Synaptic behavior and STDP of asymmetric nanoscale memristors in biohybrid systems

Abstract

We fabricate and characterize asymmetric memristors which show a very strong single-sided hysteresis. When biased in one direction there is hysteresis and in the opposite direction there is a lack of hysteresis. We demonstrate that this apparent lack is actually hysteresis on a much faster time-scale. We further demonstrate that this form of asymmetric behavior correlates very well to the asymmetric structure and function of an actual synapse. The asymmetric memristor device presented here is necessary to correctly implement spike-timing-dependent-plasticity STDP in mixed memristor/neuron hybrid systems as an artificial synapse. These devices show the required characteristics for implementing the asymmetric form of long-term potentiation (LTP) and long-term depression (LTD) of a synapse between two neurons, where symmetric memristor devices do not. Signals from a presynaptic neuron are sent via its axon across the synapse to the dendrite of a postsynaptic neuron. Postsynaptic neuron signals sent to subsequent neurons have an influence on the strength of any further presynaptic neuron signals received by the postsynaptic neuron across the synapse. These signals are grouped into spike triplets within the framework of STDP and, as we experimentally show here, can be implemented with asymmetric memristors, not standard symmetric memristors.

A Dual‐Organic‐Transistor‐Based Tactile‐Perception System with Signal‐Processing Functionality

“…Many artificial synaptic devices, which are bionic signal‐processing elements, have been explored by using both two‐terminal memristors and three‐terminal transistors . Take a two‐terminal memristor for example, the conductance of several inorganic devices can be incrementally adjusted by modulating the charge or flux through it to perform as an artificial synaptic element . In addition, transistors can provide an effective platform for the construction of artificial synapses since the channel conductivity of a transistor can be modulated by the gate electrode to emulate the synaptic functions .…”

Modeling the Formation Process of Grouping Stimuli Sets through Cortical Columns and Microcircuits to Feature Neurons

“…These discrepancies and influences in weight which change according to applied learning rules should be investigated further by alternatively extending the model to include STDP learning rules. We have, for example, demonstrated that STDP learning behavior can be realized in a biohybrid synapse built as a memristor [48]. It is feasible to implement the model in neuromorphic hardware as in parallel memristor bridge synapse-based neural networks, as in an evolving spiking neural network with temporal spike learning, or as a model with implemented dendritic delays [49–51].…”

Organic memristive device as key element for neuromorphic networks

“…However, they are considered as very important elements in new types of computational systems. In particular, several works on memristive devices were done for the mimicking of synapse properties [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]. The activity in this directions are in progress and several important results have been reported.…”