【作者】 熊慧；

【导师】 李刚；

【作者基本信息】 天津大学 ， 精密仪器及机械， 2013， 博士

【摘要】 脑深部刺激技术能够绿色有效治疗帕金森病（PD）。但由于国外的技术垄断，使得设备费用很高；而且，目前临床应用脑深部刺激器（DBS）因电池容量有限，使用数年后需再次手术更换电池或刺激器，增加了患者的痛苦和经济负担。因此，对DBS整体设计方案的研究、开发终身使用、低成本的DBS势在必行。相比DBS的在临床上的迅速应用，其治疗机理并不十分明确。因此，强刺激信号作用下的微弱诱发脑电信号的检测与分析对于DBS的作用机理研究显得尤为重要，这不仅能进一步改进DBS，而且对研究和治疗各种神经类疾病具有重要意义。本文研究了终身使用的DBS，有两种设计方案--无线无源DBS和可充电式DBS，设计了强噪声下神经电信号采集系统，并运用不同的方法提取诱发信号。主要的工作及结果包括以下几个方面：1）创新性地提出了无线无源参数可调脉冲发生器的设计方案，实现了刺激脉冲各项参数的可调；体内刺激器部分不需任何电池，有望实现DBS的一次植入、终身使用，设计并制作了无线无源的DBS样机。2）首次将超级电容作为体内刺激器的储能单元，设计了其无线充电方案以及输出稳压电路；设计了基于MOS管的脉冲发生方案，使得刺激脉冲的各项参数易于调节且系统稳定性高，设计并制作可充电式DBS样机。3）针对强电刺激和微弱脑电信号相差6个数量级难于检测的难题，设计了高精度的信号采集系统，采用过采样技术和高分辨率与高速ADC器件实现超大动态范围信号采集，完成同步刺激等强噪声背景下大鼠脑深部电位的采集。4）刺激中诱发脑电信号淹没在强噪声与自发脑电信号中，信噪比很低。本文研究了基于线性与非线性理论的大鼠脑深部诱发电位的提取方法。提出了以提升小波处理后的信号作为自适应信号增强器的参考输入的方法提取诱发电位。得到的诱发电位精度较高，而算法运算速度快，耗时短，在实时处理上效果突出。就单导脑深部诱发电位的检测提出了采用经验模态分解与盲源分离相结合的方法。该方法不仅能处理非平稳信号的分离，而且可适用于源信号数多于观测信号数的分离等。更多还原

【Abstract】 Deep Brain Stimulating could cure Parkinson’s disease without causing damageto the brain tissues. But the monopolization of foreign DBS’s technology drives up theprice of this device. Moreover, due to the limit of battery, the Deep Brain Stimulator(DBS) include its battery will have to be replaced via surgery in clinical application.And all of this will increase patient’s suffering and economic burden. So research ofthe DBS overall design and development of lifetime use and low cost DBS isimperative.Compared to the DBS technology development, its therapeutic mechanism is notclear, the detection and research of weak evoked potentials signal under strongstimulation are particularly important for its mechanism study, as the result in thatthese could not only further improve the DBS, but also be of great significance for theresearch and treatment of neurological diseases.This paper proposed two designs for lifelong DBS: wireless passive DBS andrechargeable DBS, the design of the nerve signal acquisition system in strong noise ispresented too. Then rat EEG under stimulating is detected and the evoked potentialsare extracted from them.The main parts and achievements are as following:1) A design of wireless passive pulse generator was innovatively proposed, and itcould adjust parameters of the stimulation pulses. This design didn’t need battery invivo stimulation parts, so the DBS could be lifetime use after the first implant. Thewireless passive DBS prototype was developed.2) Super-capacitor was used as the energy storage unit in vivo, the wirelesscharging scheme and output voltage regulation circuit was designed. A pulsegenerator based on MOSFET was presented which made it easier to adjust theparameters of the DBS, and the system has a high stability. The rechargeable DBSprototype was produced.3) The high-precision signal acquisition system was developed for the detectionof weak EEG which differed six orders in magnitude from stimulation signal.Combining the oversampling technology and high-resolution, high-speed ADCdevices to achieve large dynamic range signal acquisition, thus the acquisition of the rat deep brain potential was completed under the strong noise which includessynchronization stimulation.4) The signal-to-noise ratio of stimulation evoked EEG is very low because itwas buried in strong noise and Spontaneous EEG. This paper studied the extractionmethods of deep evoked potential in rat brain under different requirements based onlinear and nonlinear theory. The algorithm utilized the signal de-noised by liftingwavelet transform as the input of adaptive signal enhancer to extract evoked potential.This method could take very short time to achieve a high precise, and it hadprominent effect on the real-time processing. What’s more, the method of combiningempirical mode decomposition method with blind source separation was raised toextract single channel deep brain evoked potentials. This method can not only be usedin the separation of non-stationary signal, but also be applied to the situation wherethere were more source signals than observed signals.更多还原