In-depth analysis on electrical parameters of floating gate IGZO synaptic transistor affecting pattern recognition accuracy.

The reliable conductance modulation of synaptic devices is key when implementing high-performance neuromorphic systems. Herein, we propose a floating gate indium gallium zinc oxide (IGZO) synaptic device with an aluminum trapping layer to investigate the correlation between its diverse electrical parameters and pattern recognition accuracy. Basic synaptic properties such as excitatory postsynaptic current, paired pulse facilitation, long/short term memory, and long-term potentiation/depression are demonstrated in the IGZO synaptic transistor. The effects of pulse tuning conditions associated with the pulse voltage magnitude, interval, duration, and cycling number of the applied pulses on the conductance update are systematically investigated. It is discovered that both the nonlinearity of the conductance update and cycle-to-cycle variation should be critically considered using an artificial neural network simulator to ensure the high pattern recognition accuracy of Modified National Institute of Standards and Technology (MNIST) handwritten digit images. The highest recognition rate of the MNIST handwritten dataset is 94.06% for the most optimized pulse condition. Finally, a systematic study regarding the synaptic parameters must be performed to optimize the developed synapse device.

In-depth analysis on electrical parameters of floating gate IGZO synaptic transistor affecting pattern recognition accuracy.

The reliable conductance modulation of synaptic devices is key when implementing high-performance neuromorphic systems. Herein, we propose a floating gate IGZO synaptic device with an aluminum trapping layer to investigate the correlation between its diverse electrical parameters and pattern recognition accuracy. Basic synaptic properties such as excitatory postsynaptic current, paired pulse facilitation, long/short term memory, and long-term potentiation/depression are demonstrated in the IGZO synaptic transistor. The effects of pulse tuning conditions associated with the pulse voltage magnitude, interval, duration, and cycling number of the applied pulses on the conductance update are systematically investigated. It is discovered that both the nonlinearity of the conductance update and cycle-to-cycle variation should be critically considered using an artificial neural network simulator to ensure the high pattern recognition accuracy of Modified National Institute of Standards and Technology (MNIST) handwritten digit images. The highest recognition rate of the MNIST handwritten dataset is 94.06% for the most optimized pulse condition. Finally, a systematic study regarding the synaptic parameters must be performed to optimize the developed synapse device.

Glutamate-AMPAR interaction in a model of synaptic transmission.

Over the last several years we have investigated the excitatory synaptic response by means of a mathematical model based on a detailed description of the synapse geometry, the Brownian motion of Glutamate molecules and their binding to postsynaptic receptors. Recently, the basic model has been modified for the numbers, the size and the 3D structure of receptors according to new data from the literature. Some results of simulations performed with the updated model are shown here. They were aimed to study the synaptic response in relation to the binding probability, to the probable height of the receptors in the synaptic cleft, and to the space-time distribution of Glutamate/Receptor collisions. A first series of simulations permitted to determine a possible range of values for the binding probability of Glutamate to receptors. Other simulations, investigating the changes induced on the synaptic response by the variations of the height of AMPA receptors in synaptic cleft, allowed to identify the height producing the higher amplitude peak of the mEPSCs. Finally, two new statistical descriptors for analyzing the synaptic response were presented. The first is based on the study of the space distribution of the number of Glutamate/Receptor collisions. Simulations investigating the effects of an increasing eccentricity of the releasing vesicle allowed assessing this method. The second one considers the inter-collision times between Glutamate molecules and binding sites. The results of some of the last simulations demonstrated its capacity to highlight the subtleties and the randomness underlying the activation of the receptors. This article is part of a Special Issue entitled Neural Coding 2012.