【摘要】 神经群模型可用于模拟产生癫痫发作各阶段的脑电(EEG)信号,在癫痫发作机理分析与控制的研究中起了重要的作用。本文主要利用闭环反馈控制的方法对神经群模型中的癫痫状棘波进行控制,以期为临床治疗癫痫时刺激位置和刺激大小的确定提供一定的理论依据。针对实际测量中不可避免的测量噪声,在反馈回路中加入无迹Kalman滤波(UKF)环节来估计模型的状态,UKF控制器可由估计的状态构造得到。仿真中分别将控制作用于异常兴奋的神经群及将控制作用于每个神经群,结果表明:两种UKF控制方案均能抑制神经群模型中的癫痫状棘波,但是,当控制作用于每个神经群时所需的控制能量比控制只作用于异常兴奋的神经群时所需的控制能量小。更多还原

【Abstract】 Neural mass models are able to produce epileptic electroencephalogram(EEG) signals in different stages of seizures.The models play important roles in studying the mechanism analysis and control of epileptic seizures.In this study,the closed-loop feedback control was used to suppress the epileptic-form spikes in the neural mass models.It was expected to provide certain theory basis for the choice of stimulus position and parameter in the clinical treatment.With the influence of measurement noise taken into account,an unscented Kalman filter(UKF)was added to the feedback loop to estimate the system state and an UKF controller was constructed via the estimated state.The control action was imposed on the hyper-excitable population and all populations respectively in simulations.It was shown that both UKF control schemes suppressed the epileptic-form spikes in the model.However,the control energy needed in the latter scheme was less than that needed in the former one.更多还原