一种基于二代小波变换与盲信号分离的脑电信号处理方法

目的研究对混杂有眼电和心电干扰脑电信号的处理方法。方法首先用二代小波硬/软阈值、折衷阈值、μ律阈值方法对脑电信号消噪,然后运用FastICA算法对消噪后仍含眼电和心电的脑电信号进行盲信号分离。结果二代小波μ律阈值方法对脑电信号有较好的消噪效果,FastICA算法能成功分离出脑电中眼电和心电的干扰。结论运用二代小波μ律阈值法对脑电消噪后再用FastICA算法对独立源产生的干扰进行分离是一种有效的预处理方法。     

基于小波变换和K-means聚类算法的心电信号特征提取

目的研究一种基于小波变换和K-means聚类算法的心电信号特征提取方法,根据特征点信息判断心电是否正常。方法利用小波变换和形态学滤波方法去除工频干扰、肌电干扰和基线漂移等主要的噪声之后,利用K-Means聚类算法提取出心电信号的QRS波群,P波和T波这3个主要的特征点,实现心电智能诊断。结果实验数据取自MIT-BIH数据库,多次实验结果显示QRS波群的阳性检测度(P+)达到99.68%和灵敏度(Se)达到99.21%,P波和T波的检测准确度分别达91.43%和97.01%。结论相对于其它方法,本文心电特征提取方法准确度较高,具有一定参考价值;在移动医疗和临床医疗方面具有一定实用价值。     

基于独立分量分析的自适应最大熵算法对脑电干扰的识别与剔除

目的 提出一种在脑电信号采集过程中自动识别和剔除各种干扰的新方法.方法 将独立分量分析(ICA)最大熵算法和非线性参数阈值设定相结合.首先对ICA最大熵算法进行自适应改进,并将其用于对包含肌电、眼动等各种干扰的19导脑电信号进行独立分量分解;然后对各独立成分进行3个参数的非线性分析,通过设定的阈值,自动识别出其中的伪迹成分;在去除识别出的各伪迹的独立成分后,将其余独立成分反投影到头皮各电极处,得到去除干扰后较为纯净的脑电信号.结果 研究表明,基于盲源分离技术的自适应最大熵算法实现了脑电信号与其中所包含的多种干扰成分的分离,通过信号的重建可实现对不同来源干扰的剔除.结论 自适应最大熵算法是生理信号消噪研究中一种有潜力的方法.     

航天飞行中心电信号去噪方法研究

目的 寻找适合载人航天需要的心电信号去噪方法。方法 通过研究载人航天过程中心电信号的噪声特性，提出将小波变换和形态滤波级联的去噪方法。结果 通过实际应用和仿真对比实验结果表明，本文提出的将小波变换和形态滤波级联的方法，能有效地去除载人航天过程中心电信号的噪声。结论 将小波变换级联形态滤波的去噪方法应用到载人航天过程中是可行的。     

高采样率下简单整系数工频陷波器的设计

目的 解决心电信号分析中高采样频率下工频窄带陷波器设计难题.方法 本文提出了一种基于零极点抵消和插值理论结合的带阻滤波器的设计方法,设计出滤波器系数为简单整系数,带宽可任意调整,而且可满足高采样率下工频陷波的要求.同时为了消除窄带滤波器的吉布思现象,本文吸取前人的经验,提出了在信号线性段采用平均滤波的方法消除吉布思现象.结果 该方法设计的滤波器能有效去除高采样频率下50 Hz工频干扰,且系数为整数,实现简单.结论 基于零极点抵消和插值理论结合的带阻滤波器,可以有效地解决高采样率下窄带工频陷波器的设计难题,滤除了工频干扰信号,同时最大程度地保留了信号的有用成分,满足心电分析的要求.     

Elimination of ECG Baseline Drift Using Meyer Wavelet.

目的 提出Meyer多分辨率逼近的新方法,消除心电信号基线漂移.方法 带基线漂移的心电信号减去其在V9上的正交投影就基本消除了基线漂移.通过多分辨率逼近的数值分析,讨论了小波基的选择,正交投影的选择和数据长度的选择.对目前此类算法因缺少定量化分析方法而难以进行效果的比较,提出算法精度的定量化分析方法,构建算法效果比较的共同基础.结果 应用算法于MIT-BIH数据库中具有严重基线漂移的心电数据,采用Meyer小波,带基线漂移的心电信号在V9上的正交投影和基线漂移的差异为(0.3 ±3.6) μV.结论 应用Meyer小波在层次9上消除30 s左右ECG的基线漂移是合适的.     

