Cortical Dynamics Consensus set approximation
Keywords
Loading...
Authors
Issue Date
2021-07-01
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Consensus set approximation refers to a technique that can be used to solve
a Dual estimation problem, a problem in which one wishes to know both the
signal in a dataset and the parameters for a model. Solving such problems
for neurological systems can provide a low cost method for retrieving the
state of e.g. the cortex in the brain.
Previous work made it possible to use this method to approximate networks
using this technique. In this research this method is expanded to solve this
problem with the added constraint of attempting to discover regions with
an anatomical basis. This is done by using such an approximation applied
to an Unscented Kalman Filter with a Wilson-Cowan model in combination
with A nity Propagation clustering.
Two model variations were tested on several datasets containing voltage
traces of ve mice across their dorsal cortices. Results show that not all
data can be signi cantly approximated, however networks similar to other
studies can be found. Future research could improve the numerical stability
of the current method and optimize both noise and transition models in
order to improve current results.
Keywords: Kalman Filter, Unscented Kalman Filter, Wilson-Cowan Model,
Consensus set, Data Assimilation, Dual Estimation
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen