• Login
• Admin

Howley, Elliot (2025) Quantifying Data Quality and Its Impact on Functional Brain Imaging Experiments. PhD thesis, University of Nottingham.

Preview

PDF (Corrections) (Thesis - as examined) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader Available under Licence Creative Commons Attribution. Download (34MB) | Preview

Abstract

Functional Magnetic Resonance Imaging (fMRI) is a widely-used tool in neuroscience research. While there is general agreement on what imaging sequences and methods work best overall, there is much less agreement and consistency on how particular parameter choices are made. These parameter choices can have an effect on data quality, which can negatively affect analysis of this data. It is therefore important to characterise this effect. The thesis investigates the impact of image acceleration techniques on fMRI data quality, quantified using temporal Signal-to-Noise Ratio (tSNR), and explores how these effects vary across different brain regions. The thesis then investigates the impact of higher levels of Gaussian noise and head motion on an important and widely adopted analysis method: population Receptive Field (pRF) analysis. Assessment of the effects of applying image denoising to fMRI data are also investigated throughout.

Chapter 3 presents development and use of the fMRI ROI Analysis Tool (fRAT), software designed to provide a comprehensive Region-of-Interest (ROI) analysis toolset for fMRI data. fRAT addresses the lack of existing fMRI tools making it easy to analyse multiple ROIs with data quality metrics. This tool enables researchers to easily study spatial variations in the relationship between scanning parameters and data quality. The software's features, including statistical analysis and data visualisation capabilities are detailed, and current and potential future applications are highlighted.

Chapter 4 uses fRAT to characterise the effect of hardware (3T Philips Achieva and 3T Philips Ingenia), image acceleration (in-plane SENSE factor and through-plane Multiband factor) and a post-hoc denoising technique (using NOise reduction with DIstribution Corrected [NORDIC] PCA) on data quality across a selection of regions of interest: the Frontal Pole, the posterior Inferior Temporal Gyrus and the Occipital Pole. The relationship between these variables was found to vary between these regions, supporting the idea that region-wise data quality (tSNR) reporting provides important information.

Chapter 5 evaluates the robustness of pRF analysis in the visual domain to decreased levels of tSNR and increased levels of participant motion through adding simulated thermal noise and head motion to a pre-existing pRF dataset collected in stroke patients [@behLinkingMultiModalMRI2021]. Work in this chapter also makes use of fRAT to first quantify noise levels and then provide a convenient way to manipulate the data before pRF analysis. It is shown that in general, pRF analysis is more robust to the addition of head motion than to noise, with the polar angle of the pRF estimates being the property most consistently affected by these factors.

Overall, this thesis provides a detailed analysis of the spatially dependent effects of image acceleration on fMRI data quality and underscores the practical consequences of changes in the level of data quality and motion in pRF analysis. The findings aim to inform best practices when conducting fMRI research, and importantly, the software developed within this thesis has been made open-source with usage tutorials to enable it to be used across a wide range of applications in future research.

Item Type:	Thesis (University of Nottingham only) (PhD)
Supervisors:	Schluppeck, Denis Francis, Sue
Keywords:	Functional Magnetic Resonance Imaging, fMRI, tSNR, pRF, Population Receptive Field, brain imaging
Subjects:	Q Science > QP Physiology > QP351 Neurophysiology and neuropsychology R Medicine > RC Internal medicine > RC 321 Neuroscience. Biological psychiatry. Neuropsychiatry
Faculties/Schools:	UK Campuses > Faculty of Science > School of Psychology
Item ID:	80725
Depositing User:	Howley, Elliot
Date Deposited:	30 Jul 2025 04:40
Last Modified:	30 Jul 2025 04:40
URI:	https://eprints.nottingham.ac.uk/id/eprint/80725

Actions (Archive Staff Only)

Edit View