Two Suggested Probabilistic and Kinetic Models for Astrocytic Network in Spiking Neural Networks

Abstract

Astrocytes, the predominant glial cell type in the brain, were traditionally considered as merely passive supportive cells without any important roles in synaptic information processing. In contrast, the contemporary view was given rise to show that astrocytes play active roles in synaptic neurotransmission and information processing. Hence, recently two terms have been emerged, tripartite synapse, to describe the communication between an astrocyte and two neurons, and the term astrocytic syncytium or astrocytic network to describe the communication among the astrocytes by gap junction. Therefore, we propose mathematical models for tripartite synapseand astrocytic syncytium based on twostate kinetics models, several probabilistic methods and Spiking Neural Network (SNN) to introduce new Artificial Astrocytic Syncytium (AAS) model. The simulation results have shown that proposed model could represent the cellular intrinsic properties of astrocyte based on the spatial and temporal aspects to emulate the astrocytic network functions related to cognitive, learning and memory.

Analysis of Network Models with Neuron-Astrocyte Interactions

“…None of the analyzed neuron-astrocyte network models, however, studied this phenomenon. Nevertheless, eight of the models, equipped by diverse mechanisms and interaction schemes related to neurotransmission and gliotransmission, were primarily focused on studying signal or information transfer in neuron-astrocyte networks (Liu & Li, 2013a , b ; Yang & Yeo, 2015 ; Li et al., 2016 ; Kanakov et al., 2019 ; Nazari & Faez, 2019 ; Abed et al., 2020 ; Nazari et al., 2020 ).…”

“…This table lists several details for each study: the tool or programming language used, code availability online, brain area, experimental data used, number of neurons and astrocytes, and experimentally shown neural function or other function that the model was finetuned to capture Study Tool/ availability Brain area/data Neurons Astrocytes Function Abed et al. ( 2020 ) n/a Generic 10,000 E & I n/a Sgn./Inf. Aleksin et al.…”

“…( 2020 ), and Study NN interactions NA interactions AN interactions AA interactions Abed et al. ( 2020 ) EE/EI/IE/II: EA/IA: AE/AI: , AA: IP 3 via GJs Aleksin et al. ( 2017 ) EE/II(2), EI/IE(0–1): EA/IA(0–1): AE/AI(0–1): AA(2): Ca 2+ via GJs Allegrini et al.…”

“…Thus, we used different notation in the third column compared to the fourth column, for example EA(AE) means that we considered both, the interaction from E to A and from A to E. If both interactions existed between the same two cells, the model was categorized as (locally) recurrent, if one of them did not exist, the model was (locally) feed-forward Study Spatial organization of cells Structure of interaction scheme Direction of information flow Abed et al. ( 2020 ) All cells/E/I/A: 2D, random placement EE/EI/IE/II: all-to-all; EA/AE/IA/AI: one-to-one; AA: n/a Global: recurrent. Local: EE/EI(IE)/II/AA: recurrent; EA(AE)/IA(AI): feed-forward Aleksin et al.…”