基于Pinsky-Rinzel模型的输入输出线性化的癫痫状态控制

Dynamic control of seizure states with input-output linearization method based on the Pinsky-Rinzel model

本文的目的是设计一个输入输出线性化反馈控制器控制Pinsky-Rinzel(PR)模型的放电行为.PR模型能够模拟癫痫样式放电和正常放电,癫痫样式的行为由29Hz的高频簇放电模拟,6Hz的低频放电模拟正常的神经元放电.此研究将癫痫的治疗问题看做典型的控制问题,把胞体膜电位作为控制输出,树突作为受控对象.在这种控制策略下,可以得到稳定的期望的控制效果.同时与将控制加在胞体的控制效果进行比较,控制效果的平均控制误差,能量消耗和控制信号的波动性都比控制加在胞体时的数值小.笔者的发现证明了输入输出线性化对癫痫疾病控制的潜在价值,而且对选择树突作为癫痫治疗的靶目标有十分深远的意义.

This paper means to explore the use of input-output feedback linearization method to control the firing behavior of PR model. The epileptiform behavior is modeled by the high frequency burst of 29 Hz and low frequency spikes of 6 Hz represents normal desired behavior. The epilepsy treatment is formulated as a classic control problem with current control input. The output term is soma membrane potential, but the input term is dendrite. Under this control strategies, the stable and desired control effect is achieved. At the same time, we compare the result with which of soma input control, the average control error,energy expenditure and the standard deviation of control signal are larger than which in control law based on dendrite.Our findings demonstrate the potential value of input-output feedback linearization controller for epilepsy disease. This is also significant in selecting dendrite as the stimulation target for epilepsy treatment.