The data were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (\url{https://adni.loni.usc.edu}). Launched in 2003 as a public-private partnership, ADNI is led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessments can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer's disease (AD). The ADNI community provides diffusion-weighted imaging (DWI) data for General Electric (GE) MRI sessions from ADNI2 and all MRI sessions from ADNI3. Images were deliberately chosen from various manufacturers to ensure the inclusion of three-dimensional (3D) T1-weighted (T1W) imaging and two-dimensional echo-planar DWI. We thoroughly analyzed a total of 237 MRI and DWI brain images, encompassing Normal Cohort subjects and those at different disease stages: Early MCI (EMCI), Late MCI (LMCI), and Alzheimer's disease (AD), as per the guidelines discussed in Endy's book [cite John Endy]. The downloaded DICOM brain images were converted into NIFTI files using the Heudiconv software and organized into structured directory layouts (ANAT, DWI) in compliance with the Brain Imaging Data Structure (BIDS) format. Diffusion and anatomical images were then automatically preprocessed using a combination of software packages, including MRTRIX3 (\url{http://www.mrtrix.org}), FreeSurfer (\url{http://surfer.nmr.mgh.harvard.edu/}), Advanced Normalization Tools (ANTs) (\url{http://stnava.github.io/ANTs/}), and the FMRIB Software Library (FSL) (\url{https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/}). The preprocessing was guided by recent advancements in the field \cite{mabrouk2023contribution,wu2021interactions}, involving multiple steps: Denoising: First, denoising is performed \cite{liu2014multimodal}. The b0 images are extracted from the diffusion data acquired in the anterior-to-posterior (AP) direction. Eddy current corrections are applied to the phase-encoding AP direction of the b0 images using the FSL command. Bias field correction and skull stripping are subsequently carried out using Advanced Normalization Tools (ANTs). Tissue-specific Analysis: Various basis functions are then estimated for each tissue type (white matter [WM], gray matter [GM], and cerebrospinal fluid [CSF]) to conduct multi-shell multi-tissue constrained spherical deconvolution, generating images of fiber orientation densities (FOD) overlaid on the estimated tissues. The FODs are normalized to enable inter-subject comparisons. In our experiments, normalization is restricted to two tissues (WM, CSF) for two shells (b=0 and b=1000) in the DW images. GM/WM Boundary Creation: A gray matter/white matter boundary is created for seed analysis by converting the anatomical images to MRTRIX3 format, segmenting them into five tissue categories (1=GM; 2=Subcortical GM; 3=WM; 4=CSF; 5=Pathological tissue), co-registering the averaged b0 diffusion images, and finally creating a boundary that separates gray matter from white matter. Tractography and Connectivity Matrix: Finally, streamlines are generated using MRTRIX3's default probabilistic tractography approach and are subsequently refined. These streamlines are used to create a weighted symmetric connectivity matrix, denoted as ( W ) (84x84), where 84 represents the number of Regions of Interest (ROIs). These ROIs consist of 42 parcellations for each hemisphere, obtained through the recon-all command from FreeSurfer, following the Desikan-Killiany atlas. The elements of the connectivity matrix, denoted as Wi,j, represent the strength of connectivity between nodes. This strength is determined by normalizing the number of fibers connecting the i and j nodes.