Bio-Consistent Direction-Selective Neurons Based on Threshold Switching Devices

Direction selectivity concerning animal survival was critical for the artificial visual system to perceive complex environments. In this letter, we constructed artificial neurons to simulate direction selectivity utilizing annealed Ti/NbOx:Al/Pt devices with superior performance. By integrating excitatory and inhibitory signals for accurate modulation of the number of spikes, the preferred direction was reliably identified, and visual attention was further focused through adaptive feedback. This work designed and implemented biologically similar robust artificial sensory neurons, which provided innovative support and strategies for the artificial visual sensory system.