Multivariate comparison of voxel-based versus morphometric analysis

munchsistersΤεχνίτη Νοημοσύνη και Ρομποτική

17 Οκτ 2013 (πριν από 3 χρόνια και 10 μήνες)

62 εμφανίσεις

Multivariate comparison of voxel
-
based versus morphometric analysis

Introduction

The need for a better understanding of brain atrophy in normal and pathological conditions has
motivated the development of models derived from anatomical brain images such
as Magnetic
Resonance Imaging (MRI).

Voxel
-
based morphometry (VBM) is a well
-
established technique for medical image processing
and analysis. It allows the comparison of focal differences in brain anatomy by applying statistical
parametric mapping (SPM) ba
sed on general linear modeling (GLM) to obtain probability maps of
significant difference between groups of patients (Ashburner and Friston, 2000).

FreeSurfer is an automated pipeline that produces regional cortical thickness and volumetric
measures (Fisch
l et al., 2002). This segmentation approach has been successfully applied for
multivariate classification of Alzheimer's disease subjects and healthy controls (Westman et al.,
2011).

Objective

Compare anatomical features obtained from FreeSurfer versus sel
ected features derived from VBM
applying multivariate statistics and demonstrate their advantages and disadvantages.


Goals



Get familiar with
s
oftware used in neuroscience research, such as SPM, FreeSurfer, FSL, or
AFNI.



Get familiar with state
-
of
-
the
-
art
multivariate statistical techniques such as: SVM and OPLS.



Feature selection should be implemented using any of validated approaches, for example,
Relief
-
SVM, weights from linear classifiers, principal component analysis, among others.



Comparison of classi
fication performance in a sample consisting of two groups (normal and
abnormal) based on accuracy, sensitivity, specificity, area under the ROC curve.


Start of project March 2012


References

Ashburner, J., Friston, K. J., 2000. Voxel
-
based morphometry


T
he methods. NeuroImage 11, p.
805
-
821.

Fischl, B., Salat, D., Busa, E., Albert, M., Dieterich, M., Haselgrove, C., van der Kouwe, A.,
Killiany, R., Kennedy, D., Klaveness, S., Montillo, A., Makris, N., Rosen, B., Dale, A., 2002.
Whole Brain Segmentation::
Automated Labeling of Neuroanatomical Structures in the Human
Brain. Neuron 33, 341
-
355.

Westman, E., Simmons, A., Zhang, Y., Muehlboeck, J.S., Mecocci, P., Vellas, B., Tsolaki, M.,
Kloszewska, I., Soininen, H., Weiner, M.W., Lovestone, S., Spenger, C., Wa
hlund, L.O., for the
AddNeuroMed consortium and the Alzheimer’s Disease Neuroimaging Initiative, 2011.
AddNeuroMed and ADNI: Similar patterns of Alzheimer’s atrophy and automated MRI
classification accuracy in Europe and North America. NeuroImage, in pres
s, corrected proof,
available online 1st July 2011.


Contact

Eric Westman
eric.westman@ki.se