Recent developments in functional magnetic resonance imaging (fMRI) methods at high magnetic field strengths (≥7 T) promise the non-invasive indirect measurement of neural activation at the spatial scale of cortical columns and layers. However, the achievable spatial specificity of fMRI, which is ultimately limited by the spatio-temporal properties of the hemodynamic response, is still waiting to be fully characterized.
To examine the spatial specificity of the blood oxygenation level dependent (BOLD) contrast exploited by fMRI, the spatial point spread function (PSF) of the gradient echo (GE) and the spin echo (SE) BOLD signal tangential to the cortical surface was determined at different cortical depths. Both GE- and SE-BOLD showed a loss in spatial specificity toward the pial surface, demonstrating the impact of unspecific macrovascular contributions to both contrasts and only a minor advantage of the SE-BOLD signal for high-resolution fMRI applications.
However, unidirectional draining of deoxygenated blood mainly limits spatial specificity in the radial direction. To examine the discriminability of laminar information, ocular dominance columns (ODCs) in the primary visual cortex (V1) were mapped using fMRI sensitive to either the BOLD contrast or cerebral blood volume (CBV) changes, and the stimulated eye was decoded using a machine learning classifier at different cortical depths. Only CBV-fMRI showed increased prediction accuracies at the cortical depth that matched neurophysiological expectations, showing its improved spatial specificity and potential for layer-specific fMRI in humans.
Furthermore, the thin-thick-pale stripe pattern in the secondary visual cortex (V2) was targeted, exploiting the sensitivity to color and binocular disparity of thin and thick stripes, respectively. The structure-function relationship of the stripe architecture to cortical myelin was studied, which so far has shown inconsistent findings in multiple histological experiments. High-resolution quantitative MRI (qMRI) parameter maps of the longitudinal relaxation rate (R1) were used as a proxy for cortical myelin content. The comparison of fMRI and qMRI maps showed that both thin and thick stripes have lower R1 than surrounding cortical tissue, pointing toward higher myelin content of pale stripes.
While macrovascular contributions in fMRI must be considered cautiously, the thesis demonstrates the capabilities to study structure-function relationships and retrieval of laminar information at the spatial scale of cortical columns with high-resolution fMRI at 7 T.:List of figures
List of tables
List of acronyms
1 Introduction
1.1 Imaging the human brain
1.2 The visual cortex
1.3 Vascular supply of the cerebral cortex
1.4 Thesis outline
2 Background
2.1 Nuclear magnetic resonance
2.1.1 Nuclear magnetic moment
2.1.2 Zeeman effect
2.1.3 Bulk magnetization
2.1.4 Excitation
2.1.5 Relaxation
2.1.6 Refocusing
2.1.7 Detection
2.2 Magnetic resonance imaging
2.2.1 Gradients
2.2.2 Spatial encoding
2.2.3 Echo-planar imaging
2.3 Functional magnetic resonance imaging
2.3.1 Blood
2.3.2 Hemodynamic response
2.3.3 BOLD-fMRI
2.3.4 CBV-fMRI
2.4 Spatial specificity
2.4.1 Point spread function
2.4.2 Imaging PSF
2.4.3 Physiological PSF
3 Cortical depth-dependent spatial specificity of GE- and SE-BOLD
3.1 Introduction
3.2 Theory
3.3 Materials and methods
3.3.1 Participants
3.3.2 General procedure
3.3.3 Visual stimulation
3.3.4 Imaging
3.3.5 Data preprocessing
3.3.6 MTF model fitting using MCMC
3.4 Results
3.4.1 GE- and SE-BOLD maps
3.4.2 Percent signal changes across cortical depth
3.4.3 Cortical distances along iso-eccentricity lines
3.4.4 MCMC diagnostics
3.4.5 Estimated MTF parameters
3.4.6 MTF within veins
3.5 Discussion
4 Laminar profile of human ocular dominance columns
4.1 Introduction
4.2 Materials and methods
4.2.1 Participants
4.2.2 General procedure
4.2.3 Visual stimulation
4.2.4 Imaging
4.2.5 Data preprocessing
4.2.6 Pattern classification
4.3 Results
4.3.1 Topography of ocular dominance columns
4.3.2 Reproducibility of ocular dominance maps
4.3.3 Univariate contrasts across cortical depth
4.3.4 Decoding accuracies across cortical depth
4.4 Discussion
5 Cortical myelination of the secondary visual cortex (V2)
5.1 Introduction
5.2 Materials and methods
5.2.1 Participants
5.2.2 General procedure
5.2.3 Visual stimulation
5.2.4 Imaging
5.2.5 Data analysis
5.3 Results
5.3.1 Functional mapping of color-selective and disparity-selective stripes
5.3.2 Consistent qMRI maps across cortical regions and cortical depth
5.3.3 Higher myelination of pale stripes
5.4 Discussion
6 General discussion
A Gradient-based boundary registration
B Construction of anaglyph spectacles
C Supplementary data for chapter 3
D Supplementary data for chapter 4
E Supplementary data for chapter 5
F Analysis of registration accuracy
Acknowledgements
Bibliography
Curriculum vitae
Declaration of authorship
| Identifer | oai:union.ndltd.org:DRESDEN/oai:qucosa:de:qucosa:91551 |
| Date | 22 May 2024 |
| Creators | Haenelt, Daniel |
| Contributors | Universität Leipzig |
| Source Sets | Hochschulschriftenserver (HSSS) der SLUB Dresden |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/publishedVersion, doc-type:doctoralThesis, info:eu-repo/semantics/doctoralThesis, doc-type:Text |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0021 seconds