Microconnectomics of neocortex: a multiscale computer model

描述（由申请人提供）：我们将基于正在进行的研究中获得的丰富实验数据集开发初级运动皮层（M1区域）的多尺度模型。该模型的范围将从树突中的离子通道水平，到皮层其他区域的输入和输出水平，范围从微米到厘米，时间范围为毫秒到10秒。我们将评估跨尺度的动态相互作用，由于具有跨皮层和跨尺度的第5层锥体细胞的长顶端树突的结构而变得更加复杂。这一特征产生了复杂的结构-功能关系：顶端树突直接处理来自不同皮质层的输入，然后从局部微电路输出（直接输入/输出）。它们也在规模层次中起作用，形成局部网络的一个组成部分，通过整个第5层金字塔细胞集合提供并行处理输入以产生输出。第5层金字塔形成两组不同的细胞：向纹状体和其他皮层区域投射的皮质纹状体细胞和向下投射到脑干和脊髓的皮质脊髓细胞。我们的双输出假设将这些概念化为由两种大细胞类型锚定的部分可分离的子电路。我们认为，这些亚回路之间的单向投射（皮质纹状体到皮质脊髓）影响了一个主要的编码转换：显性皮质皮质颞编码（皮质纹状体亚回路）到显性速率编码（皮质脊髓亚回路）。重要的是，对于这一假设，皮质脊髓锥体细胞表现出线性激活特性，几乎没有适应。我们的目标从低到高，通过一系列紧密相连的实验和模拟，预测导致实验，再导致修改的模拟：模拟树突中突触信号的整合，基于不同树突位置细胞激活测量的Ih和IA密度（亚细胞尺度：1-10 μ m）；2. 基于细胞放电的模型

DESCRIPTION (provided by applicant): We will develop a multi-scale model of primary motor cortex (area M1) based on a rich experimental dataset obtained in ongoing studies. The model will range from the level of ion channels in dendrites, up to the level of the inputs from and outputs to other areas of cortex, a range of microns to centimeters, with a temporal range of milliseconds to 10 sec. We will evaluate dynamical interactions across scale, made more complicated by a structure that features long apical dendrites of Layer 5 pyramidal cells that reach across layers of cortex and thereby across scales. This feature produces complex structure-function relations: apical dendrites directly process inputs from different cortical layes for export from the local microcircuit (direct input/output). They also act within the scale hierarchy, forming a component of the local network which provides a parallel processing of inputs to produce outputs via the entire Layer 5 pyramidal cell ensemble. Layer 5 pyramids form two distinct groups: corticostriatal cells that project to striatum and to other cortical area, and corticospinal cells that project downwards to brainstem and spinal cord. Our dual-output hypothesis conceptualizes these as partially separable subcircuits anchored by the two massive cell types. We suggest that the one-way projection (corticostriatal to corticospinal) between these subcircuits effects a major code transformation: dominant corticocortical temporal coding (corticostriatal subcircuit) to dominant rate coding (corticospinal subcircuit). Importantly for ths hypothesis, the corticospinal pyramidal cells show linear activation properties with little adaptation. Our Aims proceed from low to high through sets of tightly-linked experiment and simulation, with predictions leading to experiments leading to modified simulations: 1. Simulate integration of synaptic signals in dendrites, based on densities of Ih and IA measured with cell activation at different dendritic locations (subcellular scale: 1-10�m); 2. Model cell firing based on current clamp experiments (cell scale: 30-800�m); 3. Create circuit-level models based on projection strength measurements (microcircuit scale: 2-5mm); 4. Use information theoretic and dynamical measures to evaluate SPI-STR code transformation hypothesis through input/output analysis (projection scale: 10-100mm). We predict that the neocortical circuit can either combine or multiplex signals, depending on tags based on location, frequency, phase, and amplitude of inputs.