Regulating Characteristics Of Cortical Neural Network Through Synaptic Plasticity

Cortical neural network is a complex dynamic system.Existing studies have shown that,for one thing,excitation-inhabitation balance(E-I balance)plays an important role in maintaining normal brain function;for another,excitation-inhabitation imbalance will lead to the occurrence of multiple diseases such as Alzheimer's disease,epilepsy,schizophrenia.Extensive experimental studies have revealed that firing rate homeostasis is a fundamental feature of neuroelectric activity,and it provides a substrate for neural information processing and maintaining normal neurological functions.Moreover,the structure of neural network has a direct impact on the efficiency of information processing.Therefore,the present study focuses on E-I balance,firing rate homeostasis and structure of cortical neural network.Existing studies on E-I balance are mainly based on inhibitory synaptic plasticity,while there are both excitatory and inhibitory synaptic plasticity in cortical neural network.Excitatory synaptic plasticity has an effect on E-I balance,so the E-I balance underlined by the interaction of excitatory and inhibitory synaptic plasticity remains to be explored.Firing rate homeostasis is not only a typical characteristic of neuroelectric activity but also an important indicator to measure the level of neuroelectric activity.It is of great significance for neural network to achieve neural information processing and maintain normal function.Although extensive experimental researches have found that firing rate homeostasis is ubiquitous in brain electrical activity,the mechanism of firing rate homeostasis still needs to be further studied.Cortical neural network is a complex dynamic system,and the structure of neural network will change along with cortical activities.Previous studies have shown that network structure exerts effect on computational efficiency of cortical neural network,but the underlined mechanism is still not wholly known.Therefore,in this paper,we construct a cortical neural network composed of excitatory neurons and inhibitory neurons.The study mainly includes the following two parts,the first part is the study of firing rate homeostasis and neural network structure based on inhibitory synaptic plasticity;the second part is the study of E-I balance,firing rate homeostasis and neural network structure based on excitatory and inhibitory synaptic plasticity.The detail contents are as follows:(1)To study the network characteristics based on inhibitory synaptic plasticity.Firstly,we establish a neural network containing excitatory and inhibitory neurons,and incorporate inhibitory spike-timing-dependent plasticity(STDP)into the pathway from inhibitory neurons to excitatory neurons.Then we study the firing rate homeostasis and structure of neural network.Spike raster,neural membrane potentials,neural firing rate,average firing rate,firing rate charts,strength of inhibitory synapse and inhibitory synaptic conductance are used as performance measures to evaluate firing rate homeostasis.The structure of network is evaluated and analyzed by using degree distribution and connection matrix of neural network at different moments.In addition,in order to simulate the actual situation of neural network,we further analyze the robustness of firing rate homeostasis in the face of external disturbances and parameter perturbations.The results show that,with the action of inhibitory synaptic plasticity,neural network can achieve firing rate homeostasis with a strong robustness.Under the regulation of inhibitory synaptic plasticity,the degree of neural network increases on the whole,but the connection probability between neurons does not change significantly.(2)To study the network characteristics based on excitatory and inhibitory synaptic plasticity.Firstly,we construct a neural network containing excitatory and inhibitory neurons,and incorporate STDP into the pathway from inhibitory neurons to excitatory neurons and the pathway from excitatory neurons to excitatory neurons.The sum of synaptic currents and the ratio of synaptic currents are used to evaluate E-I balance.Spike rasters,neural membrane potentials,neural firing rate,average firing rate,firing rate charts,strength of synapse and synaptic conductance are used as firing rate homeostasis evaluation indexes.For the structure of neural network,we mainly use degree distribution and connection matrix as evaluation indexes.Furthermore,we study the robustness of E-I balance and firing rate homeostasis in the face of external disturbances and parameter perturbations.Our results indicate that,under the interaction of excitatory and inhibitory synaptic plasticity,the neural network can achieve E-I balance and firing rate homeostasis in a robust manner after a long transition time.Compared with the results only under the regulation of inhibitory synaptic plasticity,the E-I balance of network is relatively low,but the network is more sensitive to stimulus.With the interaction of excitatory and inhibitory synaptic plasticity,the neural network can achieve firing rate homeostasis,but the neurons are intermittent fire and intermittent inhibition.The structure of neural network changes significantly under the interaction of excitatory and inhibitory synaptic plasticity,the synaptic connections between excitatory neurons become spare and strong.The present study results show that inhibitory synaptic plasticity is an effective mechanism to achieve firing rate homeostasis.When there are both excitatory and inhibitory synaptic plasticity in neural network,the network can achieve E-I balance and firing rate homeostasis.Meanwhile,the network structure changes significantly,and the synaptic connections between excitatory neurons become spare and strong.Moreover,the network has a higher sensitivity to signals.Our study focuses on the analysis of neural network characteristics based on excitatory and inhibitory synaptic plasticity,which has important role in the research of both computational neuroscience and physiological research.