Subject Based Registration for Individualized Analysis of Diffusion Tensor MRI.

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Diffusion Tensor Imaging

Background:

• Standard brain imaging analysis relies on registering individual brains to a common atlas template.
• This atlas-centered registration can introduce errors due to anatomical differences between subjects and the template.
• Such misregistration errors can lead to inaccurate identification of brain abnormalities.

Purpose of the Study:

• To introduce and evaluate a novel subject-centered registration approach for brain imaging analysis.
• To eliminate registration errors by using the subject's brain as the analysis template.
• To improve the accuracy of identifying diffusion abnormalities, particularly in mild traumatic brain injury (mTBI).

Main Methods:

• Developed a subject-centered registration technique where the individual brain serves as the template.
• Applied this method to analyze Fractional Anisotropy (FA) values from Diffusion Tensor Imaging (DTI) data.
• Compared results from subject-centered analysis with two separate atlas-centered analyses in 20 mTBI patients and 48 healthy controls.

Main Results:

• The subject-centered approach identified significantly fewer regions of abnormally low FA compared to atlas-centered methods.
• Atlas-centered analyses misregistered 9.8%–13.3% of gray matter (GM) and cerebrospinal fluid (CSF) onto white matter (WM), creating false positives.
• Subject-centered analysis revealed abnormalities in 9/20 subjects, while atlas-centered methods showed abnormalities in 19/20 and 20/20 subjects.

Conclusions:

• Subject-centered registration effectively eliminates misregistration errors inherent in atlas-centered approaches.
• This method provides a more accurate and reliable identification of diffusion abnormalities, reducing false positives.
• Visualizing abnormalities directly in the subject's native space enhances correlation with other imaging data and clinical findings.