Wave Digital Emulation of a Bio-Inspired Circuit for Axon Growth

Even though artificial neural networks and corresponding hardware implementations are already very powerful, they are still far away from the efficiency of biological neuronal networks. It is hence useful to better understand and model biological principles shaping these networks besides synaptic changes, such as axon growth. As the few existing implementations focus on technical abstractions of the latter aspect, in this work, we aim for a bio-inspired circuit implementing axon growth well comparable to biological processes. For this purpose, we utilize Morris-Lecar oscillators as nodes of Ranvier and memristors for the growth mechanism. In order to emulate this circuit, we derive a wave digital model, which is well scalable and runtime efficient. Emulation results show the circuit successfully mimics axon growth principles observed in mammalian visual cortex and motor cortex neurons.