Dynamic Neural Network for Recognizing Interspike Interval Sequences

The functions of neural system, such as learning, recognition and memory, are emerging from the dynamic mechanisms of neurons and synapses. A dynamic neural network is designed to discuss how the dynamic mechanisms in the neurons and synapses work in recognizing interspike intervals (ISIs). The dynamic synaptic transmission mechanisms and the properties of integrate-and-fire neurons are the key mechanisms, based on which the input interspike interval sequences are decomposed into isolated spikes. The synaptic delay times modulated by STDP learning rule is another key mechanism in the ISIs recognition, based on which the ISIs are learned and saved in the delay times. After learning, the neural network could recognize whether different input sequences include the same consecutive ISIs. This model shows that the dynamic mechanisms of neurons and synapses in brain are powerful, even under a simple network structure.