Blood oxygen level dependent imaging of cerebral mesostructure

Thesis

In this thesis I investigate blood oxygen level-dependent (BOLD) MRI methods of imaging the cerebral blood volume (CBV), mean vessel radius and oxygen extraction fraction (OEF). Through the investigation of these individual techniques a new framework is proposed for the simultaneous measurement of all three parameters, providing a comprehensive assay of the cerebral mesostructure.

A new method for the segmentation of blood filled voxels from the sagittal sinus is presented. The implemented method is completely automated and thus removes user bias in voxel selection. The segmentation method is used in a volunteer study to calculate CBV from a hyperoxic challenge according to an existing technique. CBV measurements from this study are found to be significantly overestimated. However, a new derivation of the hyperoxic CBV equation is presented that reveals significant errors in the original method, corresponding to the observed overestimates in CBV. Modelling studies are presented that investigate the discrepancy in reported BOLD MRI measurement of mean vessel size. A significant degree of the variation in the results is found to arise from the noise sensitivity of the analysis methods. This finding is confirmed with experimental data from healthy volunteers that show good agreement with the modelling studies.

Comprehensive modelling of the BOLD response to hyperoxia and hypercapnia is used to develop a new framework for OEF calculation. The new method is based on the calibration of the BOLD signal response against a change in intravascular susceptibility. The OEF calculation is extended by introducing a spin-echo readout into the acquisition scheme. This extension of the acquisition scheme provides a further independent probe of the BOLD signal, enabling the simultaneous calculation of the mean vessel size and CBV. The new framework is shown to provide OEF and vessel size estimates over a wider range of physiological parameters, providing greater scope for the clinical implementation of these techniques.

Authors: Germuska, M

• Publication date: 2013
• DOI: 10.5287/ora-44vrnbexk
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: MRI; oxygen metabolism
• Subjects: Clinical Neuropsychology; Metabolism; Neuropsychology; Biomedical engineering