Abstract
Alzheimer's disease (AD) is a neurodegenerative condition characterised by permanent cognitive decline. The financial burden associated with providing care for those affected by this ailment has escalated to a significant magnitude, amounting to billions of dollars. Moreover, projections indicate that the prevalence of this sickness is anticipated to surge by more than 200% throughout the next 15-year period. Numerous researchers have dedicated their efforts to the field of computer-aided diagnosis (CAD), whereby computational methods are used to discern and diagnose various medical conditions. Nevertheless, the identification of persons afflicted with the condition does not mitigate the adverse consequences. It would be advantageous to identify individuals who may be susceptible to the onset of Alzheimer's disease (AD) at an early stage. This might enhance the efficacy of pharmaceutical interventions and the implementation of preventive measures for Alzheimer's disease (AD). This research examined two imaging modalities, namely magnetic resonance imaging (MRI) and positron emission tomography (PET).The study focused on the investigation and development of a strategy for the selection of slices from three-dimensional (3D) scans for their subsequent conversion into two-dimensional (2D) images. Additionally, the study aimed to integrate the created images from the slice selection technique for both modalities. The purpose of developing the integration was to tackle the issue of insufficient data and consolidate the images to accurately predict mild cognitive impairment (MCI) individuals transitioning to Alzheimer's disease. The features extracted from the generated images were subjected to comparison using machine learning classifiers and pre-trained networks. The developed integration approach demonstrated superior performance compared to existing literature in the field of MRI and PET analysis for predicting the conversion of mild cognitive impairment to Alzheimer's disease, particularly when using full brain imaging.
The work identified the 6-by-6 coronal slice selection technique as the best slice selection technique from series of brain anatomical views and slice selection techniques created for representing the 3D MRI and PET images in predicting MCI’s subjects’ conversion to AD. Secondly, the coronal slice selection identified was used to develop a data integration of the two imaging modalities in predicting MCI’s subjects’ conversion to AD. The image classification process demonstrated a much quicker performance, completing the test in just 0.35% of the time needed by a 3D neural network (3D-CNN) for the identical task. The purpose of using slice selection from 3D images to convert them into 2D is to enhance computational speed and improve accuracy in performance simultaneously and was achieved in the study.
| Date of Award | 12 Apr 2024 |
|---|---|
| Original language | English |
| Supervisor | Simon Parkinson (Main Supervisor), Andrew Crampton (Co-Supervisor) & Shamaila Iram (Co-Supervisor) |