Classical neuron models have inspired decades of research and development for in silico hardware and AI software. McCulloch and Pitts developed one such model that relies on Hebbian's synaptic reinforcement principle. Today, we know that pyramidal cells, the most important excitatory neurons in the brain, include nonlinear NMDA-based dendritic integrations. Also, their synapses are reinforced in accordance to an asymmetric spike-timing-dependent plasticity (STDP) principle. Energy efficient realization of these non-linear models in millions is very challenging. In this paper, for the first time, we present a novel RF-analog hybrid energy efficient equivalent model with a feedback mechanism to represent the non-linear behavior of L5 PCs.
