tract segmentation

tract segmentation segments white matter bundles from diffusion MRI to produce tract-specific segmentations and tractography-ready outputs for bundle-specific white matter analysis.

Key Features:

• Tract Orientation Mapping (TOM): Creates tract-specific orientation maps from fiber orientation distribution (FOD) peaks that represent voxel-wise principal orientations for individual tracts.
• Automated Segmentation: Segments major white matter tracts including tract outlines and start/end regions without manual dissection.
• Custom Probabilistic Tracking: Implements a probabilistic tracking algorithm that samples from a Gaussian distribution centered on each TOM peak to generate bundle-specific tractograms.
• Performance and Validation: Validated across 72 bundles using high-quality, low-quality, and phantom data and benchmarked against seven state-of-the-art methods, showing faster and more accurate results without requiring whole-brain tractography or non-linear registration.
• Generalizability: Demonstrates applicability across scanners, acquisition settings, and pathologies, evaluated on 17 different datasets.

Scientific Applications:

• Neuroscience Research: Enables mapping of brain connectivity and investigation of white matter pathways at the bundle level.
• Clinical Diagnostics: Supports diagnosis and monitoring of neurological disorders through bundle-specific tractography and segmentation.
• Pathology Studies: Facilitates analysis of how pathologies affect white matter integrity and connectivity.

Methodology:

Generate tract-specific orientation maps from FOD peaks (TOM), segment tract outlines and start/end regions, and perform probabilistic tracking by sampling Gaussian distributions centered on TOM peaks.