Impact frequency and interval modulate brain network outcomes in a rat model of repetitive mild traumatic brain injury

Abstract

Repetitive mild traumatic brain injury (mTBI) is an increasing public health concern, however, how impact frequency and inter-injury interval shape long-term brain network function remains unclear. This study investigated the effects of varying impact number and interval duration on functional connectivity and neurological outcomes in a rat model of repetitive mTBI. Fifty adult male Sprague–Dawley rats were randomly assigned to five groups: sham, single impact, two impacts (1-hour interval), three impacts (1-hour interval), and three impacts (6-hour interval) of closed-head injury. At 50 days post-injury, resting-state functional MRI and behavioral assessments were performed. Rats exposed to repetitive impacts at 1-hour intervals showed widespread disruptions in functional connectivity across default mode, sensory, and motor networks, with alterations increasing according to impact number. These network changes were attenuated in rats receiving impacts spaced 6 h apart, which exhibited connectivity patterns closer to sham controls. Behavioral testing revealed significant motor impairments in the 1-hour interval group, whereas motor performance was preserved in the 6-hour interval group. Connectivity within sensory and motor networks correlated with motor outcomes. These findings demonstrate that injury frequency and timing influence brain network integrity and suggest that extending inter-injury intervals may mitigate cumulative neural dysfunction following repetitive mTBI.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-52215-1.

Introduction

Repetitive mild traumatic brain injury (mTBI) has emerged as a significant clinical and public health concern, particularly in high-risk populations such as contact-sport athletes, military personnel, and victims of child abuse, including shaken baby syndrome^1,2. Clinical studies have shown that repetitive mTBI is associated with a range of acute symptoms, such as headache, dizziness, and confusion, collectively referred to as post-concussion syndrome, and may also result in chronic neuropsychiatric deficits^2,3. These include depression, anxiety, emotional dysregulation, and sleep disturbances^4,5. While a single mTBI typically results in transient symptoms with full recovery, repeated injuries can lead to cumulative neuropathological damage and long-term sequelae, including increased risks of chronic traumatic encephalopathy (CTE), Alzheimer’s disease, and Parkinson’s disease^3,6,7. The brain’s heightened vulnerability between injuries has been identified as a critical period during which additional trauma may exacerbate tissue damage and worsen outcomes^8,9.

A major clinical concern in repetitive mTBI is the premature return to activity following a concussion, exposing individuals to subsequent injuries before full recovery^10,11. In the United States, over 300,000 sport-related concussions occur annually, with many being recurrent^12,13. Both the number of impacts and the interval between them significantly influence clinical outcomes⁷. Multiple injuries have been linked to persistent cognitive deficits and higher rates of depression compared to single events^14,15. The inter-injury interval is another key variable; longer recovery periods between impacts are associated with reduced cognitive and behavioral impairments^7,16. Conversely, sustaining another mTBI during the brain’s vulnerable phase, prior to physiologic recovery, has been shown to increase the risk of cerebral edema and prolonged neurological dysfunction^16,17.

Although prior research has primarily focused on the effects of one or two mTBIs, emerging evidence suggests that increasing the number of impacts exacerbates long-term outcomes^18,19. For instance, retrospective studies on retired athletes have reported a dose-dependent relationship between injury frequency and the prevalence of depression^14,20. Similarly, preclinical models have demonstrated that shortened inter-injury intervals amplify cognitive deficits and histopathological damage^1,7. Animal studies indicate that spacing injuries farther apart allows partial restoration of neural function, thereby attenuating cumulative effects^17,21. However, the combined influence of both impact number and inter-injury timing remains incompletely understood, particularly in relation to alterations in brain networks and functional connectivity.

To address these gaps, the current study employed a modified closed-head injury (CHI) model in rats to systematically investigate the effects of varying impact numbers and inter-injury intervals on brain function^22,23. This refined weight-drop model enables improved reproducibility through fixed head positioning and precise delivery of impact forces. Using this controlled preclinical paradigm, we aim to elucidate the dynamic alterations in brain functional connectivity following single and repetitive mTBI under various conditions. Understanding these patterns and underlying associations with alterations in brain networks and behaviors is critical for informing clinical guidelines on safe recovery periods and preventing long-term impairment caused by mTBI²⁴.

We hypothesize that (1) increasing the number of impacts will exacerbate disruptions in brain functional connectivity, potentially leading to greater physical and psychological deficits, and (2) extending the inter-injury interval from 1 h to 6 h in the repeated injuries may allow for partial recovery or preservation of functional connectivity and behavioral outcomes. To assess long-term recovery, MRI scans and behavioral assessments were performed at 50 days post-surgery.

Materials and methods

All experimental protocols were approved by the Institutional Animal Care and Use Committee of National Yang Ming Chiao Tung University (IACUC No. 1090508). All methods were carried out in compliance with the ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines and the recommendations of National Institutes of Health Guidelines for Animal Research (the Guide for the Care and Use of Laboratory Animals