Structural magnetic resonance imaging (MRI) studies have shown dramatic age-associated changes in grey and white matter volume, but typically use univariate analyses that do not explicitly test the interrelationship among brain regions. The current study used a multivariate approach to identify covariance patterns of grey and white matter tissue density to distinguish older from younger adults. A second aim was to examine whether the expression of the age-associated covariance topographies is related to performance on cognitive tests affected by normal aging. Eighty-four young (mean age=24.0) and 29 older (mean age=73.1) participants were scanned with a 1.5T MRI machine and assessed with a cognitive battery. Images were spatially normalized and segmented to produce grey and white matter density maps. A multivariate technique, based on the subprofile scaling model, was used to capture sources of between- and within-group variation to produce a linear combination of principal components that represented a "pattern" or "network" that best discriminated between the two age groups. Univariate analyses were also conducted with statistical parametric maps. Grey and white matter covariance patterns were identified that reliably discriminated between the groups with greater than 0.90 sensitivity and specificity. The identified patterns were similar for the univariate and multivariate techniques, and involved widespread regions of the cortex and subcortex. Age and the expression of both patterns were significantly associated with performance on tests of attention, language, memory, and executive functioning. The results suggest that identifiable networks of grey and white matter regions systematically decline with age and that pattern expression is linked to age-related cognitive decline.
Cognitive Neuroscience Division, Taub Institute for Research on Alzheimer's Disease and the Aging Brain, Columbia University Medical Center, New York, NY 20032, USA. email@example.com, , ,