The structure identification of feedforward neuronal network based on adaptive synchronization

The function of the neuronal network is neural code. In the network, neurons connect with each other by synapses. The stability of synaptic connections ensures the reliable transmission of spiking activity in the network, which is one of the key properties of candidate neural code. However, some nervous system diseases can lead to some synaptic connections lost stochastically in the neuronal network, which will disturb the reliability of transmission seriously. For studying the transmission feature of the potential neural code, it is necessary to detect whether there exist lost synapses and their position in the network. In this paper, a virtual network is built to identify the synaptic connection structure in the feedforward neuronal network. Through the adaptive estimation method, the variable connections in the virtual network detected the connected and unconnected synapses successfully in the feedforward neuronal network. Furthermore, our simulation results proved that the theoretical analysis is effective. This research provides a general method to detect the lost synapses in the feedforward neuronal network.