The Influence Of Synaptic Plasticity In The Working Memory

It is known of that one of the most important cognitive functions of the brain is learning and memory.In this paper,we first construct a mathematical model of working memory through synaptic plasticity of neurons,then we analyze the mechanism of working memory formation,the way of maintaining working memory and the cause of working memory reproduction.Working memory is regarded as the ability to store and manipulate information for completing a complex cognitive task in a short time for the specific neuron population.The generally accepted model is the sample-selective mode,i.e.A specific working memory can only be stored by a specific selective neuron population.In the paper,we adopt the stochastic HH neuron model to describe neuron,so that we can explore the influence of channel noise on short-term working memory and long-term memory.In the first chapter,we construct a short-term working memory model based on synaptic plasticity,which is composed of a stochastic HH neuron model and short-term synaptic connections.We obtain the following results:(i)By comparing the firing behaviors of the neuronal network under different membrane areas,it is found that neurons would fire very randomly as the membrane area is small(eg:=100m~2);Rare spikes are generated as the membrane area is large(eg:=500m~2);However the working memory occurs for the moderate membrane area(eg:=270m~2).(ii)We further explore the way to extend working memory.When the membrane area is large,we need to stimulate the selective population repeatedly to retrieve the memory after the formation of working memory,otherwise the memory would disappear after a period of time;However,for moderate membrane area,working memory can be maintained in the presence of subsequent stimuli at any time.(iii)Finally,the formation of the new working memory is investigated.It is found that if one working memory has already been generated and another working memory expects to be produced,the spike of the previous selective population will be inhibited.The second chapter focuses on the study of long-term memory.Analogy to first chapter,we construct a long-term memory model formed by synaptic plasticity.However,the model considers different synaptic connections and network structures.The following studies are carried out using this model:(i)First,we investigate the effect of the decay time constant of NMDA receptor and the duration of stimulation on long-term memory.By observing the spike of neuron population,it is found that larger decay time constant yields the longer duration of the long-term memory.So does for the duration of stimulation.(ii)We further explore the influence of synaptic weight and the initial release probability of neurotransmitters on long-term memory.We find that synaptic weight plays a role in different long-term memories.Meanwhile,it is also found that if the initial release probability of neurotransmitters exceeds a certain value,the effect on long-term memory is not significant.(iii)Finally,we consider the relationship between the two kinds of long-term memory.It is found that the matched memory promotes memory preservation,while the non matched memory inhibits memory preservation.