REMOVING ARTIFACTS FROM EEG RECORDED WITHIN MR SCANNER BY DYNAMICAL TEMPLATE STATE SPACE APPROACH

Andreas Galka, Laith Hamid, Ulrich Stephani

Keywords

Time Series Analysis, Filtering, State Space Modelling,Artifact Removal, Electroencephalogram, Functional Mag-netic Resonance Imaging

Abstract

We propose a novel state space modelling approach to removing scanner-related artifacts from electroencephalo- grams recorded inside MR scanners. For this purpose, dynamical templates for the actual brain activity and the ballistocardiogram are obtained from a short piece of data recorded without fMRI scanning; dynamical templates for the scanner artifacts are obtained from data recorded dur- ing fMRI scanning. Finally the two sets of dynamical tem- plates are merged. We compare our approach with Indepen- dent Component Analysis and find superior performance.

Important Links:



Go Back