Time resolved functional brain networks : a novel method and developmental perspective

Abstract

Functional neuroimaging has helped elucidating the complexity of brain function in ever more detail during the last 30 years. In this time the concepts used to understand how the brain works has also developed from a focus on regional activation to a network based whole brain perspective (Deco et al., 2015). The understanding that the brain is not just merely responding to external demands but is itself a co-creator of its perceived reality is now the default perspective (Buzsáki and Fernández-Ruiz, 2019). This means that the brain is never resting and its intrinsic architecture is the basis for any task related modulation (Cole et al., 2014). As often in science, understanding and technological advances go hand in hand. For the advancement of the functional neuroimaging field during the last decade, methods that are able to track, capture and model time resolved connectivity changes has been essential (Lurie et al., 2020). This development is an ongoing process. Part of the work presented in this thesis is a small contribution to this collective endeavor.

The first theme in the thesis is time resolved connectivity of functional brain networks. This theme is present in Study I which presents a novel method for analysis of time resolved connectivity using BOLD fMRI data. With this method, subnetworks in the brain are defined dynamically. It allows for connectivity changes to be tracked from time point to time point while respecting the temporal ordering of the data. It also provides relational properties in terms of differences in phase coherence between simultaneously integrated networks and their gradual change. The method can be used see how whole brain connectivity configurations recure in quasi-cyclic patterns. Finally, the method is able to estimate flexibility and modularity of individual brain areas. The method is applied in Study III in order to understand how premature birth effects flexibility and modularity of intrinsic functional brain networks.

Beyond the purely scientific endeavor to understand how the brain creates cognition, consciousness, perception and supports motor function, neuroimaging research has also been helpful in elucidating normal brain development and neurodevelopmental disorders. The second theme in this thesis is brain development in extremely preterm born children at school age. This theme is the focus of Study II & III. Study II investigates the prevalence of discrete white matter abnormalities at school age in children born extremely preterm and the relationship to neuro-motor outcome. The prevalence of white matter abnormalities was high but there was no relationship to an unfavorable outcome. Also, a longitudinal association to neonatal white matter injury was seen. While discrete white matter abnormalities were not correlated to quantitative measures of white matter volume and white matter integrity, neonatal white matter injury was associated with lower volume and integrity at age 8- 11 years. Moreover, neonatal white matter injury was associated with lower processing speed at 12 years.

The third and final study investigated flexibility and modularity as well as lateralization of intrinsic networks in children born extremely preterm at age 8-11 years. No significant differences in either flexibility or modularity was seen for any intrinsic network after correcting for multiple comparisons. However, at the level of individual brain areas, preterm children showed decreased flexibility in both the basal ganglia and thalamus. Also, children born extremely preterm had a decreased level of lateralization in most networks.