Frontiers in Aging Neuroscience (May 2017)

White Matter Structural Connectivity Is Not Correlated to Cortical Resting-State Functional Connectivity over the Healthy Adult Lifespan

  • Adrian Tsang,
  • Adrian Tsang,
  • Adrian Tsang,
  • Catherine A. Lebel,
  • Catherine A. Lebel,
  • Catherine A. Lebel,
  • Signe L. Bray,
  • Signe L. Bray,
  • Signe L. Bray,
  • Bradley G. Goodyear,
  • Bradley G. Goodyear,
  • Bradley G. Goodyear,
  • Moiz Hafeez,
  • Moiz Hafeez,
  • Moiz Hafeez,
  • Roberto C. Sotero,
  • Cheryl R. McCreary,
  • Cheryl R. McCreary,
  • Cheryl R. McCreary,
  • Richard Frayne,
  • Richard Frayne,
  • Richard Frayne

DOI
https://doi.org/10.3389/fnagi.2017.00144
Journal volume & issue
Vol. 9

Abstract

Read online

Structural connectivity (SC) of white matter (WM) and functional connectivity (FC) of cortical regions undergo changes in normal aging. As WM tracts form the underlying anatomical architecture that connects regions within resting state networks (RSNs), it is intuitive to expect that SC and FC changes with age are correlated. Studies that investigated the relationship between SC and FC in normal aging are rare, and have mainly compared between groups of elderly and younger subjects. The objectives of this work were to investigate linear SC and FC changes across the healthy adult lifespan, and to define relationships between SC and FC measures within seven whole-brain large scale RSNs. Diffusion tensor imaging (DTI) and resting-state functional MRI (rs-fMRI) data were acquired from 177 healthy participants (male/female = 69/108; aged 18–87 years). Forty cortical regions across both hemispheres belonging to seven template-defined RSNs were considered. Mean diffusivity (MD), fractional anisotropy (FA), mean tract length, and number of streamlines derived from DTI data were used as SC measures, delineated using deterministic tractography, within each RSN. Pearson correlation coefficients of rs-fMRI-obtained BOLD signal time courses between cortical regions were used as FC measure. SC demonstrated significant age-related changes in all RSNs (decreased FA, mean tract length, number of streamlines; and increased MD), and significant FC decrease was observed in five out of seven networks. Among the networks that showed both significant age related changes in SC and FC, however, SC was not in general significantly correlated with FC, whether controlling for age or not. The lack of observed relationship between SC and FC suggests that measures derived from DTI data that are commonly used to infer the integrity of WM microstructure are not related to the corresponding changes in FC within RSNs. The possible temporal lag between SC and FC will need to be addressed in future longitudinal studies to better elucidate the links between SC and FC changes in normal aging.

Keywords