Thermal neuristors for computing

Our research is dedicated to a novel type of thermal neuristors utilizing the quantum material VO₂. We demonstrated a wide array of reconfigurable electrical behaviors that mirror those of biological neurons, such as the all-or-nothing law, type-II neuronal rate coding law, spike-in and DC out effect, spike-in and spike-out effect, and stochastic leaky integrate-and-fire law. Remarkably, these behaviors are achieved solely through thermal interaction. Moreover, we achieved both inhibitory and excitatory functionalities within a single oxide device, enhancing its versatility. Information flow between consecutive neural layers is accomplished through thermal interactions, eliminating the necessity for traditional CMOS circuits. This groundbreaking method simplifies the incorporation of reconfigurable cascading neural layers with potential for scalable energy-efficient thermal neural networks for computing.