| Neuromuscular disease is a chronic disease,and its early mild symptoms are easy to be ignored by patients.Because some neuromuscular diseases are irreversible and incurable,the diagnosis of neuromuscular diseases has always been a difficult problem in the medical field.Most of the clinical medical devices used to evaluate neuromuscular diseases have large volumes,high costs,and complex operations,and the detection process is often accompanied by ionizing radiation,which is not suitable for repeated detection of the same patient in a short time.Therefore,a convenient,fast,and safe portable testing instrument has great research and development value.Electrical Impedance Myography(EIM)is a non-invasive,painless,radiation-free,rapid,and low-cost muscle health assessment technology proposed in recent 20 years.It has shown broad application prospects in the diagnosis of neuromuscular diseases,especially in the evaluation of the progress of amyotrophic lateral sclerosis(ALS).This thesis mainly uses EIM to measure the dynamic complex impedance of human muscle tissue and establishes the relationship between muscle motion state and muscle complex impedance through the amplitude-phase diagram.The research content of this thesis is expected to provide a new solution for the development of low-power and portable human EIM detection and analysis instrument and the early diagnosis of neuromuscular diseases in clinics.This thesis mainly designs a human tissue dynamic complex impedance measurement system based on EIM.The whole system has low power consumption and is easy to carry.It can realize the real-time measurement of human muscle complex impedance.The complex impedance measurement system is mainly composed of a mirrored current source circuit,an IQ demodulation circuit,and a high-speed dualchannel acquisition system based on ARM+FPGA.The mirrored current source provides the high-frequency carrier signal for the human tissue to be tested.The complex impedance signal of muscle tissue is modulated by the four-electrode measurement system.The modulated signal is input to the IQ demodulation circuit for demodulation to separate the amplitude and phase information of the complex impedance of muscle tissue,and then the output signal is transmitted to the computer through the acquisition platform for real-time display.In this thesis,the reliability of the complex impedance measurement system is verified by measuring the change of complex impedance of human neck muscles during swallowing-related actions.In addition,this thesis constructs a swallowing action recognition model based on convolution kernel to recognize the amplitude and phase diagram of the complex impedance signal of human swallowing-related actions output by the complex impedance measurement system.This thesis proposes a feature extraction method based on convolution kernel,which innovatively applies a convolution kernel from the field of image feature extraction to a two-dimensional curve.Compared with PCA and LDA feature extraction methods,the convolution kernel can better extract the features of swallowing complex impedance signals.The parameters in the model are optimized by the PSO algorithm.The swallowing motion recognition model establishes the direct relationship between the swallowing signal amplitude and phase diagram and swallowing-related actions,which lays a foundation for the real-time recognition of muscle movements in the future.Through the analysis of swallowing data of 19 volunteers,it is proved that the system designed in this thesis can accurately measure the change of muscle complex impedance caused by the change of human muscle movement state,and show the relationship between muscle amplitude and phase with time and the amplitude and phase diagram of different movements.At the same time,the swallowing action recognition model is trained through the swallowing action amplitude and phase diagram,and the model finally achieves the recognition result with an accuracy of94.9%. |
PDF Full Text Request