Disentangling the effects of age and mild traumatic brain injury on brain network connectivity: A resting state fMRI study

Abstract

Introduction: Cognitive complaints are common shortly after mild traumatic brain injury (mTBI) but may persist up to years. Age-related cognitive decline can worsen these symptoms. However, effects of age on mTBI sequelae have scarcely been investigated.

Methods: Fifty-four mTBI patients (median age: 35 years, range 19-64 years, 67% male) and twenty age- and sex-matched healthy controls were studied using resting state functional magnetic resonance imaging in the sub-acute phase. Independent component analysis was used to identify intrinsic connectivity networks (ICNs). A multivariate approach was adopted to evaluate the effects of age and group on the ICNs in terms of (static) functional network connectivity (FNC), intensities of spatial maps (SMs) and time-course spectral power (TC).

Results: We observed significant age-related changes for a) FNC: changes between 10 pairs of ICNs, mostly involving the default mode (DM) and/or the cognitive-control (CC) domains; b) SMs: intensity decrease in clusters across three domains and intensity increase in clusters across two domains, including the CC but not the DM and c) TC: spectral power decrease within the 0-0.15 Hz range and increase within the 0.20-0.25 Hz range for increasing age within networks located in frontal areas, including the anterior DM. Groups only differed for TC within the 0.065-0.10 Hz range in the cerebellar ICN and no age × group interaction effect was found.

Conclusions: We showed robust effects of age on connectivity between and within ICNs that are associated with cognitive functioning. Differences between mTBI patients and controls were only found for activity in the cerebellar network, increasingly recognized to participate in cognition. Our results suggest that to allow for capturing the true effects related to mTBI and its effects on cognitive functioning, age should be included as a covariate in mTBI studies, in addition to age-matching groups.

Keywords: Aging; Brain connectivity; Resting-state; fMRI; mTBI.

Fig. 1

Spatial maps of the 18 intrinsic connectivity networks identified as belonging to functional domains, thresholded at z-score > 1.

Fig. 2

Results from the multivariate tests showing the significance of the covariates of interest and nuisance predictors for intensity of spatial maps, TC power spectra and FNC. Gray squares represent model terms that were not retained in the backward selection process (α = 0.05).

Fig. 3

Effects of age on (static) functional network connectivity (FNC) between intrinsic connectivity networks (ICNs). (A): FNC matrix showing the pairwise correlations between ICN TCs (FNC) averaged across all participants. Black squares and rectangles highlight ICN pairs for which significant effects for log(Age) were found. (B): The matrix displays the significance and direction of the effects of age for each pairwise correlation (p < 0.05, FDR-corrected). (C, D): Scatterplots depicting how FNC between pairs of ICN TCs changes for increasing age based on two examples from the significant results displayed in (B): (C) is an example of positive age-related correlation (r_AGE) between pairwise FNC and age, based on the pair ICN23 (DM) and ICN19 (SMO). (D) Is an example of negative age-related correlation (r_AGE) between pairwise FNC and age, based on the pair ICN17 (VIS) and ICN13 (CB). The examples selected for the scatterplots (C) and (D) are highlighted in the FNC matrix (B) with asterisks. Age is presented in the scatterplots (C) and (D) on a log-scale.

Fig. 4

Effects of age on SM intensities. (A): Significant effects of age for each ICN SM (p < 0.05, FDR-corrected) in a representative slice. The top panel indicates an example of a cluster with significant SM intensity increase for increasing age (B in the circle) and an example of cluster with significant SM intensity decrease for increasing age (C in the circle). (B;C): Scatterplots depicting how SM intensities change for increasing age. Age is presented in the scatterplots on a log-scale.

Fig. 5

Effects of age on TC power spectra. (A): Effects of age for each IC TC power spectrum; the color bars display their significance and direction (p < 0.05, FDR-corrected). (B): Scatterplots depicting the significant results of changes in TC for increasing age for ICN23 (DM). (C): Line plots of the average TC power spectra of ICNs 5, 15 and 23 for younger and older participants based on median split (median age = 32 years old). The line plots show mean log(power) ± 1SE.

Fig. 6

Effects of Group (mTBI vs. HC). (A): the color bar displays the significance and direction of the effects of group for ICN13 (cerebellar network) in the TC power spectrum (p < 0.05, FDR-corrected). (B): Line plots of the average TC power spectra for mTBI patients (mTBI; red) and healthy controls (HC; blue). C): Line plots of the average TC power spectra for mTBI patients with PTC-present (PTC-present; red), mTBI patients with PTC-abstent (PTC-absent; green) and healthy controls (HC; blue). The line plots show mean log(power) ± 1SE. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)