Power-law spectrum and small-world structure emerge from coupled evolution of neuronal activity and synaptic dynamics

4th International Workshop on Statistical Physics and Mathematics for Complex Systems

物理学;数学

Zeng, Hong-Li^1 ; Zhu, Chen-Ping^1 ; Guo, Yan-Dong^1,2 ; Teng, Ao^1 ; Jia, Jing^1 ; Kong, Hui^1 ; Zhou, Rui^1 ; Yang, Juan-Ping^1 ; Li, Su-Quan^1

College of Science, Nanjing University of Aeronautics and Astronautics, Nanjing

210016, China^1

College of Electronic Science and Engineering, Nanjing University of Posts and Telecommunications, Nanjing

210046, China^2

关键词: Degree-degree correlations;    Electrocorticogram (ECoG);    Initial configuration;    Neuronal activities;    Orders of magnitude;    Power law spectrum;    Power-law behavior;    Synaptic plasticity;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/604/1/012023/pdf
DOI  :  10.1088/1742-6596/604/1/012023

来源: IOP

【 摘 要 】

A co-evolutionary neuronal network model based on previous ones is proposed, and both functional and structural properties are numerically calculated. Recent experiments have revealed power-law behavior in electrocorticogram (ECoG) spectrum related with synaptic plasticity and reorganization. In the present neuronal network model, the network starts its evolution from the initial configuration of random network which is the least biased and without special structure, and the interaction rules among neurons are modified from both models by Bornholdt's and Arcangelis' groups to simulate the process of synaptic development and maturation. The system exhibits dynamic small-world structure which is the result of evolution instead of the assumption beforehand. Meanwhile, the power spectrum of electrical signals reproduces the power-law behavior with the exponent 2.0 just as what is experimentally measured in ECoG spectrum. Moreover, the power spectrum of the average degree per neuron over time also exhibits power-law behavior, with the exponent 2.0 again over more than 5 orders of magnitude. Different from previous results, our network exhibits assortative degree-degree correlation which is expected to be checked by experiments.