Neuromorphic computing, inspired by the human retina, offers a promising avenue to transcend the constraints of the traditional von Neumann architecture. To facilitate high-performance training of neuromorphic hardware, artificial synapses must exhibit characteristics such as linear and symmetric programmability, bipolar operation, multistate storage capacity, high yield, extended retention times, and minimal variability. Nevertheless, current neuromorphic devices frequently encounter limitations due to their asymmetric and nonlinear conductivity profiles, which can compromise overall performance. Consequently, there has been a significant focus on developing innovative devices capable of simultaneously demonstrating persistent positive photoconductivity (PPC) and persistent negative photoconductivity (NPC). These devices can effectively mimic synaptic behavior, enhance information perception in complex environments, reduce power consumption, improve recognition accuracy, and streamline hardware design.