Ution structural MR images were acquired with 3D magnetization prepared rapid gradient echo sequence (TR = 2,530 ms, TE = 3.5 ms, TI = 1,100 ms, FOV = 256 mm, flip angle = 7u, matrix size = 2566256, 192 sagittal slices, voxel size = 1.061.061.0 mm, no gap). All the images were acquired parallel to the anterior commissure osterior commissure line. To minimize motion artifact generated during image acquisition, each subject’s head was immobilized with cushions MedChemExpress PHCCC inside the coil. Each image was carefully checked by an experienced radiologist to ensure that they had no scanner artifacts, motion problems, or gross anatomical abnormalities.2.4 DARTEL-based T1 VBM Analysis Materials and Methods 2.1 Participants and InstrumentsWe recruited 330 healthy participants in northern Taiwan (mean age: 56.2622.0 years, range: 21?2; 57.9 males). Each participant was evaluated by a trained research assistant using the Mini-International Neuropsychiatric Interview [28]. The participants were screened using the Mini-Mental Status Examination (MMSE) and the Clinical Dementia Rating Scale. The exclusion criteria included the following: (1) Any Axis-I diagnosis according to the DSM-IV, such as mood disorders or psychotic disorders; 
(2) neurological disorders, such as dementia, head injury, stroke, or Parkinson disease; (3) illiteracy; (4) participants with an MMSE score below 24; (5) any chronic illness under medical control,Individual T1-weighted volumetric images were processed using Gaser’s VBM8 toolbox (http://dbm.neuro.uni-jena.de) within Statistical Parametric Mapping (SPM8, Wellcome Institute of Neurology, University College London, UK) executed in MATLAB 2010a (The MathWorks, Natick, MA, USA) under Linux 64-bit environment with recommended settings. VBM processing was performed as following procedure: 1) the anterior commissure was set as the origin of each T1-weighted image. 2) Segmentation approach in the VBM8 toolbox was applied in the initial native space that combined the nonlocal means denoising filter [29] and adaptive maximum a posteriori segmentation approach [30] with partial volume Avasimibe web estimation technique [31]. Images were further refined by applying an iterative hidden Markov random field model [32] to remove isolated voxels which were unlikely toBcl-2 and Age-Related Gray Matter Volume Changesbelong to a determinate tissue type, and to improve the quality of tissue segmentation. 3) To achieve higher accuracy of registration between subjects, the native space GM, white matter (WM), and CSF segments were initially affine registered to the 1516647 tissue probability maps in the Montreal Neurological Institute (MNI) standard space (http://www.mni.mcgill.ca/). 4) All affine registerted tissue segments were iteratively registered to group-based templates, which were generated from all images included in the current study through nonlinear warping using DARTEL (Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra) toolbox [33] that implemented in SPM8. 5) The nonlinear deformation parameters obtained in the previous step were used to modulate the GM, WM, and CSF tissue maps of participants’ brains so as to compare actual volumetric differences across groups. 6) Finally, the modulated tissue segments were converted into an isotropic voxel resolution of 16161 mm3. All normalized, segmented, and modulated MNI standard space images were smoothed with an 8-mm Gaussian kernel ahead of tissue volume calculation and voxel-wised group comparisons.Ution structural MR images were acquired with 3D magnetization prepared rapid gradient echo sequence (TR = 2,530 ms, TE = 3.5 ms, TI = 1,100 ms, FOV = 256 mm, flip angle = 7u, matrix size = 2566256, 192 sagittal slices, voxel size = 1.061.061.0 mm, no gap). All the images were acquired parallel to the anterior commissure osterior commissure line. To minimize motion artifact generated during image acquisition, each subject’s head was immobilized with cushions inside the coil. Each image was carefully checked by an experienced radiologist to ensure that they had no scanner artifacts, motion problems, or gross anatomical abnormalities.2.4 DARTEL-based T1 VBM Analysis Materials and Methods 2.1 Participants and InstrumentsWe recruited 330 healthy participants in northern Taiwan (mean age: 56.2622.0 years, range: 21?2; 57.9 males). Each participant was evaluated by a trained research assistant using the Mini-International Neuropsychiatric Interview [28]. The participants were screened using the Mini-Mental Status Examination (MMSE) and the Clinical Dementia Rating Scale. The exclusion criteria included the following: (1) Any Axis-I diagnosis according to the DSM-IV, such as mood disorders or psychotic disorders; (2) neurological disorders, such as dementia, head injury, stroke, or Parkinson disease; (3) illiteracy; (4) participants with an MMSE score below 24; (5) any chronic illness under medical control,Individual T1-weighted volumetric images were processed using Gaser’s VBM8 toolbox (http://dbm.neuro.uni-jena.de) within Statistical Parametric Mapping (SPM8, Wellcome Institute of Neurology, University College London, UK) executed in MATLAB 2010a (The MathWorks, Natick, MA, USA) under Linux 64-bit environment with recommended settings. VBM processing was performed as following procedure: 1) the anterior commissure was set as the origin of each T1-weighted image. 2) Segmentation approach in the VBM8 toolbox was applied in the initial native space that combined the nonlocal means denoising filter [29] and adaptive maximum a posteriori segmentation approach [30] with partial volume estimation technique [31]. Images were further refined by applying an iterative hidden Markov random field model [32] to remove isolated voxels which were unlikely toBcl-2 and Age-Related Gray Matter Volume Changesbelong to a determinate tissue type, and to improve the quality of tissue segmentation. 3) To achieve higher accuracy of registration between subjects, the native space GM, white matter (WM), and CSF segments were initially affine registered to the 1516647 tissue probability maps in the Montreal Neurological Institute (MNI) standard space (http://www.mni.mcgill.ca/). 4) All affine registerted tissue segments were iteratively registered to group-based templates, which were generated from all images included in the current study through nonlinear warping using DARTEL (Diffeomorphic Anatomical Registration Through 
Exponentiated Lie Algebra) toolbox [33] that implemented in SPM8. 5) The nonlinear deformation parameters obtained in the previous step were used to modulate the GM, WM, and CSF tissue maps of participants’ brains so as to compare actual volumetric differences across groups. 6) Finally, the modulated tissue segments were converted into an isotropic voxel resolution of 16161 mm3. All normalized, segmented, and modulated MNI standard space images were smoothed with an 8-mm Gaussian kernel ahead of tissue volume calculation and voxel-wised group comparisons.
