This project aims to develop compact elements that mimic the function of brain neurons and synapses, creating Spiking Neural Networks capable of performing brain-like computations in a smaller size and with lower energy consumption than traditional methods. The goal is to build hardware that replicates neuronal behavior, integrating analog physical components and complex dynamics to enhance perception, learning, and computational capabilities.

Developing compact realizations of synaptic and neuronal functions has broad applications in brain understanding and bio-inspired computing. Spiking Neural Networks (SNNs), modeled on biological knowledge, communicate through spikes via synapses with adjustable weights. Currently, SNNs use CMOS circuits, requiring significant resources to simulate spiking behavior. By building neurons and synaptic devices with biological-like physical dynamics, we can reduce size, complexity, and energy use. In biological systems, neurons process signals by adjusting synaptic strength based on timing differences, requiring neuristor elements that emulate key nonlinear and amplification properties of natural neurons.