April 28, 2014 -
2:00 pm to 3:00 pm
Location: Cupples I, Room 199 |
Host: Prof. Jiming Ding
A Master in Statistics Thesis Defense
Abstract: Alzheimer's disease is most common form of dementia. The cause and development of Alzheimer's disease are not completely understood. Studies results show that the disease is associated with plaques and tangles in the brain (Tiraboschi, P., et al., 2004). Alzheimer's disease is diagnosed among people who are over 65 years old and the incidence rates increases exponentially after 65 years of age (Brookmeyer, 1998). Several studies have indicated that structural MRI can help study AD. The typical set of MRI data contains full brain scanned images from different individuals, with intensities on the same two-dimensional domain for each slice. The main purpose of this type of image data is to combine all the information from images to make statistical inferences about the effects on the populations represented in the images (Morris, Jeffrey S., et al. 2010). The Open Access Series of Imaging Studies (OASIS) project freely distributes magnetic resonance imaging (MRI) data sets of the brain. Shearlets are becoming one of the best frameworks for representing the multidimensional data efficiently. Shearlet decomposition is applied to the MRI data set by choosing a configuration. After decompositions, we had a large sparse coefficients matrix. We used different methods for data size reduction in order to perform a meaningful statistical analysis. Variable ranking feature selection method is applied and then Principle Components Analysis was used for feather data reduction. At last, logistic regression was applied to the OASIS data set with principle components. In the result of the logistic regression, gender, age, education and some principle components significantly affect the CDR, which is a measurement for AD.
Contact Us | Home Page
Department of Mathematics | Washington University in St. Louis | One Brookings Drive, St. Louis, MO 63130-4899 |
contact@math.wustl.edu