Abstract:
I will selectively review some recent studies that aim to address the challenges of mechanistic understanding and prediction of human brain connectivity measured using MRI. Mechanistic studies are broadly defined by their intention to identify underlying neurobiological mechanisms (cells, synapses, gene transcripts etc) that can account for the relatively coarse-grained measurements and statistical metrics of large-scale connectivity that are estimable from human MRI. They are often analogical or correlational by design, e.g., testing for spatial co-location of human MRI-derived connectivity maps with brain maps of gene expression. Predictive studies are defined in contrast by their more reductionist or causal ambition, e.g., seeking to define genetic and environmental factors that determine the trajectories of brain development, or using computational models to simulate connectome formation and attendant cognitive functions. Both strategies are important and have been progressive, especially in the context of the recently growing scale and open-ness of relevant MRI datasets and reference atlases.
