表面肌阻抗混合信号的盲源分离电特性提取方法  

Blind Source Separation Electrical Characteristics Extraction Method of Surface Electrical Impedance Myography Mixed Signals

作　　者：严洪立 黄林南 高跃明[1] Željka Lučev VASIĆ Mario CIFREK YAN Hongli;HUANG Linnan;GAO Yueming;Željka Lučev VASIĆ;Mario CIFREK(College of Physical and Information Engineering/Fujian Key Lab of Medical Instrumentation and Pharmaceutical Technology,Fuzhou University,Fuzhou 350108,Fujian,China;Faculty of Electrical Engineering and Computing,University of Zagreb,Zagreb 10436,Croatia)

机构地区：[1]福州大学物理与信息工程学院/福建省医疗器械和医药技术重点实验室,福建福州350108 [2]萨格勒布大学电气工程与计算学院,萨格勒布10436

出　　处：《华南理工大学学报（自然科学版）》2022年第12期142-150,共9页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家重点研发计划项目(2022YFE0115500);科技部政府间交流例会项目(国科外[2019]16号);福建省科技计划项目(2021I0005)。

摘　　要：表面肌阻抗图(sEIM)是肌肉失衡与肌肉疾病状态评估中的重要手段。表面电极获取的皮下多层组织阻抗混合信号包含众多冗余成分。为了提升sEIM检测对目标肌肉状态变化的敏感性,文中将sEIM获取的混合信号作为盲信号,肌肉层阻抗值作为源信号,提出了一种基于阻抗等效分析和盲源分离的肌肉层阻抗分离方法。首先,建立肢体多层圆柱体有限元模型,采用灵敏度方法计算各组织层的阻抗贡献,用于排除冗余微弱信号,并将其等效为以肌肉为目标组织的盲源分离问题;然后,通过数值仿真和在体实验,比较了独立成分分析法、主成分分析法和等变化自适应独立分离法(EASI)的分离效果,获得最优方案并验证方法的可行性。结果显示,采用EASI分离肌肉层阻抗的方法,相关系数大于0.98,抗噪性约为0.8,串音误差收敛于0.876,在体实验中分离的肌肉层阻抗值符合人体阻抗特性规律,表明采用EASI的肌肉层阻抗分离方法能够较好地分离sEIM中肌肉层阻抗值,可用于提升检测目标的肌肉状态变化敏感性。Surface electrical impedance myography(sEIM)is important for evaluating muscle imbalance and muscle diseases.The mixed signals of impedance in the subcutaneous multilayer tissue captured by surface electrodes contain diverse-redundant components.Taking the mixed signals captured by sEIM as the blind signals and the muscle layer impedance as the source signal,this paper proposed a method for separating muscle layer impedance based on the impedance equivalent analysis and blind source separation,in order to improve the sensitivity of sEIM detecting changes in target muscle state.Firstly,a limb multilayer cylindrical finite element model for simulation was constructed.Secondly,a sensitivity method was employed to calculate the impedance contribution of each tissue layer for excluding redundant weak signals,which was equated to a blind source separation problem targeting the muscle layer.Finally,the separation effects of independent component analysis,principal component analysis,and equivariant adaptive separation via independence(EASI)were compared by numerical simulation and in vivo experiments to obtain the optimal solution and verify its feasibility.The results show that the correlation coefficient is above 0.98,the noise immunity is approximately 0.8,and the error cross talking(ECT)converges to 0.876 using the EASI-based method for separating muscle layer impedance.The separated muscle layer impedance in the in vivo experiments is consistent with the law of human impedance characteristics,indicating the method for separating muscle layer impedance with EASI can better enhance the sensitivity of sEIM to detect changes in the target muscle state.

关 键 词：肌阻抗图 生物电阻抗 盲源分离 有限元法 肌肉 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]