The Dynamics of Two-Layer Continuous Attractor Neural Network Model With Moving Stimulus
Min Yan, K. Y. Michael Wong · Rare & Special e-Zone (The Hong Kong University of Science and Technology) · 2015
The dynamics of single-layer continuous attractor neural networks (CANNs) model has gained extensive attention. We generalize the neural network model to a two-layer structure based on the original features, and take feedback and feedforward effects into consideration. We apply a static stimulus on one layer and a moving stimulus on the other layer. Under various strengths of the feedback and feedforward couplings, the two-layer network will show distinct behaviors. Under a relatively weak input in the first layer, when the feedback is inhibitory, the network dynamics displays kinks. The kinks will also behave differently with distinct inhibitory strengths. When both the feedback and feedforward couplings are excitatory, the activities in the two layers will attract each other. Therefore, a stronger moving stimulus is required to drag both layers to move simultaneously. We also consider the effect of inhibitory or excitatory in both layers in turn, which leads to some opposite behaviors.