Computational Analysis of Multiscale Cortical Organization and Development

The cerebral cortex is organized at multiple scales, ranging from ion channels, to neuronal circuitsorganized across cortical layers, to the interconnected network of cortical areas. The structuraland functional properties at these scales vary widely across cortical areas. This heterogeneousorganization across different scales is the result of continuous refinement throughout the lifespan.Understanding the multiscale organization of the cerebral cortex and its maturation requiresintegrative computational approaches that bridge across scales. The goal of this work was to useadvanced computational techniques to better understand how micro- and mesoscale corticalphenomena relate to macroscale cortical organization throughout development. Specifically,we examined cortical cytoarchitecture associated with corticocortical connectivity (Study1), cortical microcircuitry inferred from functional dynamics and connectivity (Study 2),and cellular and molecular processes underlying cortical morphology (Study 3).In Study 1, we found that cortical laminar structure at the mesoscale varied along a principalaxis extending from caudal to rostral areas, along which the relative thickness of deeper layersincreased. This axis was co-aligned with the hierarchical organization of macroscale corticalconnectivity. Furthermore, similarity of laminar structure was associated with the likelihoodand strength of corticocortical connectivity, a phenomenon thought to have developmental roots.Next, in Study 2, we used an individualized computational modeling approach to infer theregional levels of excitation-inhibition balance in cortical microcircuits of developing adolescentsbased on their macroscale cortical connectivity and dynamics observed in resting-state functionalimaging. To enable the large-scale simulations required for this approach, we developed a noveland efficient implementation of the simulations, released as a Python package, cuBNM. Usingthis approach, across two independent cross-sectional and longitudinal datasets, we found awidespread age-related decrease of excitation relative to inhibition within the association areas,paralleled by its increase or lack of change in sensorimotor areas. This developmental patternwas consistent with the previously proposed sensorimotor-association spatiotemporal pattern ofneurodevelopment.Finally, in study 3, we examined the spatial co-localization between cortical maps of microandmesoscale neurobiological processes with cross-sectional and longitudinal spatiotemporalpatterns of cortical thickness changes across the lifespan to understand which cellular andmolecular processes may underlie maturation and lifespan changes in cortical morphology at themacroscale. Our results suggest that processes such as dopaminergic, glutamatergic, and cholinergicneurotransmitter systems, as well as glial cells, inhibitory neurons, and brain metabolism,may contribute to the maturation of cortical morphology.Overall, this work advances our understanding of multiscale cortical organization and itsmaturation while contributing to the development of computational tools for future research.By integrating micro-, meso-, and macroscale perspectives, our findings on normative cortical organizationand maturation provide a foundation for investigating impaired cortical developmentin mental health disorders.ii