Use this URL to cite or link to this record in EThOS:
Title: Spatial point process models for MRI lesion data in multiple sclerosis
Author: Taschler, Bernd
ISNI:       0000 0004 6423 1908
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
Over the past three decades neuroimaging techniques in general and magnetic resonance imaging (MRI) in particular have made large contributions to the understanding of human brain function and to the diagnosis and treatment of neurological diseases. One area of wide-spread clinical use of MRI is in the assessment of multiple sclerosis (MS). MS patients present with lesions – areas of decreased neuronal conductivity, akin scarred tissue – that occur across the brain and spinal cord. There has been growing interest to use quantitative measures of lesion incidence, exact lesion location and the shape of lesions in the analysis of MRI-based lesion data. Our objective is to address some of the limitations of current methods which rely on particular assumptions about the data and mostly ignore any spatial correlation and structure in the data. In this work we explore several ways of incorporating multiple sources of information into models that can be used for classification and prediction purposes. We compare and assess different machine learning and Bayesian spatial models in a classification task based on MS lesion data. Furthermore, we propose an extended doubly-stochastic spatial point process model based on Gamma random fields that includes non-imaging data as well as location-specific measures attached to xyz-coordinates, and use both simulated and real data to evaluate our methods.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: RC Internal medicine