Artificial afferent neurons based on the metal-insulator transition of VO₂

Neuromorphic computing offers significant advantages in addressing data redundancy and enhancing system energy efficiency. Although extensive research has been conducted on pulsed neural networks and bionic sensors, the development of artificial electronic afferent neurons for low-energy information transfer remains limited. This study introduces an artificial afferent neuron comprising a vanadium dioxide (VO₂) device, capacitor and resistor. The VO₂ devices exhibit stable electrically induced metal-insulator transition (MIT). Leveraging the MIT of this device, we develop an artificial afferent neuron to transform constant or sinusoidal analog signals into pulsed voltage signals. The output frequency increases with the increase of the input voltage, mimicking the faster pulse outputs of biological afferent neurons in response to stronger stimuli.