TractEdit: An open-source interactive tool for virtual dissection and manual refinement of diffusion MRI tractography.

Authors

Abstract

Accurate reconstruction of white matter pathways is essential for connectomics and pre-surgical planning. However, tractography algorithms inherently generate false positives, necessitating manual refinement ("virtual dissection") to isolate specific bundles and define ground-truth datasets. Existing tools are often hindered by format incompatibility, closed-source architectures, or a functional disconnect between 3D streamline visualization and precise slice-based anatomical editing. We sought to address these limitations by developing a lightweight, open-source tool optimized for the interactive cleaning and validation of tractography data. We introduced TractEdit, a Python-based desktop application built upon the Visualization Toolkit (VTK) and FURY visualization frameworks. TractEdit implements a hybrid protocol combining 3D streamline selection with voxel-level Region of Interest (ROI) definitions, including FreeSurfer parcellation-based ROI filtering and direct drawing tools (pencil, eraser, geometric shapes) on orthogonal anatomical slices. The application natively supports standard (.trk, .tck), next-generation (.trx), and alternative (.vtk, .vtp) file formats, utilizing efficient memory mapping for large datasets and ensuring seamless cross-format conversion. TractEdit enables real-time boolean logic filtering (inclusion/exclusion) and intuitive point-and-click manual bundle segmentation. The software automates the calculation of comprehensive bundle analytics, including centroids, medoids, and Track Density Imaging (TDI) maps, alongside an "Orientation Distribution Functions (ODF) Tunnel View" that projects in 3D Spherical Harmonic coefficients exclusively along selected streamlines to verify fiber alignment with the underlying diffusion signal. Finally, a dedicated export module enables the serialization of validated bundles into interactive HTML5 files for browser-based visualization. By bridging the gap between automated reconstruction and manual validation, TractEdit facilitates the rigorous quality control of diffusion MRI tractographic data. Its support for the TRX standard and integration of microstructural visualization makes it a versatile resource for neuroimaging researchers aiming to refine connectivity analyses or generate high-quality training data for machine learning models.

Tags

Topics