| Objective: To investigate the brain activity in resting-state of patients with mild Alzheimer's disease by voxle-wise corralation. To detect PCC functional connectivity characteristics and default-mode network activity distinguishes mild Alzheimer's disease from normal aging.Materials and Methods:1. Subjects1.1 Patients with AD were recruited from the gerontism dysmnesia clinic department, the geratology clinic institution at Guangdong General Hospital, Guangdong, China. The subjects groups consisted of 22 patients with possible or probable AD at the early or mild stage, including 10 male and 12 female. The patients age ranged between 50 and 80. All subjects were right-handed. All patients met the following criteria: (1) National Institute of Neurological and Communicative Disorders and Stroke/AD and Related Disorders Association (NINCDS-ADRDA) diagnostic criteria for possible or/and probable AD. (2) their mini mental state examination (MMSE) scores less than 24. The AD patients were diagnosed by at least two veteran geropsychiatric doctors. (3) their His scores less than 4 to exclude the vascular dementia and mixing dementia. (4)The patients'CDR scores ranged between 0.5 and 1. (5) No neuropathy or psychosis history. (6) Conventional MR findings is normal, besides cerebral atrophy and few normal in deep white matter (d<1cm) .1.2 22 control subjects were recruited from local communities, including 11 male and 11 female. Control subjects were healthy volunteers of age and education degrees similar to the patients. All control subjects met the following criteria: (1) had no memory impairment or cognitive disorders. (2) No neuropathy or psychosis history. (3) Conventional MR findings is normal, besides cerebral atrophy and few normal in deep white matter (d<1cm). (4) Their MMSE scores ranged from28 to 30. (5) CDR scores were all zero.2. Data acquisition All subjects wore earplug and were instructed to lie in a supine position and form padding was used to limit headmovement. Respiration and pulse were test before the examination and the subjects were asked about the examination to make sure that the data is valid.2.1 Equipment: All subjects were imaged by a standard head coil of a 3.0 TMR scanner in a dim scan room.2.2 Structural images (the axial TT1WI(TR/TE, 580/18), T2WI(TR/TE, 5100/130), FLAIR (TR/TE, 9600/ 110 TI=2400), slicethickness=5mm, slice gap=1 mm, FOV=240mm, included 25 contiguous slices covering wholecerebrum. The baseline parallel the line from anterior commissure to commissura posterior.2.3 The resting-state fMRI was performed, where subjects were instructed to lie quietly stilltheir eyes closed and to clear their heads of all thoughts. Functional images were acquired byusing a gradient echo-planar imaging (GRE-EPI) sequence (TR = 2000ms, TE =35ms, flipangle=80, slice thickness=5mm, slice gap=0mm, FOV=240mm, matrix=64×64), Each time-pointincluded 29 contiguous slices covering whole cerebrum. The EPI scan lasted 5min 12s.2.4 3D FSPGRIR scan was performed (TR = 7.6 ms, TE = 3.4 ms, flip angle = 20, slice th= 1mm, slice gap = 0mm, FOV= 240mm, matrix = 256×256, NEX 1).3. Data analysis3.1 Data preprocFMRI datap://www.fil.ion.ucl.ac.uk/spm) and DPARSF_V2.0 Basic Edition (Yan and Zang,2010) based Matlab R2008b. Images format was transformed from Dicom to NifTI firstly.Images were corrected for slice-timing and realigned to the middle image for rigid-body headmovement correction (patient data with movement greater than 1.5mm and 1.5 degree wasdiscarded). Afterwards, the functional images were normalized into 3D structral images and thenstandard stereotaxic anatomical Montreal Neurological Institute (MNI) space. The normalizedvolumes were resampled to a voxel size of 3mm×3mm×3mm. As a final step, EPI images werespatially smoothed using isotropic Gaussian filter (4mm FWHM). Before functional connectivityanalysis, all image data was simply filtered by a phase-insensitive bandpass filter (pass band0.01–0.08Hz) to reduce the effect of low frequency drift and high frequency physiological noise,and voxel intensities were scaled by dividing the value at each time point by the mean value ofthe whole brain image at that time point to minimizing the effect of global drift. We selected the PCregion of interest (ROI), which consist ofBrodmann areas 29, 30, 23, and 31 using the WFU-PickAtlas software.3.3 Calculating connectivity intensityIn order to perform the ROI-based correlation analysis, the mean PCC/Pcu signal intensitycalculated by averaging the time series of all voxels in the selected ROI. The resulting timecourse was used to perform Pearson linear correlation analysis with all voxels of the brain data,and the mask images were acquired with a threshold of correlation coefficients. A Fisherz-transform was applied to normalize the correlation coefficients.4. Statistical analysisStatistical analysised into REST for random-effects one-sample t-test to determine brain regions withsignificant connectivity (P<0.001,AlphaSim,K≥6 voxle)to the PCC/Pcu within each group.These individual scores were also