Xie, Long
Email Address
ORCID
Disciplines
Search Results
Now showing 1 - 1 of 1
Publication Multi-Template Analysis Of Human Perirhinal Cortex In Brain Mri: Explicitly Accounting For Anatomical Variability(2018-01-01) Xie, LongThe human brain is highly variable in terms of its folding pattern and the distribution of underlining cytoarchitecture. Current leading paradigms in brain structural MRI analysis have been almost completely based on methods that use a single template to capture variability between all subjects in a population, even when the underlying assumptions of these methods are not consistent with the actual anatomy. Failure to account for the anatomical variability in the analysis degrades the ability to reliably localize and accurately quantify brain regions in individual subjects. Although characterizing anatomical variability of the brain “on a whole” is desirable, it might not be feasible given an almost infinite number of anatomical configurations. However, when focusing on specific local brain regions, regional anatomical variability can often be described by discrete anatomical variants and thus can be characterized and modeled by multi-template analysis. This dissertation focuses on the human perirhinal cortex (PRC) in brain MRI. PRC is located in the medial temporal lobe (MTL) and plays important roles in semantic memory, episodic memory and visual processing systems. In addition, pathology studies have found that the PRC is the first site in the cortex that is affected by neurofibrillary tangle (NFT) pathology, the hallmark of Alzheimer’s disease (AD) that is more directly linked to neurodegeneration. Thus, accurate quantification of morphometry features of the PRC may have important utility in diagnosis and monitoring of early AD, as well as brain-behavior studies in the MTL. Probably due to its large anatomical variability, PRC is surprisingly overlooked by the biomedical image analysis literature. What makes PRC an ideal structure for multi-template analysis is that 97% of its anatomical variability can be accounted by three discrete anatomical variants, defined by the depth and the branching pattern of the anterior portion of the collateral sulcus (CS) adjacent to the PRC. In this dissertation, I introduced a novel multi-template analysis pipeline for the PRC in structural brain MRI. Anatomical variants of the PRC are identified automatically from structural MRI scans. Then, we explicitly construct a template and model anatomical variability for each anatomical variant. Experimental results show that the proposed technique is able to generate templates that recover the dominant discrete variants of the PRC and establish more meaningful correspondences between subjects than a single-template approach. In the application of discriminating AD patients from cognitively normal adults, the proposed pipeline generates measurements that are more sensitive to disease status and yields results that are consistent with the patterns of NFT pathology distribution. In addition, since the cortex is organized like a sheet, it is likely that the location of the earliest NFT pathology within the PRC is itself variable across different anatomical subtypes. By applying the proposed multi-template analysis pipeline to a large dataset with subjects at different stages of the AD spectrum, we find, for the first time, discrete patterns of spatial distribution of cortical thinning between anatomical variants. Incorporating this variability in future studies will likely further improve the sensitivity of MRI-derived measures of the PRC to early detection and monitoring of AD.