Consensus tractography: decreasing algorithm dependency to improve fiber reconstructions

Motivation: Diffusion MRI tractography is a powerful tool for analyzing non-invasively the brain white matter architecture. However, the reconstructions are highly algorithm- and parameter-dependent, providing only a partial view of the actual connectivity. Goal(s): To address this limitation, we developed the consensus tractography with the aim of obtaining more anatomically accurate reconstructions. Approach: The proposed method combines multiple tractograms and employs an advanced microstructure-informed tractography technique to control false positives and redundancy among streamlines. Results: Tests on synthetic data proved that consensus tractography consistently outperforms individual reconstructions, while the analysis of ten subjects showed its ability to better represent the underlying anatomical structure. Impact: We addressed the well-known algorithm dependency of tractography reconstructions and proposed a consensus-based strategy. Our solution provides clearer and more accurate white matter mapping, showing significant potential to facilitate precise findings and better support clinical applications.