VLSI implementable neuron-astrocyte control mechanism
Astrocyte model, Field Programmable Gate Array (FPGA), Izhikevich neuron model, Neural glial interactions, Synapse
Although functionalities of the brain are mostly explained using nerve cells and synapses, but there are a larger number of other types of cells providing support and protection for neurons in the central nervous system (CNS). Among a variety of these non-neuronal cells, astrocytes are the most abundant cell types which act as supportive cells in the brain. The important role of astrocyte such as extracellular regulation, synaptic information regulation, neuronal synchronization and feedback to neural activity, make the astrocytes play a vital role for interacting with neurons in the brain. On the other hand, astrocytes provide feedback to neurons for modulating their neuronal activity. This paper presents a set of equation as a model to describe the mechanisms of neural glial interactions which are implementable on digital platforms. The proposed model can be utilized in large scale VLSI implementation of the biological neural networks, i.e. neuromorphic systems. Simulation results show that the model can precisely reproduce different behavior of neural glial interactions. The proposed model is investigated in terms of digital implementation feasibility and costs while targeting low-cost hardware realization. Hardware synthesis and physical realization on FPGA shows that the proposed model can reproduce different pathway behavior in neural glial mechanisms with higher performance and lower implementation cost compared with the original model.
Haghiri, Saeed; Ahmadi, Arash; and Saif, Mehrdad. (2016). VLSI implementable neuron-astrocyte control mechanism. Neurocomputing, 214, 280-296.