Incorporating anatomical knowledge into iterative tractography

Determine how to fully incorporate anatomical knowledge into iterative tractography algorithms, including expert knowledge about bundle paths, shapes, smoothness, fanning, sheet-like organization, and relationships to surrounding bundles.

Background

The paper describes a fundamental tension in tractography between following local diffusion information and producing streamlines that are globally anatomically plausible. Although bundle-specific methods and anatomically constrained tractography incorporate some anatomical priors, the authors emphasize that expert anatomical knowledge remains difficult to formalize and inject into general whole-brain iterative tracking algorithms.

This problem is unresolved because whole-brain tractography must represent diverse anatomical structures without relying exclusively on fixed rules or bundle-specific procedures. A solution would support machine-learning and classical tracking methods that generate anatomically plausible pathways while retaining broad coverage.