Morphometric Analysis of Hippocampus based on Sparse Representation

Abstract

Alzheimer’s disease and mild cognitive impairment is an age-related neurodegenerative disease characterized by progressive loss of memory and irreversible cognitive functions. The hippocampus, a brain area critical for learning and memory process, has been known to be an important structure as a biomarker for Alzheimer’s disease and other neurological and psychiatric diseases. In this dissertation, the morphometric analysis be related to neurodegenerative disease within hippocampus was investigated. First, an automated hippocampal segmentation method based on a graph- cuts algorithm combined with atlas-based segmentation and morphological opening was proposed. The atlas-based segmentation was applied to define initial hippocampal region for a priori information on graph-cuts. The definition of initial seeds was further elaborated by incorporating estimation of partial volume probabilities at each voxel. The morphological opening was applied to reduce false positive of the result processed by graph-cuts. In the experiments with 27 healthy normal subjects and 37 subcortical vascular dementia patients, the proposed method showed more reliable results than the conventional segmentation method. Also as for segmentation accuracy, which is measured in terms of the ratios of false positive and false negative, the proposed method produced lower ratios than the conventional methods demonstrating its plausibility for accurate, robust and reliable segmentation of hippocampus. Second, a sparse representation approach applied to hippocampus using structural T1-weighted Magnetic resonance imaging and 18-fluorodeoxyglucose-positron emission tomography together to separate Alzheimer’s disease / mild cognitive impairment from healthy normal subjects was proposed. In experiments with Alzheimer’s disease Neuroimaging Initiative data, the proposed method showed more reliable classification performances. Also effect on classification accuracy of different parameters was evaluated. The multi-modal biomarkers from automatically defined hippocampal region of training subjects, which is discriminative and robust for Alzheimer’s disease or mild cognitive impairment classification was extracted. The proposed method could establish the understanding in morphometric analysis of hippocampus, and could be used for an early detection of brain disorders related to hippocampus.