A Novel Contrast for DTI Visualization for Thalamus Delineation

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Computational Neuroscience

Background:

• Automatic segmentation of thalamic nuclei using diffusion tensor imaging (DTI) is challenging due to poor contrast between nuclei.
• Manual delineation is difficult without clear visual boundaries in DTI data.

Purpose of the Study:

• To develop a novel DTI visualization technique that generates an edge map for improved manual delineation of thalamic nuclei.
• To validate the utility of this edge map for segmenting thalamic nuclei.

Main Methods:

• Computed principal fiber orientation from diffusion tensors at each voxel.
• Represented 3D eigenvector orientations as 5D vectors.
• Calculated the spatial gradient of these vectors and its Frobenius norm to create an edge map highlighting orientation changes.

Main Results:

• The novel gray-scale contrast successfully accentuates changes in principal eigenvector orientation.
• Distinct edges between major thalamic nuclei were revealed, facilitating manual delineation.
• A protocol for manual delineation using the edge map and registered T1-weighted MRI was described.

Conclusions:

• The developed DTI visualization method provides a crucial edge map for delineating thalamic nuclei.
• This technique significantly improves the feasibility of manual segmentation of thalamic nuclei.
• Preliminary evaluations suggest potential for accurate volume assessment of thalamic nuclei.