Mapping of the intrinsic and extrinsic cerebellar connectome at ultra high resolution with expert neuroanatomical curation

In this 5-year R01 project titled “Mapping of the intrinsic and extrinsic cerebellar connectome at ultra high resolution with expert neuroanatomical curation,” we propose to create the first detailed atlas of the human cerebellar connectome using sub-millimeter ultra-high-resolution diffusion MRI (dMRI). The human cerebellar connectome is affected in multiple conditions including autism, schizophrenia, Down's syndrome, Alzheimer's disease, Parkinson's disease, cerebellar mutism, and neurodegeneration. The complex anatomy of the cerebellar structural connectome includes intricate connections between the tightly foliated cerebellar cortex, the deep cerebellar nuclei, and structures external to the cerebellum including the spinal cord, brainstem, thalamus, and cerebral cortex. Remarkably, the cerebellum contains 80% of all neurons in the brain. Yet, a complete map of the intrinsic and extrinsic structural connectome of the human cerebellum is still lacking. Several challenges have prevented detailed mapping of the cerebellar structural connectome. First, the limited spatial resolution of current state-of-the-art dMRI data prevents mapping of intricate connections between the cerebellar cortex and small nuclei in the deep cerebellum, brainstem, and thalamus. Second, cerebellar connectome mapping is further limited by abundant anatomical errors in existing dMRI tractography algorithms, which do not respect known synapses and decussations or key nuclei. Third, our understanding of human neuroanatomy relies heavily on the results of invasive tracer studies in monkeys, but the detailed neuroanatomy of the cerebellar connectome in monkeys has not yet been systematically mined and compiled. We propose to address these challenges to create the most comprehensive description of the cerebellar connectome to date. Our strategy includes fast and distortion-free ultra-high-resolution dMRI acquisitions, novel anatomically constrained and curated cerebellar tractography, deep learning joint parcellation of fibers and nuclei for fine-grained atlasing, and expert neuroanatomical generation of the intrinsic and extrinsic cerebellar connectivity matrices from non-human primate tracer studies. Overall, these steps will enable robust in-vivo tracing of the cerebellar connectome of the human brain at an unprecedented spatial resolution. Our final deliverable will be a comprehensive, anatomically curated atlas of the human cerebellar connectome, which will enable the study of the cerebellar connectome in health and disease. We will publicly release the atlas, the monkey connectivity matrices, all extracted fascicles, all acquired images, and all software as open source.