The Random Response Of Neuron And Its Signal Transmission In The Feedforward Network

A series of action potentials produced by neurons under external stimulation are the carriers of nerve information.These action potentials are closely related to the transmembrane movement of ions in the nervous system.Thermal noise in the biological nervous system is ubiquitous due to ion movement and thermal fluctuation of ion channel switches.These noises not only have an important effect on the action potential spiking behavior of neurons,but also have an important significance on the transmission of nerve information and the realization of related nerve functions.In this paper,the generation of neuron random signals and the transmission of random signals in feed-forward networks were studied under the action of noise.Based on FitzHugh-Nagumo(FHN)neuron model,the stochastic distribution behavior of neurons under the action of noise and external stimulus current was studied in this paper.The results showed that,in some specific parameters,with the increase of noise intensity,the action potential sequence of FHN neurons would change from random to regular to random,that was to say,the distribution behavior of coherent resonance was shown.When the parameters of FHN neurons changed,the stochastic distribution and coherent resonance behavior also changed.On the other hand,the stochastic distribution and coherent resonance behavior of FHN neurons showed strong symmetry with respect to system parameter a and stimulus current I.Combining with phase plane analysis,the parameter intervals of coherent resonance firing behavior of FHN neurons were given.We also revealed the relationship between the symmetry of these stochastic distribution and coherent resonance behaviors with respect to system parameters a and I and the intrinsic dynamics of FHN neurons.Neural information transmission is one of the most basic,key and popular research topics in brain science and neuroscience.As a multi-layer,highly modular structural system,the brain has a significant hierarchical structure.Feed-forward neuron network can not only simulate the functional groups of neurons in different functional areas of the brain hierarchically,but also reflect the characteristics of the brain through a large number of neurons in different brain regions to coordinate the transmission of nerve information.In addition,as a special self-feedback structure,autapse plays an important role in the formation and evolution of patches,resonance and synchronization of neurons in neural networks.In this paper,a feed-forward neural network composed of Hodgkin-Huxley(HH)neurons was used to study the transmission of stochastic signals in the network.It was found that parameters such as autapse connection probability and delay time could regulate the distribution patterns of neurons in different layers of the network,and then affect the transmission characteristics of stochastic signals in the network.Firstly,autapse feedback regulation could increase or decrease the firing rate of neuron layer in the network,thereby promoting or inhibiting the transmission of stochastic signals in the network.Moreover,the autapse delay time region corresponding to this promotion or inhibition did not change with the changes of network parameters.When the delay time was between 0 milliseconds and 6 milliseconds,it was inhibited,while the delay time was between 6 milliseconds and 25 milliseconds,it was promoted.Secondly,when the probability of autapse connection was high and the delay time was moderate(6.0ms～15.0ms),the autapse feedback regulation weakened the synchronization characteristics of the network,but greatly improved the firing rate of the neuron layer in the network.Within this parameter range,it was most conducive to the transmission of stochastic signals in the network.