Modularity Induced Gating and Delays in Neuronal Networks

Neural networks, despite their highly interconnected nature, exhibit distinctly localized and gated activation. Modularity, a distinctive feature of neural networks, has been recently proposed as an important parameter determining the manner by which networks support activity propagation. Here we use an engineered biological model, consisting of engineered rat cortical neurons, to study the role of modular topology in gating the activity between cell populations. We show that pairs of connected modules support conditional propagation (transmitting stronger bursts with higher probability), long delays and propagation asymmetry. Moreover, large modular networks manifest diverse patterns of both local and global activation. Blocking inhibition decreased activity diversity and replaced it with highly consistent transmission patterns. By independently controlling modularity and disinhibition, experimentally and in a model, we pose that modular topology is an important parameter affecting activation localization and is instrumental for population-level gating by disinhibition.尽管神经网络具有高度互联的特性，但其激活过程呈现出显著的局部化与门控特征。模块化（modularity）作为神经网络的标志性特征之一，近期被提出为决定网络支持神经活动传播方式的关键参数。本研究采用由工程化大鼠皮层神经元构成的工程化生物模型，探究模块化拓扑结构在调控细胞群间神经活动传播中的作用。研究结果表明，成对连接的模块化单元可实现条件性传播（以更高概率传递更强的脉冲簇）、长时程延迟以及传播不对称性。此外，大型模块化网络可呈现出局部与全局激活的多样化模式。抑制阻断会降低神经活动的多样性，并代之以高度一致的传播模式。本研究通过实验与模型手段分别独立调控模块化程度与去抑制（disinhibition）状态，证实模块化拓扑结构是影响激活定位的关键参数，同时也是通过去抑制实现细胞群水平门控调控的重要基础。