Detailed Models of Network Bursts and Oscillations

描述（由申请人提供）：癫痫影响0.5-1%的人口，估计有1/3至近1/2的患者无法获得令人满意的癫痫发作控制。大多数患者有局灶性发作，可能继发。在癫痫病灶内，神经元的行为随时间和空间的变化而变化：这反映在神经元亚群中低振幅、高频（γ [30-70 Hz]或更高[> -70 Hz]）振荡的发生，就在更典型的癫痫样波开始之前，然后扩散。时空进化已经在体外和体内癫痫模型中重现，还有另一个不均匀性：跨皮质层的不同神经元放电模式。一个体外模型的>100赫兹振荡导致电痉挛也存在。更好地了解病灶内癫痫发作的详细病理生理学，然后是扩散，是很重要的：A)它可以改善即将发生的癫痫发作的早期预警，并提出中止癫痫发作的替代措施；B)它可能建议药物治疗，例如构成假定的轴突间隙连接的蛋白质的结合物，或所述连接的调节剂；c)它可以通过指出最基于生理的术前脑电图监测类型，例如，集中在通常不检查的频段，来改进手术候选人的术前评估。这一提议描述了我们如何首先将现有的单列网络模型用于研究非常快的振荡和癫痫发作，然后建立一个多列皮质网络模型。癫痫发作的时间演化可以用一列来研究，空间演化可以用多列来研究。我们使用详细的模型神经元元素，以及突触和间隙连接连接。这些模型是整合不同类型的实验测量（场电位，单细胞电和光学记录）的必要工具，并且可以解释a)多种细胞类型的内在特性，b)柱内突触和电连接，以及c)柱间突触连接对种群活动的贡献。一个成功的模型将产生对实验合作者有用的具体和可测试的预测，这些合作者过去一直在这个项目上密切合作，并将继续这样做。

DESCRIPTION (provided by applicant): Epilepsy affects from 0.5-1% of the human population, with an estimated 1/3 to nearly 1/2 of patients unable to attain satisfactory seizure control. A majority of patients have seizures of focal onset, possibly with secondary generalization. Within a seizure focus, the behavior of neurons evolves in time and space: this is reflected in the occurrence of low-amplitude, high-frequency (gamma [30-70 Hz], or higher [>70 Hz]) oscillations in a subpopulation of neurons, just prior to initiation of more typical epileptiform waves that then spread. Spatiotemporal evolution has been reproduced in in vitro and in vivo seizure models, as has another nonuniformity: different neuronal firing patterns across cortical layers. An in vitro model of >100 Hz oscillations leading into an electrographic seizure also exists. A better understanding of the detailed pathophysiology of within-focus seizure initiation, followed by spread, is important: a) it might allow improved early warning of an impending seizure, and suggest alternative measures to abort the seizure; b) it could suggest drug treatments, for example binders to proteins constituting putative axonal gap junctions, or modulators of said junctions; c) it could possibly refine the preoperative evaluation of surgical candidates, by indicating the most physiologically based types of preoperative EEG monitoring, for example, concentrating on frequency bands not typically examined. This proposal describes how we can first adapt an existing single-column network model for the study of very fast oscillations and seizure initiation, and then build a multi-column cortical network model. Temporal evolution of a seizure can be studied in one column, and spatial evolution in multiple columns. We use detailed model neuronal elements, along with synaptic and gap junctional connectivity. Such models are necessary as tools for integrating diverse types of experimental measurements (field potentials, single-cell electrical and optical recordings), and for accounting for the contributions to population activity made by a) intrinsic properties of multiple cell types, b) within-column synaptic and electrical connectivity, and c) between-column synaptic connectivity. A successful model will generate specific and testable predictions useful to experimental collaborators, who have worked closely on this project in the past and continue to do so.