| Bibliography | Prölß, Till: Implement the RIDE Algorithm Into the Unfold.jl Toolbox.
University of Stuttgart, Faculty of Computer Science, Electrical Engineering, and Information Technology, Master Thesis No. 5 (2025).
41 pages, english.
|
|---|
| Abstract | Event-relatedpotentials(ERPs)arecommonlyusedinEEGresearchtostudybrain responsestospecificstimuli.TraditionalERPaveragingmethodsareaffectedbytrial-to- trial latencyvariability,whichcandistorttheextractedcomponents.ResidualIteration DecompositionEstimate(RIDE)isanestablishedmethoddesignedtocorrectforthis variabilitybyseparatingcomponentsbasedontheirlatencydistributions.Whileeffective, its originalMATLABimplementationlacksflexibilityandintegrationwithmodernEEG analysis frameworks. This workpresentsaJulia-basedimplementationofRIDEanditsintegrationintoUn- fold, aregression-baseddeconvolutiontoolboxforEEGanalysis.Twoversionswerede- veloped:ClassicRIDE,whichreplicatesMATLABRIDE’sbehavior,andUnfoldRIDE, whichreplacesiterativedecompositionwithregression-baseddeconvolution.Bothim- plementationswereevaluatedusingsimulateddatasetswithknowngroundtruthanda real-worldEEGdatasettoassesstheiraccuracyandcomparethemwithMATLABRIDE. The resultsshowthatClassicandUnfoldRIDEsuccessfullyreconstructERPcom- ponents,producingreasonableresultsinbothdatasets.However,differencesexistwhen compared toMATLABRIDE,likelyduetomissingfeaturesandvariationsinfiltering,la- tency estimation,anditerationbehavior.Despitethesediscrepancies,thefindingsconfirm that Unfold’sregression-basedapproachisaviablealternativetoiterativedecomposition, offering potentialadvantagessuchasimprovedmodelingflexibility. Futureresearchshouldfocusonimplementingmissingfeaturesandachievingequiv- alence withMATLABRIDE.Additionally,randomizedtestingwithUnfoldSimcould providesystematicvalidationandhelpoptimizethealgorithm.Thisworklaysthefoun- dation forfurtherimprovementsinEEGdeconvolutionbycombiningRIDE’slatency estimation withUnfold’sflexibleregressionframework. The implementedalgorithmsareavailableasanopensourceprojectonGitHub: https://github.com/unfoldtoolbox/UnfoldRIDE.jl
|
|---|
| Department(s) | University of Stuttgart, Institute of Artificial Intelligence, Machine Learning for Simulation Science
|
|---|
| Superviser(s) | Ehinger, Jun.-Prof. Benedikt; Skukies, René |
|---|
| Entry date | May 13, 2025 |
|---|