Simulation of developing human neuronal cell networks

Abstract

BackgroundMicroelectrode array (MEA) is a widely used technique to study for example the functional properties of neuronal networks derived from human embryonic stem cells (hESC-NN). With hESC-NN, we can investigate the earliest developmental stages of neuronal network formation in the human brain.MethodsIn this paper, we propose an in silico model of maturating hESC-NNs based on a phenomenological model called INEX. We focus on simulations of the development of bursts in hESC-NNs, which are the main feature of neuronal activation patterns. The model was developed with data from developing hESC-NN recordings on MEAs which showed increase in the neuronal activity during the investigated six measurement time points in the experimental and simulated data.ResultsOur simulations suggest that the maturation process of hESC-NN, resulting in the formation of bursts, can be explained by the development of synapses. Moreover, spike and burst rate both decreased at the last measurement time point suggesting a pruning of synapses as the weak ones are removed.ConclusionsTo conclude, our model reflects the assumption that the interaction between excitatory and inhibitory neurons during the maturation of a neuronal network and the spontaneous emergence of bursts are due to increased connectivity caused by the forming of new synapses.

A Computational Model of Interactions Between Neuronal and Astrocytic Networks: The Role of Astrocytes in the Stability of the Neuronal Firing Rate

“…Moreover, noise is applied to each neuron to reproduce background activity. In this fashion, INEX is a computationally-light model that has also been shown of well-reproducing neuronal dynamics of neuronal cultures plated on MEAs (Lenk, 2011;Lenk et al, 2016). For all these reasons, we adopted it as a starting platform for neuronal networks to be complemented by astrocytic coupling.…”

“…To keep the model as computationally light as possible and to maintain biological plausibility, the previously introduced models are combined using relatively simple components that are not accurate descriptions of the processes, but rather descriptive. The parameters in INEX are phenomenological and were fixed using brute force to find sets of parameters that produced results in reasonable ranges (Lenk et al, 2016). By adding the Tsodyks-Markram presynapse model, we introduced short-term memory at the level of individual synapses.…”

“…Thus, we neglected the probability that more than one spike may be emitted by a given neuron during a single time step. At the benefit of computational efficiency, a time step as large as t = 5 ms can be adopted and the INEX network model can still reliably simulate neuronal activity recorded in MEA cultures (Lenk, 2011;Lenk et al, 2016).…”

“…In this paper, we analyzed neuronal activity in the form of spikes and bursts which are cascades of spikes. Synchronous population bursts are characteristics of matured and wellconnected networks (Giugliano et al, 2004;Wagenaar et al, 2006;Lenk et al, 2016). Spike and burst features were calculated using a modified version of the cumulative moving average (CMA) algorithm (Kapucu et al, 2012;Välkki et al, 2017).…”

Image-Based Profiling of Synaptic Connectivity in Primary Neuronal Cell Culture

“…Standardized data analysis strategies have been conceived to objectively categorize and visualize multidimensional data from high-content screens ( Grys et al, 2016 ; Caicedo et al, 2017 ). Computational models are being optimized to aid with the biological interpretation but also to benchmark new image analysis algorithms, and this in turn is supported by improved simulations of cellular models and networks ( Loew and Schaff, 2001 ; Lenk et al, 2016 ). Finally, the models, images and morphofunctional data should be bundled and centralized to make them openly available to the neuroscience community, as was recently done for neuronal morphology ( Parekh and Ascoli, 2013 ; Akram et al, 2018 ).…”