Neural Net to Neuronal Network Memristor Interconnects

Abstract

Abstract.Hardware-based Artificial Intelligence (A.I.) has many potential applications in biomedical technology; for example, connecting expert systems to biosensors for real-time physiological monitoring of a range of biomarkers, or interfacing with the brain. We suggest that memristors are well-placed to interface directly with neurons to interconnect between computer hardware and the brain for 3 reasons: memristors are widely-touted as neuromorphic computing elements; memristor theory has been successfully used to model spike-time dependent plasticity and the Hodgkin-Huxley model of the neuron; and, the d.c. response of the memristor is a current spike. In this chapter we show that connecting a spiking memristor network to electrically active neuronal cells causes a change in the memristor network dynamics by: removing the memristor spikes, which we show is due to the effects of connection to aqueous medium; causing a change in current decay rate consistent with a change in memristor state; presenting more-linear I − t dynamics; and increasing the memristor spiking rate, as a consequence of interaction with the active cells. This demonstrates that such cells are capable of communicating directly with memristors, without the need for computer translation.