Resonance In Neural Network: Signal Detection And Propagation

Neural system is the most complicated structure in human body. Neurons andneural ensembles in neural system are the foundation for brain function. The detectionand propagation of information in neuronal networks are the problems in neuralscience. Therefore, we construct network models affected by external stimuli,including feedforward network, cortical network and hippocampal network. Then, westudy the characteristics of signal propagation and resonance, in order to explore howto modulate the propagation of neural information based on external stimuli.Feedforward network models based on different neural models are firstconstructed to explore the basic law of information propagation. Due to the ubiquityof noise, it is possible for noise evoked vibrational resonance to be a mechanism ofweak signal’s detection and propagation. It is shown that noise could enhance thepropagation of the weak signal in multilayer feedforward network. Besides, the highfrequency stimulus is used to simulate the noise to evoke vibrational resonance. It isfound that there exist optimal amplitude and frequency of high-frequency driving,which lead to the maximal vibrational resonance phenomenon. We also discussed therole of network characteristic, such as connection probability, synaptic time constant,heterogeneous and feedback connection on detection and propagation of weak signal.Based on cortical network of Izhikevich model, we studied the effects of externalstimuli from the view of vibrational resonance and plasticity. It can be found that highfrequency stimulus can enhance the propagation of weak signal. Besides, thepropagation of weak signal in cortical network could be modulated by the size ofnetwork, synaptic connection probability and synaptic weight. Plasticity is one of themost important features in the neuronal network. We studied the self-sustained firingactivities of the cortical network with plastic rules in weak stimuli. It can beconcluded that, the cortical network at the end of the learning process showeduncorrelated Poissonian spike trains when there is no external stimuli, however, thenetwork displayed self-sustained firing activities when the network is stimulated byalternative current. Plasticity leads to the accumulation of the effects of the externalstimuli. Even if the external stimuli are weak, their effects can be obvious after thelearning process.We finally built two-compartmental PR neuron model of Hippocampal CA3pyramid cell. It can be shown that single neuron model displays no spiking activity,periodic spiking and irregular spiking patterns when the parameters are different. The study of network synchronization shows that the synchronous state of hippocampusCA3neuron network could be changed by external stimuli.We studied the propagation of neural information based on many neuralcharacteristics including resonance, synchronization and plasticity and found theeffects of high-frequent-signal, noises and network structure on informationpropagation. Our analysis may have potential helps on the designs of controlstrategies for modifying neural coding.