Meeting Banner
Abstract #2083

Modelling and Removing the Gradient Artefact Using a Gradient Model Fit (GMF)

Glyn S. Spencer1, Karen J. Mullinger1, Andrew Peters1, Richard Bowtell1

1SPMMRC, School of Physics and Astronomy, University of Nottingham, Nottingham, Nottinghamshire, United Kingdom

The EEG gradient artefact (GA) is formed from a linear superposition of individual artefacts generated by the orthogonal gradients. Variations in position of the head and/or EEG leads scale the relative weighting of the artefact contribution from the different gradient channels. Here we verify this concept and use it to introduce a novel GA correction method, which is based on a gradient model fit (GMF). Our results show that GMF performs better than conventional GA correction methods at high frequency when subject movement occurs, potentially providing a more robust method for GA correction when investigating gamma (30-150Hz) frequency neuronal activity.

Keywords

accurate acquisition acquisitions activity adaptively additional adjust affected agar aimed amplifier amplitude applied asked assumption attenuated averaging axes bands beginning better blue brain caused changing channel channels clocks coil component components comprising concept corrected correction customized described difficult dominance employed ensure entire experiment field filtering fitted focus focusing form formed frequency full functions gating generated gradient gradients greater green head healthy held highly identified imperative improvement improving included individual initial introduce inverse larger lead leads leave linear made matrix maximal model models moved movement movements neuronal novel nulled occur often orthogonal particularly pass performance period phantom physics position positioning potentially presence prior produced products pulse read recorded recording recordings reflected removing repeated repeats requires requiring residual residuals root sampling saturation scaled scaling scanner school selection similarly simulating slice slices spanning spherical square subject subtracted subtraction suggesting superposed superposition system table template temporal unaffected unattenuated unfiltered upon varying waveforms weightings whose widely