Physiological experiments have shown that inhibitory interneurons can induce and maintain epileptiform activities, and different interneuron subtypes may be responsible for different types of seizures. Here we aim to link these electrophysiological experimental phenomena with theoretical dynamic mechanisms based on an improved Liley mean field model. Fascinatingly, the number of synapses between inhibitory neural populations can induce a rich state transition. Typically, the system will experience normal rhythm discharges, multispike discharges, spike wave discharges (SWD), generalized periodic discharges (GPD), the beta band oscillation, and eventually return to the normal state. Interestingly, the transition process can also be reversed. Meanwhile, disinhibition circuits can cause more epileptiform activities after taking into account the delay of synaptic information transmission. Furthermore, we are committed to designing different control strategies for epileptiform activities. As we expected, both deep brain stimulation and optogenetic technology can destroy or even eliminate pathological waves. It is need to emphasis that the cell type specificity of optogenetic regulation allows it to precisely target inhibitory neural populations, which is agree with experiment. Particularly, three different optogenetic regulatory strategies targeting inhibitory neural populations are modeled and proposed, of which the intermittent is designed to save energy. These modeling results reproduce the experimental phenomena and more significantly, help understand the mechanism of epilepsy to guide the clinical practice.

生理实验表明，抑制性中间神经元可以诱导和维持癫痫样活动，不同的中间神经元亚型可能导致不同类型的癫痫发作。在这里，我们旨在将这些电生理实验现象与基于改进的 Liley 平均场模型的理论动态机制联系起来。令人着迷的是，抑制性神经群之间的突触数量可以诱导丰富的状态转换。通常，系统会经历正常节律放电、多尖峰放电、尖峰波放电（SWD）、广义周期性放电（GPD）、β波段振荡，最终恢复正常状态。有趣的是，过渡过程也可以逆转。同时，考虑到突触信息传递的延迟，去抑制电路可以引起更多的癫痫样活动。此外，我们致力于为癫痫样活动设计不同的控制策略。正如我们所料，深部脑刺激和光遗传学技术都可以破坏甚至消除病理波。需要强调的是，光遗传学调控的细胞类型特异性使其能够精确靶向抑制性神经群体，这与实验一致。特别是，针对抑制性神经群体的三种不同的光遗传学调控策略进行了建模和提出，其中间歇性旨在节省能源。这些建模结果再现了实验现象，更重要的是，