Gender differences and age-related white matter changes of the human brain: A diffusion tensor imaging study
Introduction
The normal adult brain undergoes considerable morphological changes as it ages. Both a decrease in brain weight and an increase in cerebrospinal fluid (CSF) filled spaces, caused by ventricle dilation and sulcal enlargement, have been documented by autopsy studies (Davis and Wright, 1997, Dekaban and Sadowsky, 1978). Differences between white and gray matter loss in the aging brain have also been reported. Neuropathological studies have demonstrated that with advancing age, neuronal loss in the cortex is either not significant or not as extensive as previously reported, which suggests that white matter (WM) changes may exceed gray matter changes in normal human aging (e.g., Miller et al., 1980). Postmortem studies of normal subjects have also revealed disturbances of WM microstructure, such as a decline in amount of myelinated fibers of the precentral gyrus and the corpus callosum, or even axonal loss (Meier-Ruge et al., 1992, Kemper, 1994). However, postmortem analyses of brain structures depend on many factors, such as the interval between death and fixation and the timing of measurements. As a consequence, it might be difficult to correlate the results from postmortem studies to living subjects.
Neuroimaging technology like magnetic resonance imaging (MRI) has greatly enhanced the ability to perform studies on living human subjects. In this context, numerous researchers have applied morphometric MRI techniques in normal aging studies. However, some results of these studies, such as the volume change of the cerebral WM, are not consistent. Some investigations observed no significant global volume decrease in cerebral WM (Jernigan et al., 2001, Good et al., 2001, Smith et al., 2007), whereas others did show such a significant volume change, either fitted by a linear regression model (Lemaitre et al., 2005) or even higher order regression models Liu et al., 2003, Allen et al., 2005). Gender effects on global WM are also unclear, i.e., both significant (Lemaitre et al., 2005) and nonsignificant interactions (Smith et al., 2007) have been reported. Notably, the measured WM volumes, as determined from conventional MRI, only reflect the late stage of macrostructure change and may not be sensitive to the age-related degeneration of myelin and axons in the WM microstructure. Such microstructural changes are beyond the detection of conventional MRI, but are within the reach of diffusion tensor MRI (DTI) methodology.
DTI is a powerful technique for assessing the WM axonal organization – and on a more macroscopic level the 3D fiber architecture – which yields several quantitative measures, such as the three principal diffusivities (the eigenvalues of the diffusion tensor: λ1, λ2, λ3), the mean diffusivity (MD), and the degree of anisotropy (e.g., the fractional anisotropy: FA). Combined with the directional diffusion information, as derived from the eigenvectors of the diffusion tensor, these measures reflect the fiber cohesiveness of WM fiber tracts (Pierpaoli and Basser, 1996). To date, DTI studies of normal aging have identified a decline of FA and an increase of MD in several WM regions (Pfefferbaum et al., 2000, Sullivan et al., 2001, Abe et al., 2002, Salat et al., 2005a, Salat et al., 2005b). These regions include the genu and splenium of the corpus callosum, the centrum semiovale, the prefrontal WM, the posterior limb of the internal capsule, and the posterior periventricular region. However, different results in such FA alterations – based on predefined regions of interest (ROIs) – have been revealed (Pfefferbaum et al., 2000, Nusbaum et al., 2001, Sullivan et al., 2001, Salat et al., 2005a, Salat et al., 2005b). This variability may be due to the user-dependent placement of the ROIs, the number of ROIs simultaneously examined in these studies, and the difference in sample size.
In many cases, particularly in aging or neuropsychiatric disorders, the spatial location and extent of the FA changes are not known a priori. In this context, in recent years, researchers have been interested in applying global search strategies in DTI studies, such as whole-brain voxel-wise analyses (Ashburner and Friston, 2000). In this approach, each subject's data set (here, the FA image) is first registered into a standard space, in which voxel-wise statistics are calculated to find areas that correlate to a covariate of interest. In this way, both location and extent of statistically significant changes can be determined without any hypothesis a priori. However, there are two main drawbacks to such a multi-subject voxel-based DTI study. Firstly, the accuracy of matching the WM fiber structures to the standard space from each individual's FA image is very difficult to determine (Smith et al., 2006). Consequently, this raises the question of whether individual topographical variability can be taken into account. The other problem relates to the standard practice of spatially smoothing the data before computing the voxel-wise statistics. The amount of smoothing can greatly affect the final results, but there is no uniform principle for choosing an appropriate kernel size (Jones et al., 2005). Furthermore, even with a moderate amount of smoothing, a large number of voxels within the central WM regions still had nonnormally distributed residuals, thus making valid statistical inferences with a parametric approach problematic in these areas (Jones et al., 2005). One possible way to solve this problem is using nonparametric methods for inference on FA statistics. In this way, no prior information of data and noise distributions are required to perform statistical analyses (Nichols and Holmes, 2002). Furthermore, such a nonparametric approach is robust with respect to outliers that might be present due to an imperfect coregistration result. Another way is using the ‘tract-based spatial statistics’ (TBSS) approach on a voxel-wise or cluster-based level to perform the statistical analyses (Smith et al., 2006). However, although the TBSS method might reduce the amount of false positive effects (as caused by misalignment errors), spatial extent and location of significant regions are limited to the ‘skeleton’ of the brain WM, which potentially introduces an increased amount of false negative effects.
