1819 This study aimed to evaluate the linear association of age with diffusion tensor imaging (DTI) 20 measures of white matter such as fractional anisotropy (FA), mean diffusivity (MD), axial 21 diffusivity (AD) and radial diffusivity (RD). We assessed patterns of overlap between linear 22 correlations of age with FA with RD, MD and AD to characterize the process of white matter 23 degeneration observed with ageing. 79 healthy adults aged between 18 and 75 took part in the 24 study. The DTI data were based on 61 directions acquired with a b-value of 2000. There was a 25 statistically significant negative linear correlation of age with FA and AD and a positive linear 26 correlation with RD and MD, and AD. The forceps minor tract showed largest percentage of voxels 27 with an association of age with FA, RD and AD, and the anterior thalamic radiation with MD. We 28 found 5 main patterns of overlap: FA alone (15.95%); FA and RD (31.90%); FA and AD (12.99%); 29 FA, RD and AD (27.37%); FA RD, and MD (6.94%). Patterns of overlap between diffusion measures 30 may reflect underlying biological changes with healthy ageing such as loss of myelination, axonal 31 damage, as well as mild microstructural and chronic white matter impairments. This study may 32 provide information about causes of degeneration in specific regions of the brain, and how this 33 may affect healthy brain functioning in older adults. 3 34 Introduction 35 36Diffusion tensor imaging (DTI) is a neuroimaging technique, which allows for non-invasive, 37 in vivo, investigation of white matter [1][2][3]. DTI measures are based on random motion of water 38 molecules, where within the brain diffusion of water is less restricted, or more isotropic, in areas 39 of grey matter and CSF, and more restricted, or more anisotropic, in areas of white matter. When 40 white matter structural architecture deteriorates, water molecules within white matter tissue 41 become more isotropic, making DTI a useful tool for assessing atrophy [4][5][6][7][8][9]. 43The diffusion tensor is a 3x3 covariance matrix used to model diffusion within a voxel, in 44 which there are 3 positive eigenvalues (λ1, λ2, λ3) and 3 orthogonal eigenvectors (ε1, ε2, ε3). The 45 eigenvalues of the tensor give the diffusivity in the direction of each eigenvector. Together they 46 describe diffusion probability using an ellipsoid, where the axes of the ellipsoid are aligned with 47 the eigenvectors, and the major eigenvector (λ1) represents the principal diffusion direction. 49Fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD) and axial 50 diffusivity (AD) are the four main diffusion-based measurements of white matter structural 51 architecture. FA measures the amount of diffusion asymmetry within a voxel, where a value of 0 52 is isotropic and is represented by a spherical ellipsoid with equal eigenvalues, and a value of 1 is 53 anisotropic and is represented by an elongated ellipsoid with unequal eigenvalues. FA has been 54 associated with the microstructural integrity of white m...

Adolescence is marked by changes in cognitive abilities and in several MRI-based measures of brain structure. This study took an individual-differences approach to help understand adolescent cognitive development in a large-sample longitudinal cohort, the IMAGEN study (initial n = 2,316). We used a latent change score model to assess the associations between levels and changes in the brain’s grey-matter regions and latent general cognitive ability between ages 14 and 19 years. As expected, higher cognitive ability was correlated with higher cortical volume and larger surface area, with more ambiguous results for cortical thickness. Higher-ability participants at age 14 tended to have accelerated subsequent cortical thinning, as well as cortical volume loss. There was no statistically significant link between changes in cognitive ability and changes in the brain measures we used. We also attempted to predict levels and changes in the brain and in cognitive ability using a polygenic score for genetic variants linked to educational attainment: the score was modestly associated with the baseline measures, but did not predict the trajectory of change in any measure to a statistically significant degree. Age-14 cortical volume and surface area—though not cortical thickness—mediated a portion (9-10%) of the association between the polygenic score and age-19 cognitive ability. These findings demonstrate how large-sample data can shed light on the links between brain and cognitive ability in this important phase of the lifespan.