Computational protocol: Plasma Amyloid Is Associated with White Matter and Subcortical Alterations and Is Modulated by Age and Seasonal Rhythms in Mouse Lemur Primates

Similar protocols

Protocol publication

[…] Cerebral imaging was performed by MRI in all the animals involved in the study. Brain images were recorded on a 7.0 Tesla spectrometer (Agilent, USA) using a four channel phased-array surface coil (Rapid Biomedical, Rimpar, Germany) actively decoupled from the transmitting birdcage probe (Rapid Biomedical, Rimpar, Germany). Briefly, animals were anesthetized by isoflurane (4% for induction and 1–1.5% for maintenance). Respiratory rate was monitored to ensure animal stability until the end of the experiment. Body temperature was maintained by an air heating system. Two-dimensional fast spin echo images were recorded with an isotropic nominal resolution of 230 μm (128 slices, TR/TE = 10,000/17.4 ms; rare factor = 4; field of view = 29.4 × 29.4 mm2, matrix = 128 × 128, slice thickness = 230 μm, acquisition time = 32 min).Images were analyzed by VBM using SPM8 (Wellcome Trust Institute of Neurology, University College London, UK, www.fil.ion.ucl.ac.uk/spm) with the SPMmouse toolbox (http://spmmouse.org) dedicated to animal brain morphometry (Sawiak et al., ). VBM is a reference method, widely used to identify structural changes in the brain including in humans (Whitwell, ).Brain images were segmented into three tissue probability maps (tpm) corresponding to tissues with cortical gray matter (GM), white matter and subcortical nuclei (WM-SC), and cerebrospinal fluid (CSF) characteristics, using locally developed priors. The intensity of the pixel in each probability map represents the probability of the pixel to be GM, WM, or CSF. Then brain images and tpm were spatially transformed to the standard space, defined by Sawiak et al., using a GM mouse lemur template (Sawiak et al., ). Affine regularization was set for an average-sized template, with a bias non-uniformity FWHM cut off of 10 mm, a 5 mm basis function cut off and a sampling distance of 0.3 mm. The resulting GM and WM-SC portions were output in rigid template space, and DARTEL (Ashburner, ) was used to create non-linearly registered maps for each subject and common templates for the cohort of animals. The warped GM and WM-SC portions for each subject were adjusted using the Jacobian determinant from the DARTEL registration fields to preserve tissue amounts (“optimized VBM;” Good et al., ) and smoothed with a Gaussian kernel of 600 μm to produce statistical maps (T maps) for analysis.A first general linear model (GLM) was designed to evaluate relative changes in GM and WM-SC tpm values, a parameter reflecting atrophy, as a function of age. The sex of the animals and total intracranial volumes (TIV) were considered in the design matrix and were treated as covariates of no interest. More specifically, with the GLM, if the brain of one animal is defined by the number “j,” and the location of a pixel is defined as “k.” The signal within a pixel (Yjk) can be explained by the following equationWith β1k = Mean image; β2k = Evolution of the signal according to the age of the animals (n = 21 animals); β5k = Sex effect on signal for males; β6k = Sex effect on signal for females; β7k = effect of TIV on the signal for each animal. In this matrix, xj,1 corresponds to the age of each animal j. Sj,1 and Sj,2 correspond to the sex of the animals. Sj,1 = 1 if the animal j is a male and = 0 otherwise; Sj,2 = 1 if the animal j is a female and = 0 otherwise.A second GLM was designed to evaluate the relationships between GM and WM-SC tpm values and plasma Aβ40 levels. Winter and summer plasma Aβ40 levels were added in the model. For this study, age, sex and TIV were considered in the design matrix and were treated as covariates of no interest. More specifically, with this second GLM, the signal within a pixel (Yjk) can be explained by the following equationWith β1k = Mean image; β2k = Evolution of the signal according to the age of the animals (n = 21 animals); β3k = Evolution of the signal according to winter plasma Aβ40 levels; β4k = Evolution of the signal according to summer plasma Aβ40 levels; β5k = Sex effect on signal for males; β6k = Sex effect on signal for females; β7k = effect of TIV on the signal for each animal. xj,1, xj,2 and xj,3 correspond to the age, winter plasma Aβ levels and summer plasma Aβ levels for each animal j. Sj,1 and Sj,2 correspond to the sex of the animals. Sj,1 = 1 if the animal j is a male and = 0 otherwise; Sj,2 = 1 if the animal j is a female and = 0 otherwise.As an example, on the basis of the GLM-2 model, for the first animal (a 7.8 year-old female, winter and summer plasma Aβ40 levels = 75.2 and 35.5 pg/ml with a TIV of 2150 mm3):for the second animal (a 5.6 year-old male, winter and summer plasma Aβ levels = 37.3 and 33.5 pg/ml with a TIV of 2058 mm3)and so on for the other animals.A contrast defines a linear combination of β as cTβ. For example, the test evaluating the positive relationship between plasma Aβ levels and the probability of pixels being GM is defined using a contrast cTβ = {0 0 1 1 0 0 0]T. The Null hypothesis is H0: cTβ=0, whereas the alternative hypothesis is H1: cTβ>0. This hypothesis is tested with:This analysis allows the removal of confounding effects, such as age (β2k), sex (β5k and β6k), or TIV (β7k) from the raw data. In other words, volumetric scans were entered as the dependent variable. Depending on the tested hypothesis, aging or plasma Aβ levels were the independent variables. Aging, sex, and TIV were covariates.One-tailed t-test contrasts were set up to find areas in which probability values from GM or WM-SC maps correlated with age or plasma Aβ levels. To control for multiple comparisons, an adjusted p-value was calculated using the voxel-wise false discovery rate (FDR-corrected p < 0.05), with extent threshold values of 500 voxels, meaning that clusters required 500 contiguous voxels to be selected as relevant (Genovese et al., ). Voxels with a modulated GM value below 0.2 were not considered for statistical analysis. The operator was blinded to the animal names during image processing. This type of regression technique produces t-statistic and color-coded maps that are the product of a regression model performed at every voxel in the brain. Contiguous groups of voxels that attain statistical significance, called clusters, are displayed on brain images.One of the minor limitations of this study is that brain images were recorded at different time points surrounding the winter and summer blood sampling, so it was not possible to attribute MR images to a given season. However, to our knowledge, modulation of cerebral atrophy by seasonal effects has never been described in mouse lemurs. We thus consider that brain images correspond to the state of an animal the year of the study independently of the season. [...] Paired Student's t-tests were used to evaluate seasonal effects on plasma Aβ40 levels. Pearson's tests were used to evaluate the correlation between plasma Aβ40 concentrations, or amplitude of seasonal variations of Aβ40 and age. Statistical analysis was done using Statistica 7.1 software (StatSoft, Maisons-Alfort, France). P < 0.05 was set as the level of statistical significance for each test. […]

Pipeline specifications

Software tools SPM, Statistica
Applications Miscellaneous, Magnetic resonance imaging