应用Meyer小波消除ECG的基线漂移

目的提出Meyer多分辨率逼近的新方法,消除心电信号基线漂移。方法带基线漂移的心电信号减去其在V9上的正交投影就基本消除了基线漂移。通过多分辨率逼近的数值分析,讨论了小波基的选择,正交投影的选择和数据长度的选择。对目前此类算法因缺少定量化分析方法而难以进行效果的比较,提出算法精度的定量化分析方法,构建算法效果比较的共同基础。结果应用算法于MIT-BIH数据库中具有严重基线漂移的心电数据,采用Meyer小波,带基线漂移的心电信号在V9上的正交投影和基线漂移的差异为(0.3±3.6)μV。结论应用Meyer小波在层次9上消除30 s左右ECG的基线漂移是合适的。     

基于时频分布的表面肌电信号瞬时中值频率的研究

目的 采用时频分析法分析表面肌电信号瞬时中值频率。方法 采用平滑伪-魏格纳分布、乔伊-威廉姆斯分布、谱图和尺度图等方法,估计模拟和真实表面肌电信号的瞬时中值频率,讨论这些分布在估计肌电信号瞬时中值频率时的性能,并与条件平均频率进行比较。结果 模拟实验表明瞬时中值频率曲线比条件平均频率曲线光滑,相对误差较小。真实肌电信号实验显示,在肌肉收缩过程中,瞬时中值频率比条件平均频率的波动更明显。结论 肌电信号瞬时中值频率比条件平均频率更能真实和敏感地反映表面肌电信号频率成分的变化。     

周期平均法消除心电图中50Hz干扰

在使用一般的带道滤波器时都程度不等的有“振铃”现象,影响对心电波形的正确分析判断,我们研究实验了一种新的周期平均法技术,可以用它来消除心电波中50Hz(60Hz)的工频干扰而不引入“振铃”现象。周期平均法针对50Hz(或60HZ)的工频为周期性信号的特点,采用平均技术,方法简便,效果明显。可以很方便的应用于心电图和向量心电图的实时监测及数据分析处理。     

心算任务复杂度对脑电theta,alpha和beta波的影响

目的以1Hz为频宽单位比较两种复杂度心算任务下4～30Hz范围的脑电功率谱,分析其与文献中以经典脑电波段划分所得结果的关系。方法14名健康志愿者完成两种复杂度的心算任务,同步采集其脑电数据。对前额和顶叶区域内10个代表性通道预处理后的脑电数据采用Welch算法实现谱估计,相对基线期做归一化处理,获得两种复杂度心算任务下的相对功率谱密度,并逐一配对t检验每1Hz频宽单位内功率谱差异的显著性。结果前额和顶叶区域能体现心算任务复杂度差异的脑电信号精细频率范围分别为10～25Hz和9～25Hz(P0.05)。结论较高频段alpha波和较低频段beta波能较好地表征心算任务的复杂度,以1Hz为频宽单位对脑电功率谱的精细区间分析是对基于经典脑电节律划分分析的一种补充。     

基于小波变换模板的QRS检测方法

目的 有效提高QRS复合波的检测率。方法 ECG信号经过小波变换后，其不同频率成份会映射到变换结果的不同尺度上，而在这些尺度之间存在相互的联系。本文正是利用小波变换后尺度之间的相互关系，提出了一种基于墨西哥草帽小波变换不同尺度之间关系的ECG信号QRS复合波检测方法。该方法根据变换结果的特征点提取小波变换模板，并将这个模板用于ECG信号的QRS检测，可以在滤除噪声的同时，有效提高检测率。结果 经MIT／BIH标准心电数据库检测验证，QRS波正确检测率达到99．9％。结论 综合利用小波变换不同尺度进行非线性变换所得到的模板可以显著提高QRS复合波的检测率。     

表面肌电信号指标评价低负荷等长收缩诱发屈肘肌疲劳的敏感性和稳定性分析

目的:分析表面肌电(sEMG)信号指标评定低负荷等长收缩诱发运动性肌肉疲劳的敏感性和稳定性,为疲劳评价提供参考。方法:以15名男性青年志愿者为研究对象,先测试受试者屈肘关节最大自主收缩力矩(Maximal Voluntary Contration,MVC),令受试者右臂腕部拉住可在肘关节处产生20%MVC的重物,记录右侧肱桡肌和肱二头肌sEMG,至不能坚持负荷时结束实验。计算时域指标均方根振幅(RMS),基于傅里叶变换计算中值频率(MF)和平均功率频率(MPF),基于小波包分析计算中值频率(MDF)和平均功率频率(MNF),基于非线性分析计算Lempel-Ziv复杂度C(n)。评定上述指标在低负荷等长收缩诱发屈肘肌疲劳方面的敏感性和稳定性。结果:肱二头肌与肱桡肌sEMG的RMS随负荷实验持续时间延长表现出逐渐增加的趋势,而MF、MPF、MDF、MNF、C(n)则表现出逐渐减小的趋势。MF、MPF、MDF、MNF、C(n)与负荷运动持续时间之间具有显著负相关,RMS与负荷运动持续时间之间具有显著的正相关。肱二头肌与肱桡肌RMS反映疲劳的敏感性和稳定性优于其他指标。结论:由于低负荷收缩诱发疲劳引起募集运动单位数量增加等因素作用,RMS表现出较显著的增加趋势,是评定由低负荷等长收缩诱发屈肘肌疲劳的较好指标。     

一种抑制心电信号50 Hz工频干扰的改进Levkov方法

目的 探讨一种用于抑制心电信号中５０Ｈｚ工频干扰的新方法。方法 在Ｌｅｖｋｏｖ滤波法的基础上,加入ＱＲＳ开关函数,并用此法与几种常用的滤波方法进行比较。结果 该方法算法简单,处理速度快,对强、弱噪声滤波效果都较好。结论 该方法对于抑制心电信号中５０Ｈｚ工频干扰有很好的应用前景。     

Vibromyographic recording from human muscles with known fibre composition differences.

