On October 4th MD Kirsten Korsholm succesfully defended her PhD thesis "Functional and Structural MRI in Optic Neuritis".

English summary

Optic neuritis (ON) is an optic neuropathy characterised by retrobulbar pain and transient visual impairment. ON is the presenting symptom in approximately 20% of patients with multiple sclerosis (MS), and after isolated acute ON the 10-year risk of developing MS has been reported to be approximately 40%.

A spontaneous recovery of visual function usually occurs within weeks or months after onset of ON symptoms despite of atrophy of the optic nerve as demonstrated by magnetic resonance imaging (MRI) and persistently prolonged latency as measured with visual evoked potentials (VEP). The mechanisms behind the recovery are not clear, however, if atrophy and abnormal VEPs are indicators of permanent local damage following ON, then recovery of vision must in principle occur either because of redundancy in the anterior visual pathways or because of subcortical or cortical remodelling of visual function.

As it is essential to understand these recovery mechanisms, not only because of what we can learn about neuroplasticity in the adult brain, but also to characterize properly the effects of treatment, a central aim of this study was to characterise, with functional magnetic resonance imaging (fMRI) and VEP, to what extent recovery from ON is based upon remyelination of the optic nerve or upon adaptive cortical and subcortical changes. As no studies have serially investigated what happens in the lateral geniculate nucleus (LGN) during recovery from ON, this region was of particular interest in the study.

In order to assess activity in small brain structures (e.g. in the LGN), fMRI at high field (3.0 tesla (T)) is preferred due to a better signal to noise ratio. Previously, diagnostic scans of patients with ON have only been carried out at 1.5 T requiring a methodological study comparing structural MRI findings at 1.5 T and 3.0 T to perform both the functional and the diagnostic scan of the study at 3.0 T. To accomplish the comparison, 28 patients with ON were scanned at 1.5 T and 3.0 T, and the number of lesions was assessed on T2-weighted images (fluid attenuated inversion recovery, FLAIR) and on T1-weighted images after injection of gadolinium. The study shows a statistically significant larger number of hyperintense lesions on FLAIR images detected at 3.0 T compared to 1.5 T (p=0.002), and with regard to gadolinium-enhancing lesions on T1-weighted images there is a trend towards more lesions at 3.0 T (p=0.068). In conclusion, the FLAIR sequence at 3.0 T is more sensitive to hyperintense lesions than FLAIR at 1.5 T. This difference in sensitivity could have an impact on the diagnosis of patients with ON as clinically isolated syndrome (CIS) and should therefore be considered in future large scale trials of patients with CIS.

Then, in order to clarify the role of the LGN and the relationship between the LGN and visual cortical areas during recovery from ON, visual activation was studied longitudinally in 19 patients during recovery from acute ON, and region-of-interest (ROI) analyses of the LGN and of primary and secondary visual areas (V1+V2) were performed. The lateral occipital complexes (LOC) were also examined, as other studies have found evidence of cortical adaptive changes in this area (Toosy et al. 2005; Levin et al. 2006).

In addition, a longitudinal voxelbased analysis of whole brain activation was performed in a group of patients with acute ON to study whole brain activation patterns during and after recovery from ON. This type of analysis can be problematic due to differences in scotoma location between patients making analysis of retinotopically organized areas difficult. However, a new method capable of modelling different scotoma locations is introduced and used in the present study.

The ROI-based analyses show that activation of the LGN during visual stimulation of the affected eye in the acute phase is significantly reduced (p<0.01) compared to the unaffected eye. This difference diminishes during recovery and after 180 days the difference between the affected and the unaffected eye is no longer significant. In LOC, V1 and V2 activation during visual stimulation of the affected eye in the acute phase is significantly reduced (p<0.01) compared to the unaffected eye, and during recovery the difference diminishes with no significant differences left after 180 days. As the pattern of activation in LOC, V1 and V2 resembles the development in the LGN, no evidence of cortical adaptive changes is found in these areas.

With the model developed for the voxelwise analysis of brain activation in patients with ON, it is demonstrated that the changes in visual activation during recovery are located to the visual cortex and the LGN. No changes in activation during or after recovery are found outside visual areas as reported by other studies (Werring et al. 2000; Toosy et al. 2002). During recovery from ON, VEP-latencies of the affected eye remained prolonged suggesting that remyelination had not taken place. VEP-amplitudes did increase significantly during recovery, pointing to resolution of inflammation and oedema.

In conclusion, no cortical adaptive changes were found, and it is most likely that recovery of vision in our population of ON patients is due to resolution of inflammation and oedema during the first months. In addition, redundancy in the visual system, e.g. in the number of fibres in the optic nerve required to maintain normal vision, may play an important role.