A probabilistic atlas of human brainstem pathways based on connectome imaging data

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• The brainstem is crucial for vital functions, cognition, and emotion, with its pathways connecting the brain and spinal cord.
• Diffusion MRI has limitations in imaging brainstem pathways due to resolution and distortion artifacts, despite its success elsewhere.
• Existing studies using Human Connectome Project (HCP) data often overlook significant brainstem distortions and subject variability.

Purpose of the Study:

• To develop a high-resolution probabilistic atlas of 23 major human brainstem fiber bundles.
• To overcome limitations of previous brainstem imaging studies, including susceptibility artifacts and lack of subject variability consideration.
• To provide a publicly available resource for researchers studying brainstem anatomy and function.

Main Methods:

• Rigorous quality control of Human Connectome Project (HCP) data to exclude subjects with severe brainstem distortions.
• Systematic manual delineation of 1300 regions of interest (ROIs) across 20 HCP subjects for tractography.
• Application of advanced connectome modeling, including high-order fiber orientation distribution reconstruction and topography-preserving tract filtering.

Main Results:

• Successful reconstruction of 23 major brainstem fiber bundles for each subject.
• Generation of probabilistic atlases in MNI152 space, demonstrating anatomically faithful reconstructions.
• Achieved improved performance compared to existing atlases, particularly for cerebellar peduncles.

Conclusions:

• The developed probabilistic atlas of brainstem pathways offers a significant improvement over existing resources.
• The atlas provides a valuable, publicly accessible tool for neuroimaging researchers.
• This work enhances the study of brainstem connectivity and its role in neurological and psychiatric conditions.