OBJECTIVE--To determine the relation between the vibromyographic (VMG) frequency characteristics and fibre composition in postural and non-postural human muscle undergoing a standardised voluntary contraction. METHODS--Two human muscles with different fibre compositions [soleus: postural, mainly type I (slow) fibres; biceps brachii: non-postural, mixed type I and II (fast) fibres] were recorded from 18 healthy males isometrically contracting at 50% of their maximum voluntary contraction (MVC). Muscle vibrations were recorded using a contact microphone and the frequency content of the signals calculated using fast fourier transform algorithms. RESULTS--The non-postural biceps brachii showed predominantly bimodal power spectra with significantly increased power in the 10-30 Hz bands (P < 0.01), as compared with soleus recordings which tended to be unimodal, with the majority of power below 10 Hz. CONCLUSIONS--Muscles with a large proportion of type I fibres generate VMG signals which contain an increased percentage of low frequencies as compared to muscles with a mixed population of type I and type II fibres. The VMG appears to be generated, in part at least, by the mechanical twitching of motor units within the muscle; frequency domain analysis of this signal may provide a non-invasive measure of muscle fibre composition.     

Oxygen uptake in whole-body vibration exercise: influence of vibration frequency,amplitude, and external load

Vibration exercise (VbX) is a new type of physical training to increase muscle power. The present study was designed to assess the influence of whole-body VbX on metabolic power. Specific oxygen uptake (sVO(2)) was assessed, testing the hypotheses that sVO(2) increases with the frequency of vibration (tested in 10 males) and with the amplitude (tested in 8 males), and that the VbX-related increase in sVO(2) is enhanced by increased muscle force (tested in 8 males). With a vibration amplitude of 5 mm, a linear increase in sVO(2) was found from frequencies 18 to 34 Hz (p < 0.01). Each vibration cycle evoked an oxygen consumption of approximately 2.5 micro l x kg(-1). At a vibration frequency of 26 Hz, sVO(2) increased more than proportionally with amplitudes from 2.5 to 7.5 mm. With an additional load of 40 % of the lean body mass attached to the waist, sVO(2) likewise increased significantly. A further increase was observed when the load was applied to the shoulders. The present findings indicate that metabolic power in whole-body VbX can be parametrically controlled by frequency and amplitude, and by application of additional loads. These results further substantiate the view that VbX enhances muscular metabolic power, and thus muscle activity.     

Comparison of localized proton NMR signals of skeletal muscle and fat tissue in vivo: Two lipid compartments in muscle tissue

In vivo ¹H NMR spectra of small volumes-of-interest (VOI) were localized in human soleus muscle (8 ml) and compared with volume selective spectra of subcutaneous fat tissue and femoral yellow bone marrow (2 ml). All examinations were performed by the double spin echo (PRESS) localization technique. To provide comparability, spectra of different tissues were recorded using identical sequence timing. Clearly improved resolution of the lipid signals of muscle tissue was obtained using long echo times TE > 200 ms. The spectra of muscle tissue exhibit lipid signals that stem from two compartments with a difference of their resonance frequencies of about 0.2 ppm (Larmor frequency difference 12-13 Hz at 1.5 T). The existence of two fatty acid compartments is supported by measurements of the relaxation times and line shape analysis. Both compartments contain fatty acids or triglycerides with similar composition. Probably one compartment corresponds to fat cells within muscle tissue, the other compartment with lower Larmor frequency is located within muscle cells.     

心电信号的实时数字预处理

本文根据心电实时监护的要求,应用整系数线性相位数字滤波器对心电信号作预处理,包括50Hz 工频干扰的陷波(带阻滤波)及高频噪声抑制(低通滤波)。文章叙述了整系数数字滤波器的设计过程,在实际滤波效果与计算量的减省方面均取得了较好的效果。     

Effects of steady state exercise on the power spectrum of heart rate variability

The effects of steady state exercise on the power spectrum of heart rate variability were studied in 19 healthy subjects. Continuous ECG signals were recorded during 1) 15 min of rest in the supine state, 2) 10 min of standing, 3) 10 min of steady state exercise at 50% of maximum predicted power output on a cycle ergometer, and 4) 15 min of post-exercise recovery in the supine state. Autoregressive modeling was used to determine the power spectrum of heart rate variability. While orthostatic stress produced a significant 51% increase in the ratio of low to high frequency peak spectral power, steady state exercise caused a significant suppression of both low and high frequency components. The low frequency peak power rose to significantly high levels throughout 15 min of the post-exercise recovery period. There was a significant leftward shift in the frequency of the low frequency peak with exercise and a rightward shift during the recovery supine state. These results suggest that neuroregulatory control of heart rate plays a major role in adaptive responses to orthostatic stress and post-exercise recovery, while humoral factors are probably more important in maintaining heart rate during steady state exercise.