Info
Background:
Concussion, or mild traumatic brain injury (mTBI), remains a diagnostic challenge because clinical symptoms often resolve while underlying neural mechanisms remain disrupted. Current “return-to-play” protocols rely heavily on sensory and symptomatic assessments; however, the recovery of higher-order brain networks is not well understood. Resting-state fMRI studies in humans have shown that the default mode network (DMN), a critical network involved in memory, memory consolidation, and internally directed cognition, remains disrupted beyond symptom resolution.
Hypothesis/Objective:
To address this, we leverage a translational porcine model of mTBI with controlled biomechanics and longitudinal imaging to investigate the relationship between functional network recovery and white matter reorganization. To further validate these findings, large-scale analysis will be used to assess whether these delayed network recovery trajectories generalize across a broader population.
Methods:
This study utilized two cohorts; eight concussed division one collegiate football players and 11 controls underwent longitudinal resting-state fMRI (rs-fMRI) at 24 hours, 7 days, and 30 days post-injury.
Twelve adolescent Yucatan minipigs were assigned to mTBI (n=6) or sham (n=6) groups. Injuries were delivered by an Electromagnetically Driven Impactor (EMDI), with Diffusion Tensor Imaging (DTI) acquired at 1 and 3 months post-injury to track white matter integrity (Fractional Anisotropy, FA).
Results:
In humans, DMN organization showed a delayed and prolonged disruption, with a peak at 7 days post-injury and incomplete recovery by 1 month, despite recovery of task-positive sensory systems during this period1. In parallel, the pig model demonstrated significant longitudinal white matter changes, with a 21% increase in fractional anisotropy (FA) in the corpus callosum from 1 to 3 months post-injury, compared to minimal change in shams2.
Conclusion:
DMN disruption peaks at 7 days post-injury and demonstrates only partial recovery by 1 month, FA in the corpus callosum increases progressively from 1 to 3 months, reflecting ongoing structural repair. Given the structural role of the corpus callosum in supporting large-scale networks such as the DMN, these findings suggest that early functional disruption is accompanied by delayed microstructural repair processes extending over months. Together, these results support a unified hypothesis: concussion might induce a disruption of large-scale functional networks, such as the DMN, that is underpinned by delayed white matter recovery. While clinical symptoms and sensory function may recover within days to weeks, underlying network integrity may require months to recover. The observed increase in FA likely reflects ongoing axonal reorganization and repair. This repair process involves the cingulum and hippocampal pathways, the same regions targeted in Alzheimer’s disease and CTE. The disparity between rapid symptomatic and sensory recovery (1 week) and prolonged DMN and structural reorganization (3 months) suggests that players returning to play early may remain at increased risk of second-impact syndrome or long-term implications.
To validate this hypothesis, our next step will be to apply this pipeline to the CARE Consortium (NCAA-DoD) dataset. By utilizing this large-scale human data, we aim to confirm if the DMN functional deficits and structural reorganization patterns observed in our smaller cohorts hold true across a large population, possibly informing safer return-to-play guidelines.
Image/Figure:
Click to view full size
Click to view full size
Image/Figure Caption:
Figure1. Default Mode Network (DMN) Entropy over time (P1: <24h, P2: 7d, P3: 30d). Concussed groups showed significantly increased entropy at 7 days (p < 0.01), with delayed and prolonged recovery. ANOVA revealed a significant group × time interaction (F=4.49, p=0.037). Fractional anisotropy (FA) in the corpus callosum significantly increased between 1 and 3 months post-injury in the mTBI group, compared to stable values in the sham group (n=6 for each group; mean ± SEM; p < 0.05).
Author
Sanjida Islam
Michigan State University