Doctor of Philosophy
School of Psychology
Large draining veins residing on the cortical surface are recognised as a major problem in fMRI measurements, leading to the displacement, distortion, and reduction in spatial localisation of signals. This is detrimental to the application of fMRI in advanced human brain imaging, such as in retinotopic mapping, where the BOLD signal becomes obscured by the artefacts from nearby draining veins (venous artefacts). There is consensus that at an increased magnetic field strength, the sensitivity to the contributions from these draining veins reduces, which promises the ability to capture measurements at greater sensitivity and specificity. With the advent of fMRI hardware and technologies, there is a great interest in understanding cortical layer-specific neuronal response that had been previously obscured due to the domination of signal change caused by venous artefacts. This forms the basis of support for fMRI at ultra-high magnetic field strength, with the conventional wisdom that the reduced sensitivity to macrovasculature (including draining veins) that comes with higher magnetic field strength can increase the sensitivity to underlying responses that closely reflect neuronal activation. However, this is debatable. Instead, there are instances where human neuronal response profiles displayed a pattern that seemed indicative of synaptic activities even without using ultra-high magnetic field strength. Recent trends in moving to ultra-high magnetic field strengths, especially for depth-dependent fMRI, beget the questions – is ultra-high magnetic field strength really necessary, and can it resolve the artefacts from draining veins? To elucidate these questions, this thesis aims to test the venous artefact and its impact on underlying signals across spatial resolutions and magnetic field strengths, focusing on the quality of the retinotopic organisation of the early visual cortex.
The first experimental study was conducted to test the venous artefact and its impact on fMRI signals across the grey matter using high-resolution (isotropic resolution of 1 mm) fMRI images collected at 3 T. Using two surface reconstruction packages, the findings established that venous artefact occurs at the cortical surface and spreads within the grey matter. In this study, the ability of high-resolution fMRI at 3 T to conduct depth-dependent analyses was demonstrated. The second experimental study delves into the role of spatial resolution in the depth-dependent analysis of venous artefacts and their impact. The 1 mm 3 T fMRI images were spatially smoothed to simulate two additional sets of lower-effective resolution images. Here, the results found consistency in the venous artefact but a reduction in the venous effects. This suggests that at lower spatial resolutions, the venous artefact exists, but its impact on underlying signals is concealed. The third experiment incorporates two sets of 7 T fMRI images, with spatial resolutions of 1.6 mm and 0.8 mm, to test the venous artefacts at ultra-high magnetic field strength. The results showed evidence that venous artefacts remained prominent at 7 T. This study was extended into a comparative study by including the 3 T fMRI images used in the first study. Here, the generalisability of the venous artefact was established at 3 T and 7 T, as well as across various spatial resolutions. Finally, the fourth experiment explored a non- BOLD contrast, postulated to be less sensitive to contributions from the draining veins. In this study, VASO-fMRI was conducted with a spatial resolution of 1.1 mm and at 7 T to test the venous artefact and its impact. A reduction in the venous artefact and its impact was demonstrated. However, VASO images were found insufficient to inform the retinotopic organisation of the early visual cortex due to higher noise and lower signals, but they can be beneficial when used in conjunction with BOLD images.
In summary, this thesis established the competence of high-resolution fMRI at 3 T for depth-dependent analysis. This suggests that the artefacts arising from draining veins can be avoided by avoiding surfaces or locations that are in proximity to these veins. Furthermore, this thesis sheds light on the generalisability of the venous artefact across magnetic field strengths and spatial resolutions. Finally, the results from the VASO-fMRI study demonstrated how VASO retinotopic mapping could benefit BOLD retinotopic mapping, especially when the BOLD response is highly polluted with macrovasculature.
Koo, Wei Shi Eleanor, Vascular artefacts in fMRI of early visual cortex: The effects of resolution and field strength, Doctor of Philosophy thesis, School of Psychology, University of Wollongong, 2023. https://ro.uow.edu.au/theses1/1605
FoR codes (2008)
1109 NEUROSCIENCES, 110999 Neurosciences not elsewhere classified, 170101 Biological Psychology (Neuropsychology, Psychopharmacology, Physiological Psychology)
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.