UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Automated subcortical brain segmentation using multispectral MRI for improved AD diagnosis and disease tracking

Manning, Emily Ndakola; (2018) Automated subcortical brain segmentation using multispectral MRI for improved AD diagnosis and disease tracking. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of EmilyManning_PhDthesis_corrected.pdf]
Preview
Text
EmilyManning_PhDthesis_corrected.pdf - Accepted Version

Download (12MB) | Preview

Abstract

This thesis is a detailed investigation into subcortical changes in Alzheimer's disease (AD). Hippocampal volumes, shapes and diffusion metrics were investigated in different disease stages and presentations, and the ability of these metrics for disease group classification investigated. A new method for automated thalamic segmentation using multimodal imaging was developed and applied to two different datasets. Hippocampal volumes were found to be disproportionately affected by the apolipoprotein (APOE) e4 allele. Hippocampal volumes were also found to be reduced in subjects with the posterior cortical atrophy variant of AD. These changes were localized in the hippocampal tail region and hippocampal shape metrics were found to be superior to hippocampal volumes in differentiating these subjects from controls. The manual thalamic segmentation protocol developed was found to have good reliability, and a template library of thalamic segmentations was generated for use in automated pipelines. The manual segmentation protocol used both T1-weighted and diffusion magnetic resonance imaging (MRI) scans for improved segmentation accuracy. The template library was used for automatic segmentation of the thalamus and leave-one-out cross-validation revealed good segmentation reliability, better than that reported by the most widely used automated thalamic segmentation techniques. The thalami from subjects from the Alzheimer's disease neuroimaging initiative (ADNI)-GO/2 datasets, which includes control subjects, subjects with subjective memory complaints, with early mild cognitive impairment (EMCI), late mild cognitive impairment (LMCI) and AD, were segmented using the automated thalamic pipeline. Subjects with AD and mild cognitive impairment (MCI) were found to have lower thalamic volumes, as well as lower hippocampal volumes suggesting early thalamic involvement. Differences in diffusion metrics were found and some diffusion metrics were associated with subsequently higher atrophy rates. The inclusion of hippocampal and thalamic diffusion metrics, in addition to volumes were found to improve disease group classification. In summary, the work in this thesis extends existing knowledge about how the hippocampi and thalami are affected in Alzheimer's disease.

Type: Thesis (Doctoral)
Qualification: Ph.D
Title: Automated subcortical brain segmentation using multispectral MRI for improved AD diagnosis and disease tracking
Event: UCL
Open access status: An open access version is available from UCL Discovery
Language: English
UCL classification: UCL
UCL > Provost and Vice Provost Offices
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences > UCL Queen Square Institute of Neurology
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences > UCL Queen Square Institute of Neurology > Neurodegenerative Diseases
URI: https://discovery.ucl.ac.uk/id/eprint/10046438
Downloads since deposit
210Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item