The effect of inhibitory feedback on temporal regularity in neural networks

摘要

The regularity of the spiking activity is closely related to the neural encoding and transmission of neural information. However, it is largely unknown how the inhibitory feedback, as a common interaction existed in cortex, contributes to the precise transmission of neural information. In order to explore the answer to this question, we construct a feed-forward network (FFN) model with lateral inhibitory circuits and study the effect of inhibitory feedback on the temporal regularity of spiking activities. Besides, we discuss the spike-time regularity of pure FFN model without feedback for selecting typical dynamic states of neural network. Results show that the inhibitory feedback with adaptive strength enhance the regularity of spike-time in typical states. It indicates that the spike-time regularity observed in the experiment is resulted by the interaction between feed-forward and inhibitory feedback.