Data-driven mapping between functional connectomes using optimal\n transport

Functional connectomes derived from functional magnetic resonance imaging\nhave long been used to understand the functional organization of the brain.\nNevertheless, a connectome is intrinsically linked to the atlas used to create\nit. In other words, a connectome generated from one atlas is different in scale\nand resolution compared to a connectome generated from another atlas. Being\nable to map connectomes and derived results between different atlases without\nadditional pre-processing is a crucial step in improving interpretation and\ngeneralization between studies that use different atlases. Here, we use optimal\ntransport, a powerful mathematical technique, to find an optimum mapping\nbetween two atlases. This mapping is then used to transform time series from\none atlas to another in order to reconstruct a connectome. We validate our\napproach by comparing transformed connectomes against their "gold-standard"\ncounterparts (i.e., connectomes generated directly from an atlas) and\ndemonstrate the utility of transformed connectomes by applying these\nconnectomes to predictive models based on a different atlas. We show that these\ntransformed connectomes are significantly similar to their "gold-standard"\ncounterparts and maintain individual differences in brain-behavior\nassociations, demonstrating both the validity of our approach and its utility\nin downstream analyses. Overall, our approach is a promising avenue to increase\nthe generalization of connectome-based results across different atlases.\n