Research And Analysis Of Functional Corticomuscular Coupling For Rehabilitation Assessment In Stroke Patients

Recently,the number of stroke patients caused by the brain lesions resulting in dysfunction in the limb has been on the rise,which brings a heavy burden to the family,healthcare institutions and society.To improve the function of motor control and sensory feedback,it is of theoretical and practical significance in rehabilitation field to develop and implement effective rehabilitation strategy.It is a key to make objective and precise assessment for rehabilitation strategy.The functional corticomuscular coupling(FCMC)is a physiological phenomenon to reflect the multilayered characteristics of the information interaction between the cerebral cortex and the corresponding muscle.The research of the FCMC can contribute to comprehend the cooperative function of the motor-sensory neural network,reveal the pathological mechanism in the motor-control system and provide a fresh perspective for rehabilitation assessment in stroke patients.The content in this thesis mainly focused on the FCMC research in the motor-control system.On basis of the theories of the dynamic signal processing such as information entropy,sorting patterns and coarse-graining processing,multiple synchronous methods were proposed to analyze the nonlinear,complex,multiscale and time-delay characteristics of the FCMC.Additionally,nonlinearly complex and multiscale models were also built to deeply analyze the local frequency-band and time-delay characteristics and expand the multiscale and cross-frequency coupling for FCMC.These researches excavated the intrinsic and potential mechanism of the FCMC.Furthermore,the correlation between the FCMC characteristics and Shang Tianmin(STM)assessment scale provided a theoretical foundation to the application of the FCMC in rehabilitation assessment for stroke patients.Above all,the innovative achievements in this thesis mainly contained the following aspects.(1)To analyze thoroughly the local frequency-band characteristics of the FCMC in motor-control system,a novel causality method,named transfer spectral entropy(TSE),was proposed to depict the local frequency-band characteristics of the FCMC from the view of both linearity and nonlinearity.In this method,two dimensional Fourier transform method was introduced to transform the multidimensional expansion of the spatial vectors in time-domain into spatial vectors in the frequency-domain.Then the transfer entropy was calculated based on the spatial vectors in each frequency bin,so as to achieve the information interaction between two signals.After that,the numerical simulation model possessing the multi-class local frequency-band was built.Based on the simulation data,the values of both TSE and granger causality method were calculated and compared to verify the effectiveness of the TSE method in describing the nonlinear coupling of narrow-band and broad-band signals.(2)To investigate the cross-frequency coupling(CFC)characteristics of the FCMC in motor-control system,the wavelet package transfer entropy(WPTE)method was proposed to describe the CFC characteristics of the information interaction between two signals possessing time-frequency scale.In this method,the wavelet package was used to decompose the signal into multi-group sub-band signals with different time-frequency scales,and the transfer entropy was also calculated within each pairwise sub-band signal.The effectiveness of this method was verified based on cross-frequency coupling and iso-cross frequency coupling models.(3)To explore the time delay characteristics of the FCMC in motor-control system,the time delay information transfer index(TDITI)method to quantitatively analyze the delay time in two directions between two signals with linear and nonlinear coupling.In this method,the sort patterns of the signal were ranked based on the permutation method,and then the sequence probability of the series with the same mode was distinguished based on the weight calculation.Then information transfer index was calculated for sort mode and accurate delay time estimation was obtained by calculating maximum IDIT index.Based on the linear and nonlinear time delay simulation model,the validity of the proposed method was verified by comparing with the existing methods.(4)To describe the multi-scale characteristics of information interaction between brain and muscle in motor control system,this thesis proposed complex multiscale coherence(CMSC)to quantify the synchronous coupling between two coupled signals at different scales.The multi-scale coupled signals were constructed based on the k-order coarse-grained process,and then the coherence analysis was performed on the signals at each scale to describe the multi-scale information interaction between two signals.On the other hand,a multiscale numerical model was constructed to evaluate the proposed method.(5)Last but not least,to investigate the nonlinear,complex,time-delay and multiscale characteristics of the information interaction between the brain and the muscle in control system for stroke patients,this thesis applied the TSE,WPTE,TDITI and CMSC methods to quantify the local frequency-band,cross-frequency coupling,time delay and multiscale characteristics based on the synchronization between the electroencephalogram(EEG)and electromyogram(EMG)signals for stroke patients and healthy controls.In addition,on basis of the Pearson correlation analysis between the FCMC indexes and the STM assessment scale for stroke patients,the changed trends of the indexes with high correlation with the STM scale were further analyzed.This research verified the function of the FCMC in the stroke rehabilitation,and provided a theoretical foundation to the application of the FCMC in rehabilitation assessment for stroke patients.