Voxel-based analysis derived from fractional anisotropy images of white matter volume changes with aging Elisabetta Pagani, a Federica Agosta, a Maria A. Rocca, a Domenico Caputo, b and Massimo Filippi a, ⁎ a Neuroimaging Research Unit, Scientific Institute and University Ospedale San Raffaele, Milan, Italy b Department of Neurology, Scientific Institute Fondazione Don Gnocchi, Milan, Italy Received 12 December 2007; revised 26 February 2008; accepted 15 March 2008 Available online 26 March 2008 Although age-related effects on brain volume have been extensively investigated post mortem and in vivo using magnetic resonance imaging (MRI), regional and temporal patterns of white matter (WM) volume changes with aging are not defined yet. The aim of this study was to assess the topographical distribution of age-related WM volume changes using a recently developed voxel-based method to obtain estimates of WM fiber bundle volumes using diffusion tensor (DT) MRI. Brain conventional and DT MRI were obtained from 84 healthy subjects (mean age = 44 years, range = 13–70). Linear and non-linear relationships between age and WM fiber bundle volume changes were tested. A negative linear correlation was found between age and WM volume decline in the corona radiata, anterior cingulum, body and crus of the fornix and left superior cerebellar peduncle. A positive linear correlation was found between age and volume increase of the right deep temporal association fibers. The non-linear regression analysis also showed age-related changes of the genu of the corpus callosum and fitted better the volume changes of the right deep temporal association fibers. WM volume decline with age is unevenly distributed across brain regions. Our approach holds promise to gain additional information on the pathological changes associated to neurological disorders of the elderly. © 2008 Elsevier Inc. All rights reserved. Introduction Age-related effects on the volume of the human brain tissues have been extensively studied both at post mortem (Rees, 1976; Meier-Ruge et al., 1992; Kemper, 1994; Aboitiz et al., 1996; Pakkenberg and Gundersen, 1997; Tang et al., 1997; Marner et al., 2003) and in vivo using magnetic resonance imaging (MRI) (Sowell et al., 2004; Raz and Rodrigue, 2006). The most shared hypothesis is that grey matter (GM) volume declines linearly with age (Sowell et al., 2004; Raz and Rodrigue, 2006), while white matter (WM) volume essentially remains steady or increases slowly through adulthood, peaking at the 40–50 year range (Courchesne et al., 2000; Bartzokis et al., 2001, 2004; Jernigan et al., 2001; Ge et al., 2002; Allen et al., 2005; Fotenos et al., 2005; Walhovd et al., 2005), followed by a precipitous decline starting around 60 years of age (Guttmann et al., 1998; Salat et al., 1999; Courchesne et al., 2000; Bartzokis et al., 2001, 2004; Jernigan et al., 2001; Ge et al., 2002; Liu et al., 2003; Allen et al., 2005; Fotenos et al., 2005; Walhovd et al., 2005). The topographic patterns of age-related GM decline have been investigated using both global and regional MR-based approaches (Sowell et al., 2004; Raz and Rodrigue, 2006), whereas the actual topographic distribution of WM changes with aging is still controversial (Raz and Rodrigue, 2006). The remarkable hetero- geneity between studies regarding WM volume changes with aging might be due to at least three reasons. First, these studies differ in the sample size and age ranges studied. In this context, the inclusion of adolescents should serve to clarify the impact of ongoing progressive volume changes that can be thought of as continuous with brain maturational effects (Sowell et al., 1999, 2002, 2004). This is particularly important for those WM fiber bundles that post mortem (Yakovlev and Lecours, 1967; Benes et al., 1994; Kemper, 1994) and in vivo MRI (Jernigan et al., 1991; Pfefferbaum et al., 1994; Reiss et al., 1996; Giedd et al., 1999; Courchesne et al., 2000; Bartzokis et al., 2001, 2004; Sowell et al., 2002) studies have shown to progressively increase in size throughout childhood and into young adulthood. Second, the lack of a correlation between subjects’ age and whole WM volume might relate to an uneven distribution of WM loss across different brain regions (Salat et al., 1999; Bartzokis et al., 2001, 2004; Jernigan et al., 2001; Allen et al., 2005; Lemaitre et al., 2005; Walhovd et al., 2005; Abe et al., 2008; Brickman et al., 2007; Smith et al., 2007). As a consequence, these regional changes might go undetected when using a global approach (Good et al., www.elsevier.com/locate/ynimg NeuroImage 41 (2008) 657 – 667 ⁎ Corresponding author. Neuroimaging Research Unit, Department of Neurology, Scientific Institute and University Ospedale San Raffaele, Via Olgettina, 60, 20132 Milan, Italy. Fax: +39 02 2643 3054. E-mail address: massimo.filippi@hsr.it (M. Filippi). Available online on ScienceDirect (www.sciencedirect.com). 1053-8119/$ - see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2008.03.021