Neural integration with active dendrites and inhibition

描述（由申请人提供）：在广泛的大脑区域观察到持续的神经活动，并涉及从信息存储和处理到运动控制的功能。大脑工作记忆区域持续神经活动的缺陷被认为是精神分裂症的核心特征。本研究旨在通过对一个具有持续神经活动的模型系统——金鱼动眼神经积分器进行计算建模，揭示持续神经活动背后的神经机制。动眼神经积分器接收速度编码的眼球运动指令，并将这些指令转换成控制眼睛位置的信号。在没有速度指令的情况下，积分器中的神经元在几十秒内保持稳定的放电速率。神经整合器受损的患者无法保持稳定的凝视，并且在眼球追踪行为和眼反射方面存在缺陷。以前的动眼肌系统模型忽略了一些重要的特征，这些特征使得它们无法通过实验进行明确的测试。使用一种允许数据直接合并的新框架，将构建一个实验约束和可验证的金鱼动眼神经积分器模型。该模型将用于分析网络和细胞对持续神经活动的贡献。突触兴奋、突触抑制和内在神经元兴奋性的贡献将通过模拟系统中持续神经活动的最新解剖学和药理学操作来评估。在网络水平上的初步建模表明，细胞之间的反复相互作用是由双稳态树突过程介导的，该过程被假设为树突平台电位。一个具有树突分支结构和电压敏感的突触和固有电导的网络模型将被构建，以验证电压依赖树突特性增加网络对扰动的鲁棒性的假设。该模型将受到细胞内切片和活体记录的限制，并将与顾问实验室目前正在进行的树突成像实验进行比较。通过在一个具有良好特征的系统中产生一个实验约束和可验证的模型，这项工作有望揭示持久神经活动产生的核心机制。

DESCRIPTION (provided by applicant): Persistent neural activity has been observed in a wide range of brain regions and has been implicated in functions ranging from information storage and processing to motor control. Deficits in persistent neural activity in working memory areas of the brain have been suggested as a core feature of schizophrenia. The proposed work seeks to reveal the neural mechanisms underlying persistent neural activity by computational modeling of a model system exhibiting persistent neural activity, the goldfish oculomotor neural integrator. The oculomotor neural integrator receives velocity-coded eye movement commands and converts these into signals that control the position of the eyes. In the absence of velocity commands, neurons in the integrator maintain a steady rate of firing for tens of seconds. Patients with impaired neural integrators are unable to maintain a steady gaze and have deficits in eye tracking behavior and ocular reflexes. Previous models of the oculomotor system have neglected important features that have made them unable to be tested explicitly by experiment. Using a novel framework that allows data to be directly incorporated, an experimentally constrained and verifiable model of the goldfish oculomotor neural integrator will be constructed. The model will be used to analyze network and cellular contributions to persistent neural activity. The contributions of synaptic excitation, synaptic inhibition, and intrinsic neuronal excitability will be assessed by modeling recent anatomical and pharmacological manipulations of persistent neural activity in the system. Preliminary modeling at the network level suggests that recurrent interactions between cells are mediated by a bistable dendritic process that is hypothesized to be a dendritic plateau potential. A network model with dendritic branching structures and voltage-sensitive synaptic and intrinsic conductances will be constructed to test the hypothesis that voltage-dependent dendritic properties increase the robustness of the network to perturbations. The model will be constrained by intracellular recordings in slice and in vivo and will be compared to dendritic imaging experiments currently being conducted in the consultants' laboratories. By producing an experimentally constrained and verifiable model in a well-characterized system, this work promises to reveal core mechanisms by which persistent neural activity is generated.