Modelling and quantifying brain connectivity and dynamics with applications in aging and ADHD

Abstract

Human brain is a complex organ and made up of integrative networks encompassing a large number of regions. These regions communicate with each other to share information involved in complex cognitive processes. Functional connectivity (FC) represents the level of synchronization between different brain regions/networks. Studying functional interactions of the brain creates a platform for understanding functional architecture of the brain as an integrative network and has implications for understanding human cognition. Furthermore, there is evidence that FC patterns are sensitive to different diseases. In addition, age is a significant determinant of intra-/inter-individual variability in the FC patterns. Therefore, key aims for the studies included in this thesis were to apply and develop novel resting-state FC methodologies, with applications in healthy aging and ADHD. Indeed, measures of the brain’s FC may serve as a useful tool to diagnose and predict the course of disease, and useful in developing individualized therapies.

Age- or disease-related alterations in the FC could reflect a multitude of factors, including changes in structural connectivity. However, we still have limited knowledge of the emergence of brain dynamics from the underlying anatomy. The interplay between the brain’s structure and dynamics underlies all brain functions. Therefore, in the last study we focused on the systematic modeling of the brain network dynamics. Large-scale computational models are uniquely suited to address difficult questions related to the role of brain’s structural network in shaping functional interactions. In addition, computational modeling of the brain enables us to test different hypotheses without any experimental complication while it provides us with a platform for improving our understanding of different brain mechanisms. A new macroscopic computational model of the brain oscillations for resting-state fMRI was introduced in this thesis, which outperforms previous model in the same class. Then, the effects of malfunctions in different brain regions were simulated and subsequently predicted perturbation patterns were recruited for local vulnerability mapping as well as quantification of hazard rates induced after perturbing any brain region.