ISBN 978-3-8439-2926-4

978-3-8439-2926-4, Reihe Informatik

Elisabeth Stühler
Refining and Developing Features Describing Dementia-related Changes in Structural and Functional Imaging

167 Seiten, Dissertation Universität Konstanz (2016), Softcover, A5

Zusammenfassung / Abstract

This dissertation aims at advancing computer-aided diagnosis of neurodegenerative diseases, in particular when differentiating various stages of dementia. The focus is on developing and refining disease-related features based on structural and functional imaging.

The first part of the thesis addresses the refinement of such features based on cross-sectional functional imaging data. Intensity values or volume measurements are extracted from each volume and a classification of various dementia types is premised on variance-based statistical analysis.

Prior to this, several necessary preprocessing steps are evaluated in order to compare different subjects. When applying the methods developed in the first part of the thesis to structural imaging data it is observable that preprocessing of these high-resolution images is not only time-consuming and prone to error but also that its quality cannot be assessed without the help of clinical experts. The second part of this thesis therefore discusses features which are independent from preprocessing quality in longitudinal data. These newly developed features are based on tracing punctual changes within the brain. Interesting points are furthermore examined in order to detect if they show significantly different behavior over a period of two years for various stages of Alzheimer’s dementia. A classification based on significant features is employed to differentiate dementia in various stages.