| Upper limb movements are often used as rehabilitation training for patients withdamaged movement function. However, patients could easily feel exhausted duringthe rehabilitation so as to discontinue the training. Therefore, the study of the corticalmuscle function coupling is carried on to help understand how brain controls muscletissue under various grip powers and in state of exhaustion, how muscular movementfeedbacks on brain function, as well as explain what specific physiological conditions,such as fatigue, originate from, so as to provide theoretical bases for the treatment ofmovement disorders and functional rehabilitation. Cortical-muscular functionalcoupling is defined as interaction, coherence, or time synchronism between cerebralcortex and muscle tissue. Through recording electroencephalograph (EEG),electromyogram (EMG) and grip signal of the tested at various grip force and in stateof fatigue, this project explores cortex muscle coherence as well as informationtransmission frequency and spatial characteristics.The paper, with target of specific hand grip function of fourteen healthysubjects, studies cortical-muscular functional coupling and information flow in twoaspects of coherence and causality of cortex muscle signals between cortex andmuscular tissues. The main study contains the following.1. The impact of various grip powers on corticomuscular coherence (CMC) wasexplored. By recording the EEG, EMG and grip signals of the subjectsunder variousgrip forces and in the state of fatigue, the amplitude and frequency feature of CMCbetween EEG of movement area and EMG of upper limb muscles was investigated.The result shows that grip power and status won’t affect amplitude of coherence, butobviously alter the distribution of coherence on frequency spectrum. As grip powerincreases, coherence compresses towards mid-low frequency band of beta wave band,while as the subjects get exhausted, coherence compresses towards the high frequencysegment of beta wave band and gamma wave band.2. Supporting vector regression expanded algorithm (SVRE) was proposed. InSVRE, the coherence of EEG-EMG for each trial is calculated and the coherences ofmultiple experiments are normalized at each frequency. Then the regression analysisis made for the normalized coherences using the conventional algorithm of supportvector regression (SVR). The mean of regression results is considered as the CMC at the corresponding frequency. With simulation data, SVRE is verified to be able toavoid impact of random disturbance on original signals. Compared with conventionalmethod, SVRE keeps in maximum extent the characteristics of CMC changingaccording to frequency and leads to more reliable and comprehensive result.3. In order to study signal transmission direction between cerebral cortex andmuscles, frequency spectrum causality analysis method is adopted as it reflects thetransmission characteristics of signal flowing on frequency spectrum, which isdifferent from the conventional time causality analysis method. The result shows EEGof motor cortex has strong causulity on EMG signal of upper extremity muscle ingamma frequency band (35-50Hz), while EMG signal has week causality on the EEGof motor cortex in low frequency band (3-35Hz). Besides the spectrum feature ofcausality has nothing to do with muscle contraction and status.In short, the functional coupling of cortical muscles and transmission ofinformation occur mostly in beta band and gamma band. Different status of voluntarymuscle contraction would change the spectrum distribution of coherence and has littleimpact on the causality. Causality value and its spectral distribution depend primarilyon the brain functional area and the information transmission direction. |
PDF Full Text Request