Layer-fMRI acquisition and analysis

Marsh, Daniel C. (2023) Layer-fMRI acquisition and analysis. PhD thesis, University of Nottingham.

PDF (Final corrected thesis) (Thesis - as examined) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Available under Licence Creative Commons Attribution.
Download (149MB) | Preview


Functional Magnetic Resonance Imaging (fMRI) is a widely adopted imaging modality used to study human brain function. The advancements in MR hardware and acquisition have enabled researchers to investigate brain function at sub-millimetre resolution, detecting brain activation at depths across the cortex. Layer-fMRI at ultra-high field (UHF, defined as B0 ≥ 7 T) has the potential to answer many neuroscience questions that were previously untenable due to low spatial resolution or a low signalto-noise ratio (SNR). In addition, the combination of layer-fMRI with electroencephalography (EEG) provides a tool to investigate neuronal oscillation across cortical depths.

This thesis develops methods for the analysis of layer-dependent simultaneous EEG-fMRI data acquired at 7 T using a gradient echo (GE) Blood Oxygenation Level Dependent (BOLD) sequence during an eyes open, eyes closed task to assess the origins of human alpha oscillations. An optimal pipeline is developed for the combination of EEG and fMRI data with structural MRI data in order to calculate layer-specific alpha activation profiles. These methods will be an important building block for the growth of layer-dependent EEG-fMRI as an imaging tool, with only one previous 3 T study to date known to have acquired such data [1]. A key limitation of the methods was found to be the draining vein effect present in GE-BOLD data, with multiple methods considered to correct for it.

To improve understanding of alpha oscillations, the bespoke analysis pipeline was applied to layer-dependent EEG-fMRI data acquired on 10 subjects at 7 T during an eyes open, eyes closed task. The results showed significant negative correlation between EEG alpha power and the BOLD response in the visual cortex. The cortical layer profiles of negative alpha-BOLD correlations exhibited a dip in the middle cortical depths and peaked in the deeper and superficial depths, suggesting that during an eyes open/closed paradigm alpha is predominantly generated during top-down processing through corticocortical mechanism.

This is then followed by a study using the non-BOLD fMRI contrast of vascular space occupancy (VASO) for layer-fMRI measures. High spatial resolution VASO has greater spatial specificity and does not suffer from the impacts of the draining vein effect making it a good option for layer-fMRI. However, VASO has inherently lower signal than GE-BOLD and is often SNR limited. In this thesis, a denoising method of NOise Reduction with DIstribution Correction (NORDIC) Principal Component Analysis (PCA) was assessed for its application to high resolution 3DEPI BOLD data before being applied to an optimised 1 mm isotropic VASO sequence for layer-specific measures during a finger-tapping task. The results from ten subjects show a VASO layer profile that peaks in the middle cortical depths. A double peak was expected from the literature however this disparity was most likely due to acquiring at a spatial resolution that was too coarse. When compared, the VASO and ‘deveined’ GE-BOLD layer profiles showed a similar shape for cortical depths 1 – 4 which then diverged in depths 5 and 6, highlighting the need for further work to validate corrections of the draining vein effect.

Finally, NORDIC denoising was applied to T1 mapping data for structural layer measures. It was found that there was a small improvement in the T1 fit, represented by an increase in the wellness of the fit value, R2. Additionally, there was a tightening in the peaks for both the grey matter (GM) and white matter (WM) T1 values, shown by decreases in the full width at half maximum (FWHM). WM showed greater improvements than GM for both the R2 and FWHM.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Francis, Susan
Mullinger, Karen
Keywords: Functional Magnetic Resonance Imaging, fMRI, brain, brain function
Subjects: Q Science > QC Physics > QC501 Electricity and magnetism
Q Science > QP Physiology > QP351 Neurophysiology and neuropsychology
Faculties/Schools: UK Campuses > Faculty of Science > School of Physics and Astronomy
Item ID: 76708
Depositing User: Marsh, Daniel
Date Deposited: 10 Jan 2024 14:44
Last Modified: 10 Jan 2024 14:44

Actions (Archive Staff Only)

Edit View Edit View