Construction of Chinese Human Brain MRI Probabilistic Atlas Using HAMMER Elastic Reglstration Algorithm

摘要

Neural anatomical atlases serve as a reference coordinate system in neuroscience and clinical studies. Up to date, the most commonly used human brain templates are the Talairach brain coordinate system. The standard MNI (Montreal Neuroscience Institute) brain templates generated in early 1990's were based on a series of MRI scans from people of the western society. However, fundamental genetic and environmental disparities exist between the oriental and western human subjects, resulting in the overall and local brain shape and volume disparities. Such difference may cause mis-localization of activated brain regions during functional MRI experiments. It may also cause positional mismatches in the image-guided stereotactic neurosurgery. Therefore, it is of great significance to develop Chinese probabilistic brain atlases as the basis of Chinese human brain research projects. Here we present a preliminary study to construct MRI-based Chinese human brain templates using the automatic deformable HAMMER ( Hierarchical Attribute Matching Mechanism for Elastic Registration) registration algorithm. HAMMER algorithm is a robust registration algorithm which yields sharp average brain templates with high fidelity.