Optimization of noise-induced resonance mechanismson co-evolutionary neurobiological networks

协同进化神经生物网络上噪声引起的共振机制的优化

• 批准号: 456989199
• 负责人: Dr. Marius Yamakou
• 金额: --
• 依托单位: Department of Data Science
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2021
• 资助国家: 德国
• 起止时间: 2020-12-31 至 2022-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/456989199?language=en
• 关键词: Optimization; noise; induced; resonance; mechanismson

The functional role of noise is a long-standing research question in neurobiology. While noise is generally undesirable, its resonance effect is well known and is generally accepted to be crucial for the proper functioning of neurons in terms of their information coding capabilities. The phenomenon of resonance has been observed both in individual neurons and as well as in networks of neurons. It is also known that different types of noise-induced resonance mechanisms occur under different conditions. These consist of different combinations of neuron parameters, synaptic connections between neurons, network topology, and noise sources. All previous research has been focused on understanding the optimization of each of these noise-induced resonance mechanisms: (a) in non-adaptive neural networks, and (b) independently of one another. A comprehensive understanding of optimization of information processing via the optimization of these noise-induced resonance mechanisms in (a) co-evolutionary (adaptive) neural networks, and (b) by using the interplay between two or more noise-induced mechanisms, is still completely lacking. The main objective of this project is to design and classify in terms of efficiency, a plethora of optimization schemes for three different types of noise-induced resonance mechanisms, in co-evolutionary biological neural networks. The main focus is on the noise-induced resonance mechanisms of coherence resonance (CR), self-induced stochastic resonance (SISR), and recurrence resonance (RR) in co-evolutionary network motifs, scale-free networks, small-world networks, random networks, and their multilayer networks. These neural networks will consist of the biophysical Hodgin-Huxley (HH) neuron model and evolve according to the spiking time-dependent plasticity learning rule or a temporal activity-dependent structural plasticity learning rule. Before considering the HH neurons/networks, the necessary conditions for the occurrence of CR and SISR in an isolated HH neuron will be determined using geometric singular perturbation theory and stochastic multi-dimensional reaction rate theory.Depending on the topology, the synaptic properties, and the learning rule of a network, different optimization schemes for CR, SISR, and RR will design and classified with respect to their efficiency. This is a timely and unique proposal that promises to push our understanding of optimal neural coding and information processing to new frontiers.

噪声的功能作用是神经生物学中一个长期研究的问题。虽然噪声通常是不受欢迎的，但它的共振效应是众所周知的，并且被普遍认为是神经元在信息编码能力方面正常运作的关键。共振现象已经在单个神经元和神经元网络中被观察到。我们还知道，在不同的条件下会发生不同类型的噪声诱发共振机制。这些包括神经元参数、神经元之间的突触连接、网络拓扑结构和噪声源的不同组合。所有以前的研究都集中在理解这些噪声引起的共振机制的优化上：(a)在非自适应神经网络中，(b)彼此独立。通过优化这些噪声诱导的共振机制(A)协同进化（自适应）神经网络，以及(b)利用两个或多个噪声诱导机制之间的相互作用来优化信息处理的全面理解，仍然完全缺乏。该项目的主要目标是在协同进化生物神经网络中，根据效率设计和分类三种不同类型的噪声诱导共振机制的大量优化方案。重点研究了协同进化网络基元、无标度网络、小世界网络、随机网络及其多层网络中的相干共振（CR）、自致随机共振（SISR）和递归共振（RR）等噪声诱导共振机制。这些神经网络将由生物物理Hodgin-Huxley （HH）神经元模型组成，并根据峰值时间依赖的可塑性学习规则或时间活动依赖的结构可塑性学习规则进行进化。在考虑HH神经元/网络之前，将使用几何奇异摄动理论和随机多维反应速率理论确定孤立HH神经元中CR和SISR发生的必要条件。根据网络的拓扑结构、突触特性和学习规则，将根据效率对CR、SISR和RR的不同优化方案进行设计和分类。这是一个及时而独特的建议，有望将我们对最佳神经编码和信息处理的理解推向新的前沿。