Registration of neuroimaging data: implementation and clinical applications

Area of Science:

• Medical imaging
• Computational anatomy
• Image analysis

Background:

• Accurate comparison of medical scan data requires images to be aligned, where each voxel corresponds to a specific anatomical location.
• Intrasubject registration involves aligning scans within the same individual, while intersubject registration is more complex, requiring the removal of individual anatomical variations.

Purpose of the Study:

• To describe automated methods for intrasubject image registration within and between modalities.
• To present automated methods for intersubject registration, specifically for aligning a 3D brain atlas with patient scans.

Main Methods:

• Automated image registration techniques for intrasubject alignment (translation and rotation).
• Automated image registration techniques for intersubject alignment, focusing on removing anatomical variability.
• Application of these methods to clinical examples, including 3D brain atlas registration.

Main Results:

• Successful implementation of automated intrasubject registration for scans within and between different imaging modalities.
• Demonstration of automated intersubject registration for aligning patient brain scans with a standardized 3D brain atlas.
• Clinical examples illustrating the effectiveness of the described registration methods.

Conclusions:

• Automated image registration methods are effective for both intrasubject and intersubject alignment.
• These techniques are crucial for enabling accurate comparisons of medical imaging data across scans and individuals.
• The described methods facilitate the integration of anatomical atlases with patient-specific imaging data.