Advanced

Developing and evaluating strategies to deal with motion in fMRI using Independent Component Analysis (ICA)

Peterson, Mikael (2007)
Medical Physics Programme
Abstract (Swedish)
Functional Magnetic Resonance Imaging, fMRI, is a non invasive method to map brain activation with a magnetic resonance camera. In this thesis we propose a method to deal with motion induced artefacts in functional fMRI, using Independent Component Analysis, ICA. Patient movement can induce false activation or cause loss of true activation. Movement is one of the most common causes of fMRI failure and it is therefore crucial to minimize the effects of motion on fMRI time series. It has recently been shown that data driven methods, like ICA have been successful in identifying and removing artifactual variance in fMRI time series. We suggest a method were we try to identify the motion induced variance using ICA and remove it before... (More)
Functional Magnetic Resonance Imaging, fMRI, is a non invasive method to map brain activation with a magnetic resonance camera. In this thesis we propose a method to deal with motion induced artefacts in functional fMRI, using Independent Component Analysis, ICA. Patient movement can induce false activation or cause loss of true activation. Movement is one of the most common causes of fMRI failure and it is therefore crucial to minimize the effects of motion on fMRI time series. It has recently been shown that data driven methods, like ICA have been successful in identifying and removing artifactual variance in fMRI time series. We suggest a method were we try to identify the motion induced variance using ICA and remove it before analysis.Our results indicate that our method does not increase the analysis sensitivity for true activation. In most cases standard analysis is a better or equally good alternative. (Less)
Please use this url to cite or link to this publication:
author
Peterson, Mikael
supervisor
organization
year
type
H2 - Master's Degree (Two Years)
subject
keywords
MRI
language
English
id
2157031
date added to LUP
2011-09-13 15:23:56
date last changed
2011-09-13 15:23:56
@misc{2157031,
  abstract     = {Functional Magnetic Resonance Imaging, fMRI, is a non invasive method to map brain activation with a magnetic resonance camera. In this thesis we propose a method to deal with motion induced artefacts in functional fMRI, using Independent Component Analysis, ICA. Patient movement can induce false activation or cause loss of true activation. Movement is one of the most common causes of fMRI failure and it is therefore crucial to minimize the effects of motion on fMRI time series. It has recently been shown that data driven methods, like ICA have been successful in identifying and removing artifactual variance in fMRI time series. We suggest a method were we try to identify the motion induced variance using ICA and remove it before analysis.Our results indicate that our method does not increase the analysis sensitivity for true activation. In most cases standard analysis is a better or equally good alternative.},
  author       = {Peterson, Mikael},
  keyword      = {MRI},
  language     = {eng},
  note         = {Student Paper},
  title        = {Developing and evaluating strategies to deal with motion in fMRI using Independent Component Analysis (ICA)},
  year         = {2007},
}