Design of a Biomolecular Neuristor Circuit for Bioinspired Control

Abstract

The nervous system serves as an inspiration for control and processing systems but can be difficult to replicate or directly implement using biological neurons in bioinspired and biohybrid systems. An alternative path is to use synthetic biomolecular neuristors that are inspired by biological neurons and can closely mimic their spiking behavior. A neuristor is built using two dynamical lipid-bilayer-based devices that exhibit volatile memory and negative differential resistance arising from the dynamics of voltage-gated ion channels and the ion gradients across the lipid membranes. The firing pattern and frequency can be easily tuned by engineering the composition of each neuristor. To investigate the viability of using biomolecular neuristors to design neural control circuits, a neuristor computational model was implemented and compared to that of the Izhikevich neuron model. Both models were assessed for individual cells and then arranged to form mutually inhibitory circuits, as central pattern generators commonly found in motor control circuits. Both models can replicate the alternating firing behavior, but further parameter tuning is needed for the neuristor model to better match the firing frequency of the biological neuron model.