Comparison of two VBM arithmetic based on Alzheimer's Disease MRI analysis

Abstract

Objective In order to study the difference of process and result, both optimized VBM and DARTEL arithmetic were used in the MR images of Alzheimer’s Disease . Materials and Methods Baseline and 3 years longitudinal MCI controls were enrolled in the study. Gray matter differences of the whole brain were assessed using both method. Results. Both method has grey matter atrophy in bilateral superior temporal gyrus, parahippocampa gyrus, right anterior cingulate,right cerebellum anterior lobe. But the clusters of the optimized VBM is smaller than the DARTEL. Further more, some anatomic region can’t be reported in the optimized VBM.Conclusion DARTEL is more robust than the optimized VBM based on AD MRI analysis. Key wordsDARTEL; Alzheimer’s Disease; optimized VBM; MRI I.INTRODUCTION As the most common form of dementia, Alzheimer’s disease (AD) currently affects more than 60 millions peoples in China. The patients may have some pathological damage, such as neuritic plaques, neurofibrillary tangles and synapse loss of cortical neurons[1]. With the development of aging society in our country, AD patients will give a heavy burden to the families and society. In recent years, scientific interest has also focused on mild cognitive impairment (MCI), a pre-dementia stage increased risk of future diagnosis of dementia, relative to the general population[2]. MCI is considered as a transitional stage between normal aging and dementia. So the early diagnosis and the treatment will be necessary to control the convertion from MCI to AD. Voxel-Based Morphometry (VBM) [3]method was used to evaluate the gray matter (GM) and white matter (WM) morphological changes of the living brain based on the structure MR images. This method could objectively detect the differences in local brain regions and the brain tissue composition. In this study, we compare the differences between optimized VBM method and VBM-DARTEL(Diffeomorphic Anatomical Registration Through Exponentiated Lie) method. In addition, based on the same longitudinal MCI groups’scans, we used two different VBM methods respectively and carried out a longitudinal analysis. This study explore the probable results with different algorithms. II. MATERIALS AND METHODS A. Image acquisition The Alzheimer’s Disease Neuroimaging Initiative (ADNI) builded in 2003 is a consortium study to observe NC, MCI, and AD[4]. All data were all got from ADNI’s MRI examinations of the brain were performed on a 3.0 T MRI scanner. We acquired a high-resolution T1-weighted MagnetisationPrepared Rapidly Acquired Gradient echo (MP-RAGE) 3D-sequence. Including TR/TE=8600ms/3.8ms, FOV=240mm*220mm, a pixel matrix=240mm*220mm, matrix size= 256×224, TI=900 ms, flip angle=9. TABLE 1:Demographic variables and CDR for the different groups. Group Baseline 3-year Sample size (male/female) 12/5 12/5 Age(years±SD) 72.8±10 75.9±10 CDR(0.5/1/2) 19/0/0 7/6/4 The baseline was the group which did MRI scans for the first time. The 3-year was the same group which did MRI scans after 3 years.The CDR was a numeric scale used to quantify the severity of symptoms of dementia. It characterized six domains of cognitive and functional performance: memory, orientation, judgment & problem solving, community affairs, home & hobbies, and personal care [5]. CDR score was useful to vary the level of impairment: 0 = No impairment, 0.5, 1, 2, and 3 indicated very mild, mild, moderate and severe dementia. Details are shown in table 1。 B. Image processing The ways of two method were processed in SPM8 (www.fil.ion.ucl.ac.uk/spm/, London, UK) .  VBM and optimized VBM method Wright[6] presents an initial conception of structural brain MR image analysis based on voxel in 1995. Ashburner and Friston[7] formally presented Voxel-Based Morphometry in 2000.From then on,VBM method attracts more and more researchers' eyes in recent years. VBM is a morphometry based on voxel. It is a whole-brain,unbiased technique for characterizing regional cerebral volume and tissue concentration differences in structural magnetic resonance images. The data processing of VBM is showed in Fig 1. Figure 1 Procedure of VBM method Spatical normalization based on the MRI template, involves transforming all the subjects’ data to the same stereotactic space.and then doing affine trasform from original image to template image,and then correcting in part of non-linear deforrmation.the purpose of spatical normalization is correcting the difference of whole brain.After normalization, scans were segmented into gray matter (GM) , white matter(WM) and cerebrospinal fluid(CSF) based on prior probability of brain and the information of gray-scale image. Next, the image after segmented were smoothed with the full-width at half maximum (FWHM) of Gaussian smoothing kernel.The general kernel of gaussian function is