Activation Mechanism of Directly Coupled N-Type Memristive Neuron Circuits

Neuron activation, which involves the state transitioning from inhibition to excitation between adjacent neuron units due to minor perturbations, can induce intricate dynamics such as synchronized oscillations and local excitation. These phenomena arising from coupling-induced changes are vital for information processing in Cellular Neural Networks. This study takes the coupled memristive neuron circuits based on the N-type Locally Active Memristor (LAM) as an example to elucidate the activation mechanism. We predict the oscillation conditions of neurons under static bias and investigate the factors affecting the system stability under dynamic parameters. Then, by analyzing global dynamics before and after coupling, we find that even slight initial differences in identical inactive neurons can lead to reactivation and periodic spiking. The dynamics map visually represents the parameter domain where the neuron activation occurs. Moreover, the impact of initial condition differences between the neurons on the dynamics of the coupled system is explored. Finally, according to the mathematical expression of N-type LAM, an equivalent emulator is designed, and the fourth-order coupled circuit is subsequently constructed. The hardware implementation demonstrates the correctness of theoretical analysis and simulation results. This work provides a systematic method for the coupling of neuron circuits, laying a stepstone for neuromorphic computing.