An investigation into the human vestibular system with magnetic resonance imaging

Raiser, Theresa (2020): An investigation into the human vestibular system with magnetic resonance imaging. Dissertation, LMU München: Graduate School of Systemic Neurosciences (GSN)

Abstract

This thesis for receiving the doctorate PhD degree presents three projects about the vestibular system and one concerning eye-tracking analyses. Tracking the eye-movements elicited by the vestibulo-ocular reflex provide insights into vestibular function. Eye-tracking is therefore a major method used for investigating the vestibular system. The first two projects present studies concerning the structure of the vestibular system measured with magnet resonance imaging (MRI). The first focused on the peripheral vestibular system, lying within the temporal bone. An atlas of the inner ear (IE-Map) was established using multiple non-invasive MRI sequences. This validated atlas comprises the three semicircular canals, their ampullae, the otolith organs (saccule and utricle), and the cochlea (with scala tympani, scala vestibuli, cochlear cupula and cochlear duct), as well as their inner and outer dimensions. The IE-Map has the highest resolution and most accurate measurements of substructures of any inner ear atlas to date retrieved with non-invasive MRI. The IE-Map can be used as a widely applicable tool for future studies in neurology, neurosurgery and otorhinolaryngology. The second structural project investigated the corticocortical connections of a priori localized vestibular regions. Structural and functional corticocortical vestibular connectomes (CVC) were derived from state-of-the-art multi-modal neuroimaging data. We derived their modules and common network measures and compared the structural CVC to findings previously reported in non-human primates (using gold standard tracer injections). Our results show that the modularity of the structural corticocortical vestibular connectome of humans is extremely robust. Comparisons with non-human primate data revealed substantial differences in the organization across the two species. The structural CVC was characterized by a strong connectivity within each hemisphere, whereas the functional connectome emphasized a substantial synchronicity for homotopic nodes. Overall a right laterality preference in vestibular processing could be observed from both functional and structural data. The third project of this thesis investigated whether the resting state functional MRI BOLD signal or the functional connectivity is altered by the magnetic field of the MRI. Strong magnetic fields like the one in the MRI scanner elicit a nystagmus (eye-movements with a slow drift and a quick corrective movement) in complete darkness due to the Lorentz force that acts upon the peripheral vestibular system. This phenomenon is known as the magnetic vestibular stimulation (MVS). It was not clear, whether the MVS effect occurs during functional MRI (fMRI) scanning and whether it affects the BOLD signal or the brain connectivity. In 88% of our participants we could detect the MVS induced nystagmus during resting state fMRI (rsfMRI) scanning. Not only horizontal semicircular canal orientation with respect to the magnetic field, but also vestibular sensitivity (as measured by means of caloric irrigation) seems to influence the strength of the MVS effect across participants. Comparing rsfMRI with and without visual fixation, functional connectivity only differed in visual occipital areas and cerebellar regions. The BOLD signal fluctuations were not related to the slow phase velocity over time. However, participants with greater horizontal slow phase velocity on average showed higher activation in vestibular, executive control and attention involved cortical areas. The fourth project deals with improving eye-tracking analyses. As already mentioned, this step is very often of great importance when investigating the vestibular system. By making use of deep learning algorithms, a novel tool “DeepVOG” was developed. Eye-tracking in complete darkness for the MVS project resulted in eye-tracking videos with heterogeneous light conditions and a black rim in the field of view. Other already established eye-tracking analyses software failed estimating horizontal and vertical eye-position with a high confidence and low noise confound. DeepVOG was highly generalizable to other data sets. It results in pupil center localization, elliptical contour estimation and blink detection. For the MVS project we used the horizontal and vertical position of the pupil center as well as their confidence and blink detection. With this thesis and its heterogeneous four projects, important aspects of the peripheral and central vestibular system in humans were discovered in-vivo. It also provides methodological advances that are important for future vestibular and oculomotor research.

• Dokumententyp: Dissertationen (Dissertation, LMU München)
• Themengebiete: 500 Naturwissenschaften und Mathematik ; 500 Naturwissenschaften und Mathematik > 570 Biowissenschaften, Biologie ; 600 Technik, Medizin, angewandte Wissenschaften ; 600 Technik, Medizin, angewandte Wissenschaften > 610 Medizin und Gesundheit
• Fakultäten: Graduate School of Systemic Neurosciences (GSN)
• Sprache der Hochschulschrift: Englisch
• Datum der mündlichen Prüfung: 20. Mai 2020
• 1. Bericht­erstatter:in: Flanagin, Virginia