Realizing diverse STDP learning rules in synaptic circuit based on memristor

Being a nonlinear circuit component with memory function, a memristor is similar to a synapse in human brain, whose conductance can be changed after electrical stimulation. It can be used to simulate a synaptic behavior in the process of learning and memory. In this work, a synaptic circuit based on memristor is realized, which includes an enhancement module, a suppression module, and a memristive synapse module. The enhancement module and the suppression module are composed of op-amps, logic gate, and analog switch, etc., while the memristive synapse module consists of a memristor and an analog switch. By inputting a pair of pulsed DC actuators in the enhancement and suppression modules, the stimulation of bio-synapse from pre-neuron to post-neuron is simulated. By adjusting the time interval between pulse input signals, it is found that the shorter the signal interval, the more remarkable the change in conductance will be achieved by the memristor, which is consistent with the change in spike-timing dependent plasticity (STDP) learning curve of bio-synapse. In order to achieve diversity of the synaptic simulation, four replacement circuits of the memristive synapse module are proposed, each of which can simulate different learning rules on its own. Therefore, the circuits can simulate four kinds of synaptic STDP learning rules, which can solve problems such as single-type simulation, harsh input conditions, etc. in synaptic circuit research, and these circuits are expected to be applied in the development of neuromorphic chips in the future.