Javascript must be enabled for the correct page display

Single-spiking and bursting activity in a three layer Liquid-state network model: Where is the information?

Klijn, W.F.A. (2010) Single-spiking and bursting activity in a three layer Liquid-state network model: Where is the information? Master's Thesis / Essay, Industrial Engineering and Management.

[img] Text
Wouter_Klijn_Master_Thesis.pdf - Published Version
Restricted to RUG campus only

Download (2MB)

Abstract

This paper presents exploratory results on the influence of bursting spiking behavior on the information content of neuron spikes in a three layered model of a cortical micro column. The bursting behavior is introduced in three parts of this model: the input, the recurrent layers and the output of the cortical micro column. Our hypotheses are as follows: 1. Bursting behavior in the input increases the information content of the neuron spikes in the model. 2. Input injection in layers closer to the output layer of the model increases the information content of the spikes. 3. Bursting in the recurrent layers of the network decreases the information content. 4. Bursting in the output of the column increases the information content of the output neurons spikes. The information content of the neurons in the model is analyzed with two methods: firstly a newly created method for assessing the correlated firing of neurons. This method is based on the FFT response of the pooled action potentials of these neurons. The second analytical tool is the Shannon mutual information metric. This metric is adapted for high temporal precision. This allows quantitative and qualitative tracking of information flow through the network. During the simulations large quantitative difference were found between the different neurons types in the model. Inhibitory neurons lacking direct input contained no information regarding the injected input. The bursting neurons in the model layers and the output contained a larger amount of information when compared with the regular spiking neurons. The bursting neurons retained information for a longer period. The introduction of bursting behavior in the input resulted in a decrease of information contained in the model. Two theoretical advances have been made: first the explicit inclusion of the input as a part of the network and secondly new insights in the visualization of the connections and layering of the cortical micro column. The application of these theoretical approaches allowed an improved functional analysis of the results.

Item Type: Thesis (Master's Thesis / Essay)
Degree programme: Industrial Engineering and Management
Thesis type: Master's Thesis / Essay
Language: English
Date Deposited: 15 Feb 2018 07:45
Last Modified: 15 Feb 2018 07:45
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/9489

Actions (login required)

View Item View Item