CRCNS: Understanding Single-Neuron Computation Using Nonlinear Model Optimization

这项建议的长期目标是将实验设计和计算 大规模参数估计以促进对神经科学中的一个中心问题的理解：如何 以及离子通道在树突室中的亚细胞定位是否有助于单个 神经元计算。这个问题只能使用其生物物理特性是 具有良好的特点，在感官信息的处理和行为的产生中的作用 是很好理解的。重点将放在三种在突触整合中发挥关键作用的通道类型上： 超极化激活、环核苷酸门控、混合钠/钾通道、瞬时树突 钾通道和钙通道参与爆发式放电的产生。具体目标将是 重点(I)确定这些通道，特别是钙通道如何促进树突状细胞 海马锥体细胞的兴奋性；以及(Ii)确定相同的通道如何参与 碰撞检测神经元中的视觉对象分割。此外，该项目还将(Iii)开发一个 更广泛、更集成的大规模建模优化框架，以研究渠道本地化的影响 关于树状结构的计算。这一框架在所研究的两个系统中的应用将使(四) 将通过模型优化获得的信道分布与实验得出的信道分布进行比较，从而 阐明它们在神经元信息处理中的作用。最终的具体目标将是(V)传播 新开发的优化方法面向更广泛的受众，允许广泛应用最新状态- 神经科学研究中的艺术数学知识。该项目将应用高级数学 基于多次射击或一次配置的二阶优化算法 与所建模的神经元相关的动力系统的离散化。该项目还将使用 用于确定亚细胞通道的电生理和免疫染色解剖技术 试验性的本地化。总体而言，该项目将有助于促进以下方面的基本知识 亚细胞通道定位如何有助于单个神经元内的信息处理。

The long-term objective of this proposal is to integrate the design of experiments and computational large-scale parameter estimation to advance the understanding of a central problem in neuroscience: How and whether the subcellular localization of ion channels in dendritic compartments contributes to single neuron computation. This problem can only be addressed using neurons whose biophysical properties are well characterized and whose role in the processing of sensory information and the generation of behavior is well understood. The focus will be on three channel types that play a key role in synaptic integration: hyperpolarization-activated, cyclic nucleotide-gated, mixed sodium/potassium channels, transient dendritic potassium channels, and calcium channels involved in the generation of burst firing. The specific aims will focus (i) on determining how these channels, in particular calcium channels, contribute to the dendritic excitability of hippocampal pyramidal cells; and (ii) on determining how the same channels contribute to visual object segmentation in collision detecting neurons. Additionally, the project will (iii) develop a broader, integrated large-scale modeling optimization framework to study the impact of channel localization on dendritic computation. The application of this framework in the two systems studied will allow (iv) to compare channel distributions obtained by model optimization to experimentally derived ones, thus shedding light on their role in neuronal information processing. A final specific aim will be (v) to disseminate the newly developed optimization methods to a broader audience allowing the wide application of state-ofthe- art mathematical knowledge in neuroscience research. The project will apply advanced mathematical methods centered on second-order optimization algorithms based on multiple-shooting or a collocation discretization of the dynamical system associated with the modeled neurons. The project will also use electrophysiological, and immunostaining anatomical techniques to determine subcellular channel localization experimentally. Overall, the project will contribute to advance the fundamental knowledge on how subcellular channel localization contributes to the processing of information within individual neurons.