- Title
- Atlas-based assessment and diagnosis of Alzheimer's Disease from neuroimages
- Creator
- Qian, Guoyu
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2010
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- The human brain is the most complicated biological organ. Despite the intensive research on the brain, it still remains a great mystery. Neurodegenerative diseases such as dementia are recognized as a major health problem and are becoming increasingly common at present time, particularly with aging population. This thesis investigates the problems and limitations of current approach of assessment and diagnosis for Alzheimer’s disease (AD), a major form of dementia, and presents a rapid and automatic method of cognitive assessment and disease diagnosis by processing the neuroimages and performing the statistical analysis. The neuroimage processing in this thesis is based on a set of fully automatic image processing algorithms and a digital brain atlas with accurate brain structures segmented and labeled, including the AD-specific structures. The image processing algorithms extract the brain areas from the neuroimages, detect the landmarks on the images, and segment the AD-specific structures from the images by using the digital AD-specific brain atlas. They are presented chapter by chapter in this thesis. The brain atlas is constructed based on a high resolution magnetic resonance imaging volumetric dataset by using a set of powerful and intelligent tools also presented in this thesis. The algorithms for automated brain extraction from structural and functional neuroimages are presented in the thesis. They include a domain knowledge based brain extraction algorithm for structural computed tomography images and a rapid cerebral and cerebellar region extraction from functional positron emission tomography (PET) images. To increase the accuracy of PET images registration into the atlas space by the piecewise linear transformation, a new landmark is defined in this thesis to extend the existing Talairach landmarks, a set of commonly used landmarks in human brain registration, in order to include the cerebellum into the space. The cerebellum is an important brain structure for our research due to its role as an intensity normalization reference. The algorithm for automatic detection of the new landmark as well as the other Talairach landmarks is presented. According to the linkage between the AD diagnosis and cognitive scores like mini mental state examination, and the correlation between the cognitive scores and the changes of several specific brain structures on the neuroimages, the statistical models of stepwise regressions and discriminant classification are performed on the regions of AD-specific structures to calculate the cognitive scores and classify the experiment subjects into different diagnostic groups automatically. The approach has been applied to hundreds of cases and shown promising results. This is the first effort to quantitatively calculate the cognitive scores by processing the neuroimages automatically. It provides an objective, efficient, less expensive, and extendable way for potential clinical diagnosis in the patients with dementia by the fully automatic computer programs.
- Subject
- brain; Alzheimer's disease; neuroimages
- Identifier
- http://hdl.handle.net/1959.13/805577
- Identifier
- uon:6892
- Rights
- Copyright 2010 Guoyu Qian
- Language
- eng
- Full Text
- Hits: 1266
- Visitors: 1449
- Downloads: 350
Thumbnail | File | Description | Size | Format | |||
---|---|---|---|---|---|---|---|
View Details Download | ATTACHMENT01 | Abstract | 127 KB | Adobe Acrobat PDF | View Details Download | ||
View Details Download | ATTACHMENT02 | Thesis | 4 MB | Adobe Acrobat PDF | View Details Download |