Gating function based on transmission delays and stochastic resonance in motif network with FPGA implementation

Attuned function of different brain regions relies on the precise yet flexible communication between their subsystems, and flexible gating of information flow among brain motif network could increase neuronal response which is selected. However, the underlying mechanism remains to be investigated. In this paper, we present a hardware efficient, scalable, and real-time computing strategy for the implementation of large-scale motif cortical network via field-programmable gate array (FPGA). Based on FPGA network model, we explore the information transmission in cortical network by investigating the gating function from the perspective of stochastic resonance, which is an important way for neuronal population to communicate. Remarkably, it is found that the cortical network can only transmit signal when the excitatory-inhibitory (E/I) balance is lopsided, which is called as "E/I semi-balanced state." Moreover, time delay between populations affects the information transmission mainly by determining the phase timestamp at which the signal is transmitted from one region to another. In short, E/I semi-balance and time delay together realize the gating function to input in cortical motif network. The experimental results demonstrate that the gating function of information is determined by E/I semi-balance and time delay in cortical motif network. Besides, mean-field theory is applied to verify the simulation results, and the cortical network is found to resonate in the form of cycle dynamics. Verification result shows that the simulation speed on hardware is 27.4 ms, which is much better than the 90 s on MATLAB. Consequently, the hardware synthesis results indicate high computational speed and low area utilization, which helps to deploy the function realization of biological neuronal network in intelligent control and remote communication.