Automated longitudinal intra-subject analysis (ALISA) for diffusion MRI tractography

Area of Science:

• Neuroimaging
• Medical Physics
• Computational Neuroscience

Background:

• Fiber tractography (FT) reconstructs white matter (WM) pathways from diffusion tensor imaging (DTI) data.
• Manual FT segmentations are reliable but labor-intensive and time-consuming.
• Automated methods for FT exist, but manual segmentation is often considered more accurate.

Purpose of the Study:

• To develop and evaluate an automated longitudinal intra-subject analysis (ALISA) approach for FT.
• To assess if ALISA preserves the reliability of manual FT segmentations.
• To compare ALISA with automated inter-subject analysis for longitudinal DTI studies.

Main Methods:

• Proposed ALISA for automated longitudinal intra-subject FT segmentation.
• Collected DTI data from healthy children scanned multiple times over several months.
• Compared ALISA results with manual FT segmentations and automated inter-subject analysis.

Main Results:

• ALISA demonstrated high precision and accuracy, comparable to manual FT segmentations.
• The efficiency of ALISA did not compromise the quality of FT results.
• Automated inter-subject analysis did not achieve similar accuracy levels.

Conclusions:

• ALISA provides an efficient and reliable automated approach for longitudinal intra-subject FT analysis.
• ALISA is a viable alternative to manual segmentation for tracking WM changes over time.
• Automated inter-subject FT analysis is less accurate for longitudinal studies in subjects without gross pathology.