Exploiting the electrothermal timescale in PrMnO3 RRAM for a compact, clock-less neuron exhibiting biological spiking patterns

Spiking Neural Networks (SNNs) are gaining widespread momentum in the field of neuromorphic computing. These network systems integrated with neurons and synapses provide computational efficiency by mimicking human brain. It is desired to incorporate biological neuronal dynamics, including complex spiking patterns which represent diverse brain activities within neural networks. Earlier hardware realization was (1) area intensive because large capacitors circuit design, (2) were demonstrated clocked at device level. To achieve more realistic neuron behavior, emerging memristive devices considered promising alternatives. In this paper, we propose, PrMnO3(PMO) -RRAM device-based neuron. The voltage-controlled electrothermal timescales compact PMO RRAM replace electrical charging a capacitor. timescale used implement an integration block multiple coupled refractory generate dynamics. Here, first, Verilog-A implementation thermal model demonstrated, captures current-temperature dynamics device. Second, driving circuitry designed mimic different cortical neurons, Intrinsic bursting (IB) Chattering (CH). Third, simulated, includes demonstrate asynchronous behavior. Finally, hardware-software hybrid analysis done experimentally characterized work presents realizable biologically comparable hardware-efficient solution for large-scale SNNs.