The diffusion-simulated connectivity (DiSCo) dataset
The methodological development in the mapping of the brain structural connectome from diffusion-weighted mag-netic resonance imaging (DW-MRI) has raised many hopes in the neuroscientific community. Indeed, the knowledge of the connections between different brain regions is funda-mental to study brain anatomy and function. The reliability of the structural connectome is therefore of paramount im-portance. In the search for accuracy, researchers have given particular attention to linking white matter tractography methods - used for estimating the connectome - with infor-mation about the microstructure of the nervous tissue. The creation and validation of methods in this context were ham-pered by a lack of practical numerical phantoms. To achieve this, we created a numerical phantom that mimics complex anatomical fibre pathway trajectories while also accounting for microstructural features such as axonal diameter distri-bution, myelin presence, and variable packing densities. The substrate has a micrometric resolution and an unprecedented size of 1 cubic millimetre to mimic an image acquisition matrix of 40 x 40 x 40 voxels. DW-MRI images were obtained from Monte Carlo simulations of spin dynamics to enable the validation of quantitative tractography. The phantom is composed of 12,196 synthetic tubular fibres with diameters ranging from 1.4 mu m to 4.2 mu m, interconnecting sixteen regions of interest. The simulated images capture the microscopic properties of the tissue (e.g. fibre diameter, water diffusing within and around fibres, free water compartment), while also having desirable macroscopic properties resembling the anatomy, such as the smoothness of the fibre trajectories. While previous phantoms were used to validate either tractography or microstructure, this phantom can enable a better assessment of the connectome estimation's reliability on the one side, and its adherence to the actual microstructure of the nervous tissue on the other. (C) 2021 The Authors. Published by Elsevier Inc.
WOS:000709964400011
2021-10-01
38
107429
REVIEWED