To spike or not to spike: from a change of the excitability type in single neurons to exploring coding properties in network models

Keywords

Loading...
Thumbnail Image

Issue Date

2022-08-13

Language

en

Document type

Journal Title

Journal ISSN

Volume Title

Publisher

Title

ISSN

Volume

Issue

Startpage

Endpage

DOI

Abstract

From single neurons to large neural networks, the brain is constantly carrying and transforming information. Neural coding is a core aspect of information representation, which relies on the stimulusresponse relationship. However, the neural coding properties at a single neuron and network level are still not fully understood. At a single neuron level, the distinction between type 1 (integrator) and type 2 (resonator) neuron excitability is well known and relates to what causes a neuron to spike. However, it is unclear whether the excitability type of a single neuron can shift from type 1 to type 2 due to Acetylcholine (ACh) neuromodulatory effects. Pushing this further, understanding how such single neuron coding properties could arise and interact at a network level is a great challenge considering the available network recording techniques. Here, network modeling is crucial for understanding network behaviour in the brain. With such models, one can study larger populations of neurons and still not lose information about single neuron properties. For instance, with the Poisson Balanced Spiking network (Poisson BSN) model, one can reproduce bio-like network spiking behaviour and still be able to observe single neuron properties. Yet, how the coding properties of neurons from network models arise and are affected by network properties is still unknown. To study the coding properties in single neurons, we used the spike-triggered average (STA) analysis. With this, we investigated the stimulus-response relationship from single neuron in vitro recordings and neurons from network model simulations. From the experimental data analysis, we found that ACh did not cause a shift from type 1 to type 2 excitability. And regarding the network modelling we found that in the Poisson BSN model, network size, spiking cost, and time delays had an effect on changing the neurons’ stimulus-response relationship given by the STA analysis. Our results will allow for future research to model networks with neurons with different coding properties (such as type 1 and type 2) and account for how the interactions between neurons might affect these coding properties. Thus, this is one more step to developing and understanding network models to make them more biologically plausible and more insightful for studying and reproducing how the brain represents information.

Description

Citation

Faculty

Faculteit der Sociale Wetenschappen