Evaluation of surface-based hippocampal registration using ground-truth subfield definitions

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Anatomy

Background:

• The hippocampus, crucial for learning and memory, has distinct subfields.
• Understanding hippocampal microstructure is vital for neurological disorder research.
• Aligning hippocampi with varying morphologies presents a significant challenge.

Purpose of the Study:

• To present a novel surface-based registration method for aligning hippocampi.
• To enable accurate parcellation, registration, and retrieval of homologous points between hippocampi of different shapes.
• To improve mechanistic understanding of learning and memory and aid clinical applications.

Main Methods:

• An analytical unfolding of the entire hippocampus into a 2D space.
• Registration of samples in the 2D unfolded space based on thickness, curvature, and gyrification.
• A contrast-agnostic and topology-preserving approach.

Main Results:

• Demonstrated superior alignment of hippocampal subfields compared to conventional methods.
• Successfully registered seven 3D histology samples with diverse morphologies.
• Validated the effectiveness of the surface-based registration technique.

Conclusions:

• The proposed surface-based registration method effectively addresses the challenge of aligning hippocampi with different morphologies.
• This method offers a robust tool for comparative neuroanatomy and clinical research.
• Improved hippocampal subfield alignment enhances our understanding of brain function and disease.