Cerebral small vessel disease (SVD), including white matter hyperintensities (WMH), lacunes and microbleeds, and brain atrophy, are related to cognitive impairment. However, these magnetic resonance imaging (MRI) markers for SVD do not account for all the clinical variances observed in subjects with SVD. Here, we investigated the relation between conventional MRI markers for SVD, network efficiency and cognitive performance in 436 nondemented elderly with cerebral SVD. We computed a weighted structural connectivity network from the diffusion tensor imaging and deterministic streamlining. We found that SVD-severity (indicated by higher WMH load, number of lacunes and microbleeds, and lower total brain volume) was related to networks with lower density, connection strengths, and network efficiency, and to lower scores on cognitive performance. In multiple regressions models, network efficiency remained significantly associated with cognitive index and psychomotor speed, independent of MRI markers for SVD and mediated the associations between these markers and cognition. This study provides evidence that network (in)efficiency might drive the association between SVD and cognitive performance. This highlights the importance of network analysis in our understanding of SVD-related cognitive impairment in addition to conventional MRI markers for SVD and might provide an useful tool as disease marker.
Keywords: cerebral small vessel disease; cognition; graph-theory; structural brain networks.
© 2015 Wiley Periodicals, Inc.