We investigated whether the MD and FA of the diffusion tensor, measured in a large number of normal subjects, are correlated with age taking potential gender differences into account. This study of normal aging combined with the exploration of gender differences can provide the baseline characteristics of MD and FA changes in normal healthy subjects across the age range of older adults, which is prerequisite for interpreting DTI data of individuals with neurological disorders or other conditions affecting the brain. Basically, the analysis in this work is performed in two steps. Firstly, we apply an image coregistration method to the FA images, and secondly, we use a permutation/randomization method to search for brain voxels that are correlated with age and gender (Nichols and Holmes, 2002). Spatial smoothing of images for whole-brain analyses remains a debatable issue (Jones et al., 2005). Therefore, in this analysis, we examined the combined age-related effect of and gender difference in diffusion MR properties in spatially nonsmoothed data sets using robust statistics.
Section snippets
Participants
The subjects in this study were recruited from a health screening program in the Shin-Kong Wu Ho-Su Memorial Hospital, Taipei, Taiwan. All participants received detailed health examinations, including physical and neurological examinations, a biochemistry study, a chest X-ray, an electrocardiogram, and an electroencephalogram. Participants were excluded if they had a history of major neurological, psychiatric, or serious cardiovascular diseases. Diffusion tensor images were obtained from 145
Mean global MD and FA analysis
Scatter plots for the association between the mean global MD, FA, and age are depicted in Fig. 2. The mean global FA change showed a significant negative correlation with age (r = − 0.44, p < 0.0001). We also studied the quadratic and cubic regression model for the mean global FA as a function of age. The results showed that the linear regression model was the best model fit (i.e., only the first-degree term in the quadratic and cubic regression models showed a significant effect). Therefore, the
Global effect of age
In this study, we applied an image coregistration procedure to study the effect of age on FA, MD, and the eigenvalues in an objective and automated way, which was in good agreement with the complementary performed ROI-based analysis. The mean global FA of the normalized images showed a significant negative association with age, which has also been observed in previous studies (Salat et al., 2005a, Salat et al., 2005b, Abe et al., in press). The mean global MD showed a significantly gender
References (45)
- et al.
Normal aging in the central nervous system: quantitative MR diffusion-tensor analysis
Neurobiol. Aging
(2002) - et al.
Normal neuroanatomical variation due to age: the major lobes and a parcellation of the temporal region
Neurobiol. Aging
(2005) - et al.
Voxel-based morphometry—the methods
NeuroImage
(2000) - et al.
Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI
J. Magn. Reson., B
(1996) - et al.
Power and sample size calculations for studies involving linear regression
Control. Clin. Trials
(1998) - et al.
A voxel-based morphometric study of aging in 465 normal adult human brains
NeuroImage
(2001) - et al.
Effects of age on tissues and regions of the cerebrum and cerebellum
Neurobiol. Aging
(2001) - et al.
The effect of filter size on VBM analyses of DT-MRI data
NeuroImage
(2005) - et al.
Age- and sex-related effects on the neuroanatomy of healthy elderly
NeuroImage
(2005) - et al.
A probabilistic atlas of the human brain: theory and rationale for its development. The International Consortium for Brain Mapping (ICBM)
NeuroImage
(1995)
Tractography-guided statistics (TGIS) in diffusion tensor imaging for the detection of gender difference of fiber integrity in the midsagittal and parasagittal corpora callosa
NeuroImage
Age-related alterations in white matter microstructure measured by diffusion tensor imaging
Neurobiol. Aging
Tract-based spatial statistics: voxel-wise analysis of multi-subject diffusion data
NeuroImage
Age and gender effects on human brain anatomy: a voxel-based morphometric study in healthy elderly
Neurobiol. Aging
Diffusion tensor imaging detects early Wallerian degeneration of the pyramidal tract after ischemic stroke
NeuroImage
Effects of handedness and gender on macro- and microstructure of the corpus callosum and its subregions: a combined high-resolution and diffusion-tensor MRI study
Brain Res. Cogn. Brain Res.
Topography of the human corpus callosum using diffusion tensor tractography
J. Comput. Assist. Tomogr.
Spatial transformations of diffusion tensor magnetic resonance images
IEEE Trans. Med. Imag.
Diffusion anisotropy in subcortical white matter and cortical gray matter: changes with aging and the role of CSF-suppression
J. Magn. Reson. Imaging
Myelination of a key relay zone in the hippocampal formation occurs in the human brain during childhood, adolescence, and adulthood
Arch. Gen. Psychiatry
A new method for measuring cranial cavity volume and its application to the assessment of cerebral atrophy at autopsy
Neuropathol. Appl. Neurobiol.
Cited by (263)
Associations between glutamic acid decarboxylase antibodies, oxidative stress markers, and cognitive capacity in adolescents who stutter
2023, Saudi Journal of Biological SciencesWhite matter microstructure associations to amyloid burden in adults with Down syndrome
2022, NeuroImage: Clinical