entered into REST for random-effects analyses andtwo-sample t tests to identify regions with significant differences(P<0.005,AlphaSim,K≥12voxle) in connectivity to the PCC/Pcu between groups. Regions with statistical significance weremasked on MNI brain templates, display in axial, coronal and Sagittal. Observe and record theencephalic region, volume (cluster), intensity and coordinate (MNI) with significant differences.Results1.1. Dem22 Patients with mild AD and 22 normale study. Demographic characteristics and neuropsychological scores were shown in Table1-1.No significant differences in gender, age and educational level were noted between bothexperimental groups (P＞.05). But MMSE was significantly different between groups using thetwo-sample t test (P< 0.000).1.2. PCC connectivity: within-group analysesData from two patients were eliminated from the subject group because of excessive head movement. Within-group analysis was performed using REST, and the statistical threshold wasset at the level of corrected with AlphaSim P value <0.001 and extent threshold = 6 voxels.Regions with connectivity to the PCC/Pcu in each of groups were obtained. Primary regionsinvolved in the network in control group include the precuneus,cuneus,inferior parietal lobules,dorsomedial prefrontal cortices (DMPFC),dorsolateral prefrontal cortices (DLPFC), part oftemporal and occipital lobes, bilateral thalami,hippocampi,parahippocampi, insulae, (P<0.001,AlphaSim,K≥6 voxels). These regions coincide with regions underlying the default modenetwork.. Primary regions involved in the network in mild AD group include bilateralprecuneus,PCC,inferior parietal lobules(P<0.001,AlphaSim,K≥6voxle).1.3. PCC connectivity: between-group analysesBetween-group analyses were performed using an REST two-sample t-test, with the statistichold of corrected with AlphaSim P value<0.005 and extent threshold = 12 voxels. Thebetween-group difference in PCC/Pcu connectivity was obtained. Compared control group withthe AD group, a pattern of reduced interaction with PCC/Pcu in AD groups or areas ofdisconnectivity was found. Decreased connectivity was detected in the regions of the bilateralvisual cortex, the ITC, the posterior orbital frontal cortex (POFC), ventral MPFC and precuneus.Conversely, compared the AD group with the control group, a pattern of higher interaction withPCC/Pcu were detected in AD brain groups or increasing connectivity. Significant coactivationwas found in bilateral DLPFC, left basal ganglia, left ITC, bilateral posterior orbital prefrontalcortex, MPFC in mild AD brain. These areas may provide compensation for the malfunction ofmemory system.1.4 Among all the areas of DMN, only few areas such as left hippocampus presented gray matteratrophy as well as decreased resting-state activity in the mild AD group. All the other areasmanifested decreased resting-state activity before the gray matter atrophy was presented.Conclusion: Voxel-based correlation analysis can detect the abnormal functional connectivPCC/Pcu with other brain areas in resting state. The distinctive altearation of functionalconnectivity network in mild AD before the structural alteration can be demonstrated, which canbe important biomarker for the early diagnosis of AD.[Keywords] Alzheimer's disease; functional MRI; resting state; correlation analysis Objective: To detect white matter abnormalities in patients with mild Alzheimer's disease (AD) by voxel-based DTI analysis. To compare the distribution pattern between abnornal white matter and the abnormal functional connectivity in resting state. Materials and Methods: 22 Patients with mild AD and 22 normal aging volunteers as control subjects were enrolled in the study. All the subjects were scanned with 3.0T MRI. The original images of DTI was corrected to move the constructed defect due to motion and eddy current. Fractional anisotropy (FA) maps were computed on Slice 3D, then preprocessed using SPM8 to make voxel-wise comparison of anisotropy in whole brain between patients and the control group.1. Subjects1.1 Patients with AD were recruited from the gerontism dysmnesia clinic department, the geratology clinic institution at Guangdong General Hospital, Guangdong, China. The subjects groups consisted of 22 patients with possible or probable AD at the early or mild stage, including 10 male and 12 female. The patients age ranged between 50 and 80. All subjects were right-handed. All patients met the following criteria: (1) National Institute of Neurological and Communicative Disorders and Stroke/AD and Related Disorders Association (NINCDS-ADRDA) diagnostic criteria for possible or/and probable AD. (2) their mini mental state examination (MMSE) scores less than 24. The AD patients were diagnosed by at least two veteran geropsychiatric doctors. (3) their His scores less than 4 to exclude the vascular dementia and mixing dementia. (4)The patients'CDR scores ranged between 0.5 and 1. (5) No neuropathy or psychosis history. (6) Conventional MR findings is normal, besides cerebral atrophy and few normal in deep white