【Abstract】 The primary visual cortex is the first station of visual processing mechanism in cerebral cortex, and it is significance to in-depth analysis of response characteristics of signals for study visual processing mechanism in primary visual cortex. Most of the researches were based on the activity potential (spike) and local filed potential (LFP) recorded from primary visual cortex. However spike is susceptible to noise interference when it was detected, which resulting in missed and false detection, and the subsequent analysis was affected seriously. Therefore, research of response characteristics based on LFP was gradually studied by researchers. Studied have shown that visual stimulation was encoded by LFP, however, no coherent conclusion exists about which frequency ranges exhibit feature tuning. Extracting of response ranges of LFP has far-reaching significance to study visual processing mechanism.Addressing this, the response characteristic frequency bands of LFP which is non-stationary signal were extracted based on Hilbert-Huang Transform (HHT) in this paper, and response tuning characteristics underlying the stimuli of spatial frequency and orientation were researched, and wavelet transform was compared. The main contents are as follows:1. Signal acquisition and signal analysis. Spike and LFP signal were collected and recorded from anaesthetic Long Evans rat, and the characteristics of LFP and analysis methods used commonly were expounded, which laid the foundation for the study of tuning characteristics of LFP.2. Response characteristics of local field potential based on wavelet transform was researched. We studied the response characteristics for spatial frequency and orientation according to the Gamma band of LFP which extracted by wavelet decomposition, and the spike recorded from the same electrode was compared. The results showed that the Gamma band reflected different response to different visual stimuli, and it was consistent with spike. However, wavelet transform depends on the choice of the wavelet function which is the difficulties and disadvantages of wavelet transform analysis. 3. Response characteristics of local field potential based on Hilbert-Huang transform was researched. In contrast to previous studies, which often reported correlates of visual processing only in a limited frequency range of LFP, here the entire frequency range of LFP were studied by HHT in visual processing. The response characteristics for spatial frequency and orientation were studied based on the intrinsic mode functions of LFP which decomposed by HHT. The results showed that the second intrinsic mode function reflected response to visual stimuli compared with the other intrinsic mode functions, and it was consistent with spike. The method avoids the trouble of the wavelet transform which requires a proper wavelet basis, and a more accurate response characteristic was obtained, providing an effective means for feature extraction and the follow-up study of LFP, and it has a high value of promotion.4. The tuning characteristics of LFP underlying the stimuli of spatial frequency and orientation based on wavelet transform and HHT was compared. The results showed that the average tuning index obtained from HHT is higher than wavelet transform. Therefore, extracting response band of LFP based on HHT is more advantages.更多还原