Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

The Use of Multiple Group Outlier Detection Methods to Identify Informative Brain Regions in Magnetic Resonance Images

Pugh, Nathan (2013) The Use of Multiple Group Outlier Detection Methods to Identify Informative Brain Regions in Magnetic Resonance Images. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

This is the latest version of this item.

[img]
Preview
PDF
Primary Text

Download (3MB) | Preview

Abstract

The discovery of genetic markers that exhibit differential expression is of great interest in cancer studies. Researchers have now looked to find ways to identify genes with different expression patterns that exist only in a subset of the disease samples. Recently, a class of outlier detection methods has been developed to search for genes with outlier subsets. Using this approach, results in increased power to detect differences across groups relative to standard methods for group comparisons. Outlier detection has also been extended to handle multiple disease groups that are relevant to many more studies. The purpose of this research is to provide a comprehensive review of the class of two-group outlier detection methods which has been limited to date. From these results a modification is proposed to an existing method and the performance of this modification is examined via simulation studies. In addition, three extensions of two-group outlier detection methods are proposed to handle multiple group comparisons. Lastly, a novel application of these methods to structural magnetic resonance imaging data to identify informative brain regions related to cognitive decline in elderly adults is discussed.
Public Health Significance: Outlier detection is a significant contribution to public health as a method that allows researchers to investigate high-dimensional data where issues such as heterogeneity and multiple comparisons are problematic. These methods allow for the identification of factors, such as genes or brain regions, that are related to group membership while identifying homogeneous subpopulations in the data.


Share

Citation/Export:
Social Networking:
Share |

Details

Item Type: University of Pittsburgh ETD
Status: Unpublished
Creators/Authors:
CreatorsEmailPitt UsernameORCID
Pugh, Nathannap23@pitt.eduNAP23
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairWeissfeld, Lisa Alweis@pitt.eduLWEIS
Committee MemberLin, Yanyal14@pitt.eduYAL14
Committee MemberAizenstein, Howard Jaizensteinhj@upmc.eduAIZEN
Committee MemberChang, Joycechangejh@umpc.edu
Date: 27 June 2013
Date Type: Publication
Defense Date: 17 April 2013
Approval Date: 27 June 2013
Submission Date: 3 April 2013
Access Restriction: 5 year -- Restrict access to University of Pittsburgh for a period of 5 years.
Number of Pages: 67
Institution: University of Pittsburgh
Schools and Programs: Graduate School of Public Health > Biostatistics
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Outlier Detection, Structural Magnetic Resonance Imaging, High Dimensional Data, Detection Power, False Discovery Rates
Date Deposited: 27 Jun 2013 19:00
Last Modified: 01 May 2018 05:15
URI: http://d-scholarship.pitt.edu/id/eprint/18597

Available Versions of this Item

  • The Use of Multiple Group Outlier Detection Methods to Identify Informative Brain Regions in Magnetic Resonance Images. (deposited 27 Jun 2013 19:00) [Currently Displayed]

Metrics

Monthly Views for the past 3 years

Plum Analytics


Actions (login required)

View Item View Item