Fig. 4: Selective communication scheme between neurons.

A Schematic of selective communication of biological neurons. Postsynaptic neurons selectively transmit information from presynaptic neurons by responding to specific firing features with a resonance behavior. B Schematic of an artificial neural circuit emulating the biological selective communication circuit. The three memristive H-H neurons are configured with different circuit parameters. C The selective communication scheme in the memristive neural circuit. The resonance characteristic of the BDN neuron enables it to produce spike responses only to inputs with low frequency but not to inputs with high frequency. The integration property, on the other hand, makes the SDN neuron capable of producing spike output only for inputs with a high frequency. D The selective communication scheme of the neural circuit under continuous inputs. The SDN neuron with resonance characteristics can extract single spikes from the output of the mixed pattern of the MCN and generate spike responses, and the BDN neuron with integration characteristics extract bursts from the output of the MCN. E The firing frequencies of SDN and BDN under different input voltages on the MCN neuron. The simulation data is well-fitted with the experimental results. The error bars represent the standard deviation (σ) of each group.