Structural MRI covariance patterns associated with normal aging and neuropsychological functioning☆
Introduction
Voxel based morphometric (VBM) approaches to the analysis of structural magnetic resonance images allow for between- and within-groups comparison of grey and white matter volume or density [6]. Because of its high reliability and semi-automated procedures, VBM is well suited for large-scale cross-sectional and longitudinal studies that examine normal age-related neuromorphologic change. In a typical neuromorphologic study of aging, structural magnetic resonance images (MRI) are acquired, spatially normalized to common stereotactic coordinates, and segmented into grey matter, white matter, and cerebrospinal fluid (CSF). Comparisons between groups (i.e., in cross-sectional aging studies) or within groups (i.e., in longitudinal analyses) are made on the segmented images on a voxel-by-voxel basis. Statistical parametric maps (SPM) are generated that reflect differences between or within groups (or the relationship with a continuous variable) in each individual voxel.
Although several studies have used manual and semi-manual tracing techniques for the assessment of brain volume change across the adult lifespan (e.g. [28]), relatively few have employed VBM approaches. Those studies that have used VBM suggest dramatic normal age-associated changes in both grey and white matter volume. In the largest study to date, Good et al. [21] examined grey matter, white matter, and CSF volumes and concentrations in a group of 465 neurologically healthy adults ranging in age from 18 to 79 years. In addition to regional effects, the authors examined global effects by summing grey matter, white matter, and CSF voxel values and tested linear and non-linear contrasts with age. The authors found a linear decline in global grey matter volume with age and several focal areas of relatively greater age-associated loss. These regions included the superior parietal gyri, pre- and post-central gyri, insula, cerebellum, anterior cingulate, middle frontal gyrus, Heschl's gyrus, and planum temporale. Though global white matter volume was not significantly associated with age, there were regional effects of age, including bilateral frontal lobe and posterior limbs of the internal capsule.
Other VBM investigations of age-related change have been consistent in demonstrating a tendency for grey matter tissue loss, although they have varied in results regarding white matter and the regional distribution of change. In a longitudinal analysis, Resnick et al. [41] examined grey and white matter change in healthy older adults ranging in age from 59 to 85 and found wide-spread areas of decline for both tissue types. Total grey and white matter volume loss was approximately 5.4 and 2.4 cm3, respectively, with the greatest regional effects in frontal and parietal lobes. Similar to Good and colleagues’ findings [21], other cross-sectional analyses of subjects across the adult lifespan revealed a global decline in grey matter density and non-significant age associations for white matter [50], [52]. Tisserand et al. [51] reported a significant age-associated decline in grey matter volume with VBM, but they did not evaluate white matter and only analyzed anterior regions.
Although voxel-based approaches to the study of normal aging have generally been consistent with manual tracing or region-of-interest approaches in demonstrating age-associated grey matter decreases [9], [29], [36], [37], evidence from both human neuroimaging studies and animal studies using other approaches suggests that normal age-related volumetric loss may be primarily due to a decrease in whiter matter[10], [18], [24], [41]. Age-associated changes in white matter volume may be reflective of a disturbance in myelin, leading to faulty communication across axons [2]. This change may have particular relevance to cognitive changes associated with normal aging, which may in part be due to decreased communication among critical brain regions. Nonetheless, whether white matter significantly decreases more than grey matter with normal aging has not been demonstrated definitively. Further, the relative impact of grey versus white matter volume on cognitive function requires the comparison of each to cognitive abilities.
Despite the statistical analysis of all of the voxels in the brain, VBM approaches do not consider the interrelationship among voxel densities or volumes. That is, the univariate SPM generated for a VBM analysis provides results for individual regions and does not address structural connectivity in the brain. For example, although two regionally separate voxels may show aging effects, their relationship to each other is not explicitly examined. In the current study, we employed multivariate spatial covariance methodology to deal with this issue.
An attractive feature of multivariate techniques is their emphasis on structural networks. Because they evaluate the inter-correlation or covariance of tissue density across voxels, the results yielded from the analyses can be interpreted as representational of structural networks throughout the brain. In the current study, we employed the subprofile scaling model (SSM) analytic approach [34] to capture the patterns of age-associated changes in tissue density. This method tests whether, across subjects, tissue densities in individual voxels show a covarying relationship with age as well as amongst each other. Subtle changes in tissue density in the course of aging that also show interregional correlation will be detected by our method, while they might go undetected in a univariate analysis. Different rates of age-related change at different voxel location are specified by regional weights of both signs (for increasing and decreasing density with age) in a covariance pattern. Identification of such an age-related covariance pattern might be a useful signature of the aging process in the healthy brain. The extent to which each individual subject manifests the age-related topographic pattern can be captured in a single number, similar to a factor score in a factor analysis. The SSM approach has been used in a number of functional neuroimaging studies (e.g. [11], [25], [26], [43], [44], [48], [49]), but no study to our knowledge has employed the technique to structural neuroimaging data. Thus, a primary purpose of the current study was to examine whether SSM could identify a covariance pattern of age-related grey and white matter tissue change and to qualitatively compare this approach with a standard univariate VBM analysis of the same dataset.
