Feedback

Faculté des Sciences appliquées
Faculté des Sciences appliquées
MASTER THESIS
VIEW 124 | DOWNLOAD 492

Multivariate Statistics for the Joint Analysis of Quantitative Maps

Download
Salvoni, Geoffrey ULiège
Promotor(s) : Phillips, Christophe ULiège ; Callaghan, Martina
Date of defense : 25-Jun-2018/26-Jun-2018 • Permalink : http://hdl.handle.net/2268.2/4535
Details
Title : Multivariate Statistics for the Joint Analysis of Quantitative Maps
Translated title : [fr] Statistiques multivariées pour l'analyse conjointe de cartes quantitatives
Author : Salvoni, Geoffrey ULiège
Date of defense  : 25-Jun-2018/26-Jun-2018
Advisor(s) : Phillips, Christophe ULiège
Callaghan, Martina 
Committee's member(s) : Van Steen, Kristel ULiège
Vandewalle, Gilles ULiège
Language : English
Number of pages : 72
Discipline(s) : Engineering, computing & technology > Multidisciplinary, general & others
Institution(s) : Université de Liège, Liège, Belgique
Wellcome Trust Centre for Neuroimaging, Londres, Royaume-Uni
Degree: Master en ingénieur civil biomédical, à finalité spécialisée
Faculty: Master thesis of the Faculté des Sciences appliquées

Abstract

[en] Life expectancy increase makes the need for understanding the mechanisms underlying ageing more essential. The ongoing development of magnetic resonance imaging (MRI) provides researchers with an innovative tool to study the brain non-invasively. In particular, quantitative MRI produces high-resolution quantitative maps of parameters dependent of relevant biological measures such as myelin, iron and water concentrations. In addition, a voxel-based quantification (VBQ) procedure corresponding to a spatial normalisation and a smoothing of the maps facilitates group analyses by creating images that are spatially comparable among participants.
On one hand, this work aims at studying the correlation with age in the brain microstructures of a group of healthy volunteers (from 19 to 75 years old) using multivariate statistics. To do so, a joint analysis of the quantitative maps of the entire cohort was carried out using permutation tests implemented in the PALM (Permutation Analysis of Linear Models) toolbox and the results were then compared to univariate methods applied in the Statistical Parametric Mapping (SPM) framework. Whole-brain voxel-wise analysis showed extensive correlations with age in the gray and the white matter in line with histologic reports. Nonetheless, a severe loss of sensitivity in the joint analysis was observed due to the lack of signal present in the longitudinal relaxation rate (R1) modality.
On the other hand, the VBQ approach was analysed in order to evaluate the pertinence of this normalisation method combining weighting and smoothing. The age effect embedded in the parameter maps was removed while an age-study was performed on the weights, supposed to be independent of ageing. Voxel-wise analysis revealed correlations with age around the ventricles and in the corpus callosum, corresponding to morphological effects, but strong robustness in the cortex. The effect size proved to be tiny compared to the real age effects though.


File(s)

Document(s)

File
Access SalvoniGeoffrey_Report.pdf
Description: -
Size: 18.11 MB
Format: Adobe PDF

Author

  • Salvoni, Geoffrey ULiège Université de Liège > Master ing. civ. biomed., à fin.

Promotor(s)

Committee's member(s)

  • Van Steen, Kristel ULiège Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Bioinformatique
    ORBi View his publications on ORBi
  • Vandewalle, Gilles ULiège Université de Liège - ULiège > Département de chimie (sciences) > Département de chimie (sciences)
    ORBi View his publications on ORBi
  • Total number of views 124
  • Total number of downloads 492










All documents available on MatheO are protected by copyright and subject to the usual rules for fair use.
The University of Liège does not guarantee the scientific quality of these students' works or the accuracy of all the information they contain.