matter (d<1cm) .1.2 22 control subjects were recruited from local communities, including 11 male and 11 female. Control subjects were healthy volunteers of age and education degrees similar to the patients. All control subjects met the following criteria: (1) had no memory impairment or cognitive disorders. (2) No neuropathy or psychosis history. (3) Conventional MR findings is normal, besides cerebral atrophy and few normal in deep white matter (d<1cm). (4) Their MMSE scores ranged from28 to 30. (5) CDR scores were all zero.2. Data acquisition All subjects wore earplug and were instructed to lie in a supine position and form padding was used to limit headmovement. Respiration and pulse were test before the examination and the subjects were asked about the examination to make sure that the data is valid.2.1 Equipment: All subjects were imaged by a standard head coil of a 3.0 T Signa Excite HD MR scanner in a dim scan room.2.2 Structural images (the axial T1WI, T2WI and T2 FLAIR sequences) were performed first T1WI(TR/TE,580/18), T2WI(TR/TE,5100/ 130), FLAIR (TR/TE, 9600/ 110 TI=2400), slice thickness=5mm, slice gap=1 mm, FOV=240mm, included 25 contiguous slices covering whole cerebrum. The baseline parallel the line from anterior commissure to commissura posterior.2.3 The DTI was performed, where subjects were instructed to lie quietly still with their eyes closed. The baseline parallel the line from anterior commissure to commissura posterior. Diffution tensor images were acquired by using a spin echo-planar imaging (DW-EPI) sequence (TR = 8000ms, TE = 76ms, slice thickness=4mm, slice gap=0mm, FOV=240mm, matrix = 128×128), NEX 1. b values were 0 and 1000s/mm2, respectively. The gradient magnetic field was exerted on the 25 different non-linear direction. Each direction of gradient magnetic field included 30 contiguous slices covering whole cerebrum.3. Data analysis3.1 Data preprocessing All the raw DTI data was subjected to a quality control (QC) using the DTIPrep tool (http://www.nitrc.org/plugins/mwiki/index.php/dtiprep:MainPage) to identify any artifacts in the diffusion weighted images (DWI), as well as to correct for motion and eddy current artifacts. The datasets were also cropped or embedded into consistent image dimensions. FA value and MD value were calculated for each voxel and than FA images (.nii) and B0 images were saved. Then FA images preprocessing was performed with SPM8 (http://www.fil.ion.ucl.ac.uk/spm) based Matlab R2010b. The FA images were normalized into standard stereotaxic anatomical Montreal Neurological Institute (MNI) space (EPI template). The normalized volumes were resampled to a voxel size of 3mm×3mm×3mm. As a final step, EPI images were spatially smoothed using isotropic Gaussian filter (8mm FWHM).4. Statistical analysisStatistical analysis was performed on SPM8. The file of SPM.mat were entered into SPM8 for random-effects one-sample t-test to determine white matter regions with significant abnormality (P<0.001,uncorrected,K≥20voxel). Regions with statistical significance were masked on FA template.Results1.1 Demography and neuropsychological test were as the same as the ChapterⅠ.1.2 Significant reductions in FA values in control group were found in the splenium of corpus callosum and corticospinal tract.1.3 Significant reductions in FA values were found in the corpus callosum,bilateral anterior part of cingulum and the white matter of bilateral frontal lobes, temporal lobes, occipital lobes, inferior parietal lobules, right precuneus and cuneus and the white matter adjacent to bilateral hippocampi, the triangular part of the bilateral lateral ventricle and the right calcarine sulcus(P<0.001,AlphaSim,K≥6 voxel) in patients with mild AD.1.4 Most of the areas with reduced FA values were not consistent with those with reduced brain activity in DMN in normal aging. While some of the areas with reduced FA values were in concordance with those with reduced brain activity in DMN in patients with mild AD. Conclusion: ROI-based correlation analysis of resting-state fMRI and DTI are two valid techniques for investigating the connectivity of the human brain. Voxel-wise comparison of anisotropy in whole brain between mild AD patients and the control group can reveal wide spread reduction of white matter anisotropy in mild Alzheimer's disease objectively. The fact that the pattern of anisotropic changes of white matter was not inconcordance with that of brain activity in DMN in resting state reveals that white matter anisotropic changes may be outcome from a couple of factors rather than wallerian degeneration solely. Connectivity matrices obtained using both techniques showed significant agreement in DMN, which provided the evidence that the damage of structural connectivity might be the physical foundation in the impairment of functional connectivity. Combination of resting-state correlations and DTI provides a crucial multi-modal validation and reveals the chang of the neuron and white matter, which can be valuable for the early diagnosis of mild AD. |
PDF Full Text Request