A Study Of Alzheimer’s Disease Neural Fingerprint Based On Diffusion Magnetic Resonance Imaging

As the worlds’ population aging speed accelerating, Alzheimer’s Disease(AD) has become one of the most severe diseases that affect the health and life quality of aged people in the 21 st century. At present, there is no complete cure for AD and we can only delay the development of AD. So the early diagnosis of AD is particularly important. The modern clinical diagnostic approach to AD depends on clinical symptoms and MMSE scores. The appearance of Magnetic Resonance Imaging(MRI) provided a non-invasive method for the detection of neurological diseases. Although the volume increase of AD patients was observed through MRI, the quantization problem of AD features remains to be soluted. In this study, cerebral structures that may be highly related to AD were quantitatively analyzed based on diffusion MRI. The quantified features based on morphology and tissue characters were studied and AD’s neural fingerprint was constructed. This study could be the scientific basis of further study on AD.Firstly, since there is no visible lesions on AD patients’ MRI images, single cerebral structure should be studied for the quantitative analysis of these images. So it is especially important to segment the images to anatomical structures precisely. This study collaborated with the lab of neuroimaging in Johns Hopkins University. The multi-atlas parcellation method for T1 images adopted in this study is a state-of-the-art multi-atlas based segmentation method in the neuroimage segmentation area. A method that could map the T1 images’ segmentation result to other parametric images was proposed, which was the basis of the multi-parametric-based study of neural fingerprint.Secondly, based on the segmentation of different kinds of parametric images, we proposed a method for constructing AD neural fingerprint. For those highly related cerebral structures, different kinds of quantitative analysis methods were applied on parametric images. By combining these analysis results, a quantitative description method which could reflect AD patients’ cerebral characters was proposed and representative AD neural fingerprint was constructed.The result showed the potential of constructing AD neural fingerprint based on multi-parametric images and this could be the basis of further comprehensive study of AD. In addition, this study provided a technical solution for quantize AD patients’ cerebral features and constructing AD neural fingerprint.