Oscillations, networks, and their development: MEG connectivity changes with age

Abstract

Magnetoencephalographic (MEG) investigations of inter-regional amplitude correlations have yielded new insights into the organization and neurophysiology of resting-state networks (RSNs) first identified using fMRI. Inter-regional MEG amplitude correlations in adult RSNs have been shown to be most prominent in alpha and beta frequency ranges and to express strong congruence with RSN topologies found using fMRI. Despite such advances, little is known about how oscillatory connectivity in RSNs develops throughout childhood and adolescence. This study used a novel fMRI-guided MEG approach to investigate the maturation of resting-state amplitude correlations in physiologically relevant frequency ranges within and among six RSNs in 59 participants, aged 6-34 years. We report age-related increases in inter-regional amplitude correlations that were largest in alpha and beta frequency bands. In contrast to fMRI reports, these changes were observed both within and between the various RSNs analyzed. Our results provide the first evidence of developmental changes in spontaneous neurophysiological connectivity in source-resolved RSNs, which indicate increasing integration within and among intrinsic functional brain networks throughout childhood, adolescence, and early adulthood.

Keywords: alpha-band; beta-band; development; functional connectivity; functional magnetic resonance imaging; magnetoencephalography; neural network; neural oscillations; neural synchrony; resting-state networks.

Figure 1

Network regions for fMRI‐guided MEG connectivity analysis. Regions of interest for 42 coordinates from the six RSNs were adapted from de Pasquale et al. [2012] and are shown on brain surface plots in axial, sagittal left, and sagittal right views. To prevent bias in analysis across a wide age range, seed points were dilated according to each individual's fMRI activity patterns, and coordinates representing maximal local connectivity were selected as seed locations for MEG functional connectivity analysis (see Table 1). Regions are grouped into six RSN, which are denoted by color. Depicted here are fMRI‐guided coordinates averaged across all participants, shown in MNI space. Regions were visualized using BrainNet Viewer software (http://www.nitrc.org/projects/bnv). [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.]

Figure 2

Network connectivity contrasts between age groups. Group contrasts of inter‐ and intra‐ network connectivity for the six RSNs: DAN, VAN, DMN, VIS, MOT, and LAN were determined by averaging all node pairs within one network comparison (i.e., all DMN‐to‐VIS connections). Network connectivity contrasts between age groups are shown for high gamma (80–150 Hz), low gamma (30–80 Hz), beta (15–30 Hz), alpha (8–14 Hz), and theta (4–7 Hz) frequency bands. Connections surviving Bonferroni correction across all network comparisons and frequency bands are shown with asterisks (*α = 0.05; **α = 0.01). Color bar depicts average between‐group differences in inter‐regional correlations. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.]

Figure 3

Spontaneous network connectivity increases throughout development. Associations between network connectivity and age in the beta (15–30 Hz), alpha (8–14 Hz), and theta (4–7 Hz) frequency bands across participants are shown. A: Global connectivity, as calculated through the average connectivity of 42 regions, correlates positively with age. B, C: Increased intra‐network connectivity (i.e., all regions within the DMN) and average inter‐network connectivity (i.e., DMN‐DAN) is associated with increased age. P‐values and R ² –values for linear regression are reported. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.]

Figure 4

Development of oscillatory network connectivity. Correlations of connectivity strength with age for each region, as well as inter‐regional connectivity in the beta (15–30 Hz) and alpha (8–14 Hz) frequency bands. Regions depicted are group averaged fMRI‐guided coordinates from all participants. Region colors illustrate network affiliation. Region size reflects the correlation between age and that regions' connectivity strength with all other areas in the analyzed network. Thus, larger regions denote greater connectivity increases with development. Age‐related increases in connection strength were significant for all regions in the beta and alpha frequency bands (Bonferroni corrected across all regions and frequencies at an alpha level of 0.05). Inter‐regional connections reflect increased connectivity strength that is positively correlated with age. Line thickness represents correlation strength. Only connections surviving Bonferroni correction across all region pairs and frequencies at an alpha‐level of 0.05 are shown. Regions and connections were visualized using BrainNet Viewer software (http://www.nitrc.org/projects/bnv). [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.]

Figure 5

Average alpha band MEG and fMRI connectivity. The strength of inter‐regional connectivity shown for (A) alpha band MEG and (B) fMRI data, averaged across all participants. White cells represent region pairs closer than 35 mm, which were excluded from the analysis. Note adjusted colourbars for MEG and fMRI average connectivity matrices. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.]