The Measure and Meaning of Structural Similarity in the Brain

The concept of similarity has recently emerged as a central and often-overlooked organizing principle of brain network structure. At the same time, recent advances in structural MRI analytics have enabled large-scale mapping of anatomical similarity within individual brains. However, key questions remain that limit the current applicability and interpretability of research using similarity mapping approaches. What is the theoretical basis for using similarity to study brain network architecture? How should similarity be measured? What are its genomic influences, neurobiological implications, and clinical consequences? This thesis tackles these fundamental, unresolved questions about the measurement and meaning of structural similarity in the brain. In Chapter 1, I review the historical, technical, and theoretical foundations of structural MRI similarity. I begin by introducing the historical roots and technical methodology of MRI similarity analysis and comparing it with the distinct MRI-based approaches of structural covariance and tractography analysis. In particular, I focus on evaluating the evidence supporting two key assumptions that often underlie the interpretation of structural MRI similarity: (i) that MRI similarity represents architectonic similarity between cortical areas; and (ii) that similar areas are more likely to be axonally connected, as predicted by the homophily principle. Next, I contextualise this empirical work with two generative models of homophilic networks: an economic model of cost-constrained connectional homophily, and a heterochronic model of ontogenetically phased cortical maturation. I then review studies of the genetic and transcriptional architecture of MRI similarity in population-averaged and disorder-specific contexts, and developmental studies of normative cohorts and clinical studies of neurodevelopmental and neurodegenerative disorders. Finally, I prioritise knowledge gaps that must be addressed to consolidate structural MRI similarity as an accessible, valid marker of the architecture and connectivity of an individual brain network. Building on this foundational review, in Chapter 2 I propose Morphometric INverse Divergence (MIND), a novel method to estimate within-subject similarity between cortical areas based on the divergence between their multivariate distributions of multiple MRI features. Compared to the prior approach of morphometric similarity networks (MSNs) on N>11,000 scans spanning three human datasets and one macaque dataset, MIND networks were more reliable, more consistent with cortical cytoarchitectonics and symmetry, and more correlated with tract-tracing measures of axonal connectivity. MIND networks derived from human T1-weighted MRI were more sensitive to age-related changes than MSNs or networks derived by tractography of diffusion-weighted MRI, suggesting that MIND can successfully identify relevant inter-individual variation. Gene co-expression between cortical areas was more strongly coupled to MIND networks than to MSNs or tractography, offering strong new evidence for the close link between cortical transcription and MRI-estimated brain structural architecture. MIND network phenotypes were also more heritable and thereby represent promising candidates for genetic analysis of brain network organization. I argue that MIND network analysis provides a biologically-validated lens for cortical connectomics using readily-available MRI data. In Chapter 3, I study the common genetic variation of MIND structural similarity in order to shed fresh light on the organization and evolution of the cortex, the causal relationships between brain structure and function, and the pathogenesis of heritable disorders. In a large normative adult cohort (N>30,000), I performed a genome-wide association study (GWAS) for each of the 276 edges in MIND similarity networks constructed using multivariate distributions of 4 complementary MRI features at each of 23 cortical areas. These edge-level genetic effects were highly replicated by parallel GWAS of an independent validation dataset (N>18,000 adults). I observed that the strong genetic correlations between multiple edges were largely reducible to two gradients of genetically-determined cortical similarity, each of which was aligned with geodesic distance from one of the two phylogenetically primitive areas (paleocortex and archicortex) predicted by the Dual Origin theory of cortical evolution. Genetic MIND gradients were more heritable than comparable gradients derived from GWAS of functional MRI connectivity networks, and the paleocortical trend was genetically correlated with, and causally predictive of, functional connectivity. Finally, I identified multiple global and local genetic correlations between both MIND genetic gradients and 9 clinical disorders or biological traits, indicating that the normative genetic architecture of human brain networks is pleiotropically associated with inherited risk of neuropsychiatric disorders. These results provide new insight into the dual origin of cortex and its implications for human brain function and disorders. Finally, in Chapter 4, I summarize the key theoretical, technical and neurobiological takeaways of this body of work. Based on these insights, I discuss next steps for establishing structural similarity as a framework for studying the fundamental organizing principles of the brain and their translational relevance in neuropsychiatry and beyond.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2024

Author dc:creator

• Sebenius, Isaac

Advisors dc:contributor.advisor

• Morgan, Sarah
• Bullmore, Edward

• MRI
• imaging-genetics
• networks
• MIND
• dual origin theory
• structural similarity

Licence

• https://www.repository.cam.ac.uk/bitstreams/e5130721-576d-4531-8e0f-71c43ac1a230/download
• https://creativecommons.org/licenses/by/4.0/

Language dc:language

eng

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.117257

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/382458