Analysis And Control Of Parkinson's Disease Based On Model

Parkinson's disease (PD) is a degenerative disorder of central nervous system which endangers the olds'health seriously. However, the cause and treating mechanism of PD remain uncertain. Hence, the research and clinic of PD is a challenge in neuroscience. By combining the computational model and the control theory, we propose a Deep Brain Stimulation (DBS) waveform shape optimization strategy and a close-loop control scheme to explore the mechanisms of PD and how DBS waveform influences PD neural network based on the computational model.First of all, we analyze the latest studies on PD and DBS, propose a PD neuronal network model exposed to the external electrical field to simulate the modulation mechanism of DBS on the brain region relating to PD. Based on this model, we introduce the relay reliability of thalamic cells (TC) and target at studying the firing patterns of this model exposed to external electrical field for both single neuron and neural network. It is discovered that DBS can reduce the inhibitory input from globus pallidus interna (GPi) neurons to TC neurons, and thus restore the relay reliability of TC neurons. When stimulating subthalamic nucleus (STN) with different DBS waveform, different nucleus of basal ganglia will show different rhythms. The results demonstrate the modulating effectiveness of external electrical field on PD neural network model.Because DBS STN with different parameter sets would induce different responses in nucleus of basal ganglia and finally take effect on the thalamus through GPi, we propose the concept of thalamic selectivity in response to different input. We simulate inhibitory input from GPi neurons to thalamic cells and an excitable input from sensorimotor cortical with some simple waveforms. This method ignores the complexity of the neural network and focus on the response of thalamic cell. When the inhibitory input from GPi neuron and the excitable input from sensory motor cortical are injected into the thalamic cell, the thalamic activities depend on the competition between these two inputs. Analysis results show that when a fixed excitable input is applied, TC will response to this input under some inhibitory projection. The relay ability is reserved under the high-frequency inhibitory input with low-amplitude and some inputs with high-amplitude.This thesis proposes the optimization of DBS waveform shape to address the problem that the optimal energy consumption cannot be obtained by the proposed method to choose DBS waveform. Firstly, we apply a particle swarm optimization to select the parameters of the periodic DBS waveform. This optimization algorithm can reduce the power consumption and selective time. Then, based on the area equivalent principle of pulse width modulation (PWM), we firstly propose a new rule to generate DBS waveform. Simulation results show that this waveform would take equivalent effect on periodic square wave and can reduce power consumption easily. We also discover the minimum energy effect of DBS during optimization.Since lack of understanding the mechanism of DBS and the individual difference, parameters of DBS cannot be adjusted automatically. Thus, a close-loop control scheme of PD model based on the identified model is proposed. We design a DBS control scheme for single-compartment model and multi-compartment model of thalamic relay cells using terminal sliding mode. Simulations of the theory model and the identified model of thalamic relay cells are made. Results show that terminal sliding mode controller can restore the relay ability of thalamic relay cells and eliminate the tremor in Parkinsonian state.The obtained results can not only be directly used for designing the close-loop DBS controller, but can also provide theory foundation for understanding the mechanisms of external electrical field modulates the brain neural network.