Neuromyelitis optica spectrum disorder (NMOSD) is a relapsing demyelinating neurological condition that requires prompt diagnosis and treatment to prevent the accrual of significant disability, such as blindness and immobility. It is caused by antibodies to the aquaporin-4 water channel, which can be detected within the serum of patients and are therefore a highly specific biomarker for the condition. However NMOSD can have overlapping clinical features with the more prevalent demyelinating disease relapsing remitting multiple sclerosis (RRMS). Thus in the situation of a negative aquaporin-4 antibody assay result diagnostic distinction can be difficult to achieve. This in turn leads to a treatment dilemma as NMOSD requires long term immunosuppression, and disease modifying therapies used in RRMS can worsen disease activity in NMOSD. The work presented in this thesis aimed to find further biomarkers of NMOSD, focusing on conventional and quantitative magnetic resonance imaging, and markers of neurodegeneration. The principle findings are: <ul><li>Lesion probability mapping analysis revealed that T2 lesion distribution and morphology can distinguish RRMS from NMO with a 92% sensitivity and 96% specificity. Radiological criteria to aid diagnosis have been proposed.</li><li>Quantitative imaging methods that provide complementary micro-structural measures within CNS tissue have shown that RRMS is associated with widespread neurodegenerative changes in the normal appearing tissue, where as NMO is lesion focused.</li><li>NMO is not associated with diffuse global neurodegeneration in the non-lesioned CNS tissue over a one-year period as determined by serial MRI.</li><li>Normalised thalamic and whole brain volume can be used in a function to help discriminate NMOSD and RRMS</li></ul> This thesis therefore contributes the knowledge that conventional and quantitative imaging can distinguish NMOSD and RRMS, and provides practical methods in which to apply this to patient management. It is the first longitudinal quantitative imaging study of NMOSD and lends further proof that global or diffuse neurodegeneration of the normal appearing brain or spinal cord tissue is not a feature of this condition.

Multiple sclerosis is a progressive autoimmune disease where myelin is gradually stripped from axons. Axon degeneration inevitably follows protracted myelin loss ultimately leading to irreversible neurological decline. To better understand the cellular mechanisms associated with the axon loss phase of the disease, spinal cord axons from the experimental autoimmune encephalomyelitis (EAE) animal model of multiple sclerosis were examined using correlated in vivo time-lapse microscopy and serial section transmission electron microscopic (ssTEM) reconstruction. A novel technique, termed near infrared burning (NIRB), was developed that took advantage of a femtosecond-pulsed mode locked laser’s ability to create photoconvertable fiducial markers for routine identification of previously imaged axons for ssTEM reconstruction. This combination of imaging techniques revealed the subcellular milieu that underlies axon degeneration at both the light and electron microscopic level. In particular, paranodal regions of axons in EAE animals contained a significantly higher population of mitochondria with large rounded, electron lucid, vesiculated mitochondria with unorganized cristae compared to controls. This effect was largely restricted to the paranodal region and was not always associated with direct immune cell interaction or myelin loss. Together, these results suggest a novel mechanism for axon degeneration that is not only focal in nature, but decoupled with myelin loss in the EAE animal model of multiple sclerosis. / Department of Biology