FORMATION, REFINEMENT, AND REORGANISATION OF COMPLEX BRAIN NETWORK TOPOLOGY ACROSS THE LIFESPAN

Abstract

Large-scale connections in the brain are organised in complex patterns. The unique distribution of connections across the whole brain is related to important cognitive, behavioural and mental health outcomes. In this thesis, I examine how the organisation of these connections form and changes across the lifespan. In Chapter 1, I provide theoretical background about the study of complex systems and the common properties observed across them – small-worldness, scale-free degree distributions and modularity. Diving deeper in brain networks specifically, I briefly introduce the study of neural organisation across micro-, meso- and macro-scales. Expanding on macroscale human network organisation, I highlight the relationship between brain organisation and cognition, behaviour, and mental health, and typical variations within these relationships. I then introduce the practical steps we take to study macroscale brain organisation, discuss methodological limitations, and review the current field of lifespan structural topological development. In Chapter 2, I explore the early formation of topology in a diverse sample of preterm and term-born infants. In this project, I explore how network density affects topological measures, highlighting how connectivity differences influence interpretations of fundamental topological architecture. I show that while infants born early have fewer connections than those born later, some but not all topological differences are explained by total connectivity differences. Importantly, compared to those born at term, infants born early have differently organised networks characterised by the relative conservation of vital connections in sparser networks. These differences suggest a renegotiation of optimal wiring related to the timing of birth. To extend beyond observation, in Chapter 3 I model these neonatal networks using a biophysically constrained computational approach. This method selects new connections in a network based on a basic economic trade-off that calculates how costly or valuable a connection is to the network. Despite the model’s simplicity, I demonstrate this method is accurate in replicating real-world neonatal network topology. Crucially, to replicate preterm topology, these models must alter the economic trade-off, highlighting the mechanisms that drive differences in topology in infants born early versus at term. Renegotiation of wiring in infants born preterm results in a complex topology that preserves vital, long-range connections while forming overall shorter connections in the network. This modelling project highlights the benefit of extending basic observation by simulating network generation, allowing for a better understanding of what potential principles drive topological formation. The brain's topology reorganises across the human lifespan, well after the neonatal years. In Chapter 4, I model the lifespan development of topology. Here, I demonstrate the developmental variability of different organisational principles. Some topology metrics show small changes across the lifespan, others change linearly, and some show multiple fluctuations. While topological change is highly measure-specific, two general trends emerge across the lifespan. First, there is a peak of network efficiency and integration at the beginning of the fourth decade of life. Ageing is then related to increasing network sparsity and modularity. While these two trends are prominent, there are many other non-linear intricacies in lifespan topological change. In the final empirical chapter, Chapter 5, I introduce the application of data-driven machine learning methods to better understand high-dimensional patterns. Here, I take a multi-methods approach to characterise general topology change and identify important topological turning points in the lifespan. By compressing high-dimensional topological data into lower dimensions, I show that topological development has distinct phases from infancy to childhood, adolescence, adulthood, early ageing and late ageing. Importantly, topological turning points align with important cognitive, behavioural and health milestones. This work further reinforces the importance of taking lifespan approaches to explore the relationships between the development of the brain and observable outcomes. In Chapter 6, I summarise key takeaways from this thesis and integrate them into the larger field of research. I point out important areas for future research and promising prospects for future developments in the investigation of the brain as a complex system.