Summary:
This diploma thesis deals with the study and analysis of a functional Magnetic
Resonance Imaging experiment to social-cognitive deficits in normal aging.
Moreover, it
focuses on the development and investigation of mathematical approaches for
modeling
certain cerebral function mechanisms. In particular, one fMRI experiment is
conducted
by activating the human brain under conditions of false belief.
Advanced spatial processing techniques are applied to all collected data. In
particular, fMRI images are realigned so as to remove movement artifact,
normalized to
transform the images into a certain template and smoothed by convolving the
images
with a Gaussian kernel of a specified width so as to suppress noise. The
localization of
activated brain regions is achieved through the use of appropriate regression
models
and statistical tests methods. The methods which were used to do statistical
analysis
were the General Linear Model (GLM) and the Independent Component Analysis
(ICA).
Specifically, the experiment's participants were divided into two age groups,
one
of young adults and one of elderly persons. All participants were scanned using
the
funcional Magnetic Reasonance Imaging while being subjected to false belief
techniques. The produced images were pre-processed using the aforementioned
techniques, and were subjected to statistical analysis using the two
aforementioned
methodologies separately using the SPM12 toolbox and the Group ICA toolbox.
Finally the results gathered from the two age groups were compared, first
individually using each methodology subsequently leading to the result that
socialcognitive
deficits exist in normal aging, and then between methodologies, which
highlighted several common points in their results and underlined some key
differences.
Keywords:
funcional Magnetic Reasonance Imaging, General Linear Model, Independent Component Analysis, Independent Component Analysis, false-belief task, SPM12 and Group ICA
