Memristor-based neuromorphic circuit for visual emotional non-associative learning

Most of the nonassociative learning neural networks based on memristor only consider habituation or sensitization and ignore the combination with emotions. According to biological characteristics, a nonassociative learning memristive neural network circuit of emotions induced by vision is designed in this paper. The circuit comprises the voltage control module, the synapse module, the potentiation of habituation module and the spontaneous recovery module. The functions such as habituation and spontaneous recovery, habituation of dishabituation, potentiation of habituation and sensitization are implemented by the circuit. Emotions through visual stimuli of different intensities are induced in designed circuit. The designed circuit is able to implement emotional habituation and sensitization like humans after processing the sensory signals from vision. At the same time, based on picture stimuli as inputs, PSpice simulations confirm the feasibility and accuracy of the proposed circuit at the circuit-simulation level. The proposed circuit may serve as a potential neuromorphic module for robotic affective perception and emotional learning, providing a reference for future research on brain-inspired emotional intelligence.