Estimation of brain internal structures by deforming brain atlas using finite element method

Area of Science:

• Neuroimaging
• Medical image analysis
• Computational anatomy

Background:

• Conventional brain atlas deformation relies on surface landmarks, which can be difficult to accurately correspond due to variations in sulcal patterns.
• The relationship between external brain shape and internal structures is not fully understood, limiting the reliability of existing atlas-based estimation methods.

Purpose of the Study:

• To develop a more reliable method for estimating internal brain structures using a standard brain atlas.
• To overcome the limitations of conventional deformation techniques that depend solely on surface contours.

Main Methods:

• A novel finite element method (FEM) is proposed for deforming a standard brain atlas to match a patient's brain.
• The method utilizes landmarks from both the brain surface and detectable internal structures visible in MRI scans.
• This approach enhances the accuracy of atlas fitting by incorporating internal anatomical information.

Main Results:

• The proposed FEM-based deformation method allows for more precise fitting of the brain atlas to patient-specific anatomy.
• By incorporating internal landmarks, the method improves the reliability of estimating internal brain structures compared to surface-only methods.
• This technique addresses the challenges associated with sulcal correspondence and the unclear surface-to-internal structure relationship.

Conclusions:

• The developed FEM approach offers a more robust and accurate method for patient-specific brain structure estimation.
• This technique has the potential to improve neurosurgical planning and understanding of individual brain anatomy.
• The use of both internal and surface landmarks represents a significant advancement in computational neuroanatomy.