A secondary aim of the current study was to examine the relationship between the expression of the identified SSM age-associated covariance pattern and performance on a brief battery of neuropsychological tests. Normal age-associated neurocognitive decline has been well documented (e.g. [42], [45]). Although no unified theory of normal cognitive aging has emerged from the extant literature, there is a growing consensus that the domains of executive functioning and memory are disproportionately affected [15], [56]. Despite the growing literature on normal age-associated morphologic changes and the parallel line of research on normal cognitive aging, relatively few studies have investigated the interrelationship among age, neurocognition, and brain morphology (e.g. [23], [38], [52], [54]). These efforts have generally demonstrated a relationship between age-associated decline in regional brain volume and performance on tasks of executive functioning and memory, although age typically accounts for more variance in task performance than morphology. In the current study, we sought to identify a grey and white matter covariance patterns that captured the effects of age on neuropsychological functioning.
Section snippets
Subjects
Data for the current study came from ongoing neuroimaging studies of normal aging. Subject groups included 113 participants, comprising younger (n = 84, mean age ± S.D. = 24.02 ± 3.83, range = 19–35) and older (n = 29, mean age ± S.D. = 73.14 ± 6.72, range = 60–84) individuals. Participants were recruited through local advertisement and word-of-mouth. All subjects were screened with medical, neurological, psychiatric, and neuropsychological evaluations to ensure that they had no neurological or psychiatric disease
Covariance patterns
Grey and white matter patterns that discriminated between younger and older subjects are displayed in Fig. 1, Fig. 2, respectively. The optimal number of principal components, based on the Akaike criteria [16], was 7 for grey matter and 6 for white matter. Talairach coordinates and their corresponding regional labels for grey matter areas involved in the covariance pattern are displayed in Table 1. Negative factor loadings, indicating collateral age-associated decreases in density, were
Discussion
In the current study, grey and white matter density was compared between younger and older neurologically healthy participants with a standard univariate VBM approach [6] and a multivariate SSM approach [33], [34]. Consistent with previous studies of normal age-related changes in neuromorphology, we found age-associated reductions of both tissue types using both analytic techniques. Furthermore, the expression of the identified covariance pattern in each participant was significantly related to
Acknowledgements
This work was supported by National Institute on Aging grants AG024708 (AMB) and AG261858 (YS).
References (56)
- et al.
The inferential impact of global signal covariates in functional neuroimaging analyses
Neuroimage
(1998) Neuropsychological and neurophysiological changes in healthy adult humans across the age range
Neurobiol Aging
(1993)- 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.
Why voxel-based morphometry should be used
Neuroimage
(2001) Age-related myelin breakdown: a developmental model of cognitive decline and Alzheimer's disease
Neurobiol Aging
(2004)Voxel-based morphometry should not be used with imperfectly registered images
Neuroimage
(2001)Memory and executive function in aging and AD: multiple factors that cause decline and reserve factors that compensate
Neuron
(2004)- et al.
Voxel-based morphometry versus region of interest: a comparison of two methods for analyzing gray matter differences in schizophrenia
Schizophr Res
(2005) - et al.
A voxel-based morphometric study of ageing in 465 normal adult human brains
Neuroimage
(2001)
Automatic differentiation of anatomical patterns in the human brain: validation with studies of degenerative dementias
Neuroimage
Neuroanatomical correlates of selected executive functions in middle-aged and older adults: a prospective MRI study
Neuropsychologia
Relation of cognitive reserve and task performance to expression of regional covariance networks in an event-related fMRI study of nonverbal memory
Neuroimage
An event-related fMRI study of the neural networks underlying the encoding, maintenance, and retrieval phase in a delayed-match-to-sample task
Brain Res Cogn Brain Res
Hippocampal head size associated with verbal memory performance in nondemented elderly
Neuroimage
Effects of age on tissues and regions of the cerebrum and cerebellum
Neurobiol Aging
Covariance PET patterns in early Alzheimer's disease and subjects with cognitive impairment but no dementia: utility in group discrimination and correlations with functional performance
Neuroimage
Voxel-based morphometry of human brain with age and cerebrovascular risk factors
Neurobiol Aging
Regional frontal cortical volumes decrease differentially in aging: an MRI study to compare volumetric approaches and voxel-based morphometry
Neuroimage
Thalamic volume predicts performance on tests of cognitive speed and decreases in healthy aging. A magnetic resonance imaging-based volumetric analysis
Brain Res Cogn Brain Res
Memory and executive function in older adults: relationships with temporal and prefrontal gray matter volumes and white matter hyperintensities
Neuropsychologia
Effects of age on volumes of cortex, white matter and subcortical structures
Neurobiol Aging
Individual differences in PET activation of object perception and attention systems predict face matching accuracy
Neuroreport
Diagnostic and Statistical Manual of Mental Disorders
Age-related changes in frontal and temporal lobe volumes in men: a magnetic resonance imaging study
Arch Gen Psychiatr
White matter structural integrity in healthy aging adults and patients with Alzheimer disease: a magnetic resonance imaging study
Arch Neurol
Identification and differential vulnerability of a neural network in sleep deprivation
Cereb Cortex
Multilingual aphasia examination
Cited by (115)
Age- and sex-related differences in baboon (Papio anubis) gray matter covariation
2023, Neurobiology of AgingLongitudinal changes in grey matter and cognitive performance over four years of healthy aging
2022, Neuroimage: ReportsBrain patterns of pace – but not rhythm – are associated with vascular disease in older adults
2022, Cerebral Circulation - Cognition and BehaviorGrey-matter networks in aging
2021, Factors Affecting Neurological Aging: Genetics, Neurology, Behavior, and Diet
- ☆
This research was reviewed and approved by the institutional ethics committee (Institutional Review Board) and written informed consent was obtained from all participants.