In this study, we examined fMRI-ICNs in a sample of OA-mTBI and age-matched OA-HCs. We employed an ICA-based multivariate approach to identify the effects of mTBI on brain network connectivity, severity of complaints and their interaction at older age.
We identified three out of 15 ICNs (i.e. ICN13, 14 and 15) with high activations in more than one functional domain. Additionally, cerebellar activations were found in several networks, including but not limited to the visual one. It is known that functional networks become more integrated with increasing age (Geerligs et al. 2014). Kundu et al. (2018) demonstrated that the number of ICNs decreases exponentially with aging and identified the connectivity pattern of the cerebellar components as the most strongly affected (by aging), suggesting a potential pivotal role for the cerebellum in functional network reorganization with age. Our results add evidence that, at older age, networks of several domains are more strongly integrated, and the cerebellar network is distributed over networks; the latter being likely the reason for not finding a separate cerebellar network.
In a previous study, we found that age affects both within- and between-network connectivity during mid-adulthood, whereas effects of mTBI were much smaller (Bittencourt-Villalpando et al. 2021). Here, we identified effects of mTBI on network connectivity within visual(-cerebellar) and cognitive-language networks, suggesting that these regions are the most affected in older adults that have sustained an mTBI. Additionally, we identified no effect of age on network connectivity, concurring with the notion that changes in functional connectivity occur at a slower pace at older age.
We found increased within-network connectivity for OA-mTBI in clusters located in the left middle temporal gyrus (lMTG) of the cognitive-language ICN. The lMTG is involved in language processing, which requires a complex integration of sensory inputs (e.g. visual and auditory) (Fridriksson et al. 2016; Hickok and Poeppel 2007). Hyperconnectivity is a common finding after TBI, but its adaptive or maladaptive nature remains unclear (Hillary and Grafman 2017; Morelli et al. 2021). Previous mTBI studies using task-fMRI in younger cohorts found that hyperconnectivity in functional networks involved in attentional and visual processes might be associated with increased subjective effort and task-related fatigue (Prak et al. 2021). In our cohort, fatigue (together with dizziness) was the most prevalent complaint after mTBI, being present in 60% of the OA-mTBI (see Appendix, Fig. S1). Perhaps increased awareness of external sensory stimuli, could partly explain increased fatigue that is commonly experienced in elderly after mTBI.
Additionally, we found decreased within-network connectivity for OA-mTBI in a cluster located posteriorly in the right fusiform gyrus (rFG) of a visual-cerebellar ICN. The rFG, particularly, is known for its role in visual pattern recognition (Grill-Spector et al. 2006), including facial recognition (Kanwisher et al. 1997). Our findings are consistent with previous studies that identified hypoactivity within visual networks of patients in the (sub)acute phase after mTBI (Amir et al. 2021; Stevens et al. 2012). Moreover, it has been suggested that the subacute phase after mTBI is marked by hypoactivation in several (non-DMN) areas across the brain, which normalizes in those patients who show good recovery (Bharath et al. 2015; Rosenthal et al. 2018). We encourage future research to investigate the longitudinal effects of mTBI on brain connectivity in the visual cortex.
In this study, we found interactions between the severity of complaints and group (OA-HC and OA-mTBI) for within-network connectivity. In clusters anteriorly located in the fusiform gyri (FG), in the middle occipital gyri and in the cuneus of two visual(-cerebellar) ICNs, within-network connectivity increased with severity of complaints score in OA-mTBI, but decreased with severity of complaints in OA-HCs. The anterior FG and the middle occipital gyri are involved in both semantic and visual processing whereas the cuneus is part of the primary visual cortex. Furthermore, in clusters located in the cerebellum bilaterally, within-network connectivity and severity of complaints score were negatively correlated in OA-mTBI, but positively correlated in OA-HCs.
The question is how our findings can be related to the functional role of the cerebellum and of the visual cortex in light of previous research. Previous studies identified an association between PTCs and altered brain connectivity after mTBI (see van der Horn et al. 2016 and Biagianti et al. 2020 for reviews). A few of these studies identified connectivity within visual and/or cerebellar areas among those associated with PTCs (Nathan et al. 2015; Palacios et al. 2017; Stevens et al. 2012), although no clear pattern for their association emerged. It is known that, after mTBI, vision impairment including blurred vision and altered oculo-vestibular reflexes are common, although the etiology of these symptoms is still not well understood (Fife and Kalra 2015). In our OA-mTBI cohort, dizziness was the most prevalent self-reported symptom and complaint (see Appendix, Figs. S1 and S2) and complaints of balance disorders were (significantly) higher than in the OA-HC group, indicating that vestibular impairments might have been present during the subacute phase after trauma. It is therefore tempting to suggest a possible association of activity in the visual(-cerebellar) domain with vestibular impairments after mTBI. Perhaps the observed increasing hyperactivation in the visual cortex of OA-mTBI with severity of complaints score indicates increased effort to compensate for functional deficits, including oculo-vestibular impairments. However, as we do not have a direct association of complaints in the vestibular and/or visual domain with brain activity, such suggestion is speculative. Future studies are required to verify if a direct association of hypoactivation in the cerebellar area with vestibular impairments in OA-mTBI exists and elucidate if (attempted) compensation via increased activation in the visual cortex is part of this scene.
To the best of our knowledge, this is the first study to investigate (whole-)brain network connectivity in OA-mTBI. Our approach for denoising using multi-echo ICA resulted in all ICs identified as ICNs (as opposed to artifacts). Nevertheless, this study entails some limitations. First, in older adults, co-existing visual deterioration (that generally occurs as part of the aging process; Chou et al. 2016) could partly contribute to altered connectivity in visual networks and motor-balance control. However, objective measures of pre- and post-injury visual and motor-balance functioning were not available, which limits our analysis. Second, due to the cross-sectional nature of this study, it is unknown how the identified alterations on brain connectivity and their interaction with PTCs evolve over time. Longitudinal studies are required to identify whether the observed effects on brain connectivity and the interaction between altered connectivity and severity of complaints can be replicated and predict the persistence of complaints. This knowledge might help defining more effective rehabilitation strategies for older adults at risk of developing persistent PTCs.