White matter atrophy and lesion formation explain the loss of structural integrity of white matter in aging
Marius de Groot,ab Meike W Vernooij,cd Aad van der Lugt,c M Arfan Ikram,d Gabriel P Krestin,c Albert Hofman,d Monique MB Bretelerd and Wiro J Niessenab
Biomedical Imaging Group Rotterdam, Departments of Radiology & Medical Informatics; c Department of Radiology; d Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam. b Imaging Science & Technology, Faculty of Applied Sciences, Delft University of Technology, Delft.
Introduction The importance of macrostructural white matter changes, including white matter lesions and atrophy, in intact brain functioning is increasingly being recognized. Diffusion tensor imaging (DTI) enables measurement of the microstructural integrity of white matter. Loss of white matter integrity in aging has been reported, but whether this is inherent to the aging process itself or results from specific white matter pathology is unknown. Using DTI in persons aged 60 years and older, we studied whether and how white matter atrophy and white matter lesions related to microstructural integrity in normalappearing white matter and whether this was independent from age.
What is ...? White matter atrophy
low Materials and methods In 832 persons (mean age 67.3 years) from the population-based Rotterdam Study, we derived fractional anisotropy (FA) measurements from DTI scans. FA values of the central voxels in the major white matter tracts of all subjects were projected onto a common white matter skeleton using Tract Based Spatial Statistics , an automated robust mapping technique that prepares data for direct voxelwise comparison. Separately, automated tissue segmentation  identified voxels originating in white matter lesions. These voxels were excluded from the analysis to allow a specific focus on the normalappearing white matter. Statistical analysis was performed with multiple linear regressions, adjusting for age, sex and white matter atrophy or white matter lesion load and corrected for multiple comparisons.
figure legend FA images procedure(s)
White matter atrophy is expressed in a scalar value per subject. White matter lesion load
White matter lesion load is expressed in a scalar value per subject.
image(s) Crosswise registration
Determine most typical subject
Create mean FA image
With increasing age, multiple regions showed significant decreases in FA in normal-appearing white matter. However, nearly all of these regional decreases in FA were explained by either white matter atrophy or by white matter lesions as depicted below and in . Both processes were found to be related to distinct brain regions: white matter atrophy was primarily related to FA decrease in the body of the corpus callosum and in the limbic system (cingulate tract and hippocampal region), whilst white matter lesion load related mainly to periventricular integrity loss.
Direction info Map subj. FA on skeleton search ⊥ to track
832 FA datasets mean FA image FA skeleton 832 corresp. skeletons
(per skeleton voxel)
FA per subject
model parameter, e.g. Age
multiple linear regression, per voxel
The white matter skeleton (black) is projected onto the axial MR images. Yellow-to-red colors represent normal-appearing white matter regions with reduced FA in relation to (a) increasing age, adjusted for sex only, (b) global white matter atrophy, adjusted for age, sex and white matter lesions, (c) white matter lesions, adjusted for age, sex and white matter atrophy and (d) increasing age, adjusted for sex, white matter atrophy and white matter lesions.
Conclusions Loss of white matter integrity in aging is primarily explained by atrophy and lesion formation and not by the aging process itself. Furthermore, white matter atrophy and white matter lesion formation relate to loss of integrity in distinct brain regions, indicating the two processes are pathophysiologically different.
 Smith, S.M., M. Jenkinson, H. Johansen-Berg, et al., Tract-based spatial statistics: voxelwise analysis of multisubject diffusion data. Neuroimage, 2006. 31(4): p. 1487-505.  Vrooman, H.A., C.A. Cocosco, F. van der Lijn, et al., Multi-spectral brain tissue segmentation using automatically trained k-Nearest-Neighbor classification. Neuroimage, 2007. 37(1): p. 71-81.  Vernooij, M.W., M. de Groot, A. van der Lugt, et al., White matter atrophy and lesion formation explain the loss of structural integrity of white matter in aging. Neuroimage, in press.