Research On Frequency-difference-dependent Stochastic Resonance In Neural Systems

The brain is a complex nonlinear system and noise plays an important role in neural information processing.In particular,it has been found that stochastic resonance can help neurons improve their capabilities to detect weak periodic signals.The classical SR studies in neuroscience have mainly focused on neural systems driven by an isolated periodic force.However,real biological neurons may be driven by multiple periodic signals from different brain regions.With the help of noise,a recent study has shown that neural systems may process weak envelope modulation signal that is superimposed by two periodic signals with different frequencies.Inspired by this finding,the theory of frequency-difference-dependent stochastic resonance(FDDSR)has been proposed in neural systems.However,so far it is still unclear whether the phenomenon of FDDSR can be observed in neurons with different intrinsic firing properties and whether this behavior is regulated by the topology of neuronal networks.In this thesis,we investigate the FDDSR and its regulatory mechanisms from both the single-neuron and population levels by computational modeling.Our main results are summarized as follows:Firstly,we focus on four types of spiking neurons at the single-neuron level,including the regular spiking(RS)neurons,the intrinsically bursting(IB)neurons,the fast spiking(FS)neurons and the low-threshold spiking(LTS)neurons.It shows that the phenomenon of FDDSR exists in all these types of neurons.In the case of lower noise,all these neurons can enhance their responses to the information carried by the weak envelop modulated signals at the beat frequency through increasing the subthreshold direct currents.In addition,different types of neurons respond optimally to external periodic stimuli at a respective range of beat frequency.FS neurons have a stronger response to the lower beat frequencies,while the other types of neurons have an optimal response to the higher beat frequencies.Secondly,by using the improved NW small-world network algorithm,we establish a small-world neuronal network of leaky integrate-and-fire neurons.On this basis,the regulation mechanism of the FDDSR in small-world neuronal networks is explored systematically.In the case of the low noise intensity,properly increasing of the shortcuts can improve the response of the neuronal network to envelope modulation signal at the beat frequency.When the intensity of noise is relatively high,the neuronal network with fewer shortcuts has a better response at beat frequency,but when more shortcuts are added,the response of network at beat frequency deteriorates.Furthermore,we show that properly increasing the weight of excitatory synapses can improve the FDDSR of networks,and an appropriate fine-tuning of the excitation-inhibition balance can further optimize the response of a neural ensemble to the superimposed signal.In addition,we find that the performance of the neuronal network for FDDSR can also be improved when the beat frequency matches the intrinsic oscillatory frequency of the neuronal network.Overall,our findings indicate that the FDDSR is a common nonlinear dynamic phenomenon in the neural systems,and the mechanism of the FDDSR contributes to the processing of information carried by the weak envelope modulation signal at the beat frequency.Meanwhile,we illustrate that the intrinsic oscillatory properties of neurons,network topology and network parameters all modulate the performance for FDDSR.Our study gives an insight into understanding how brain processing multiple frequency information simultaneously and present some sights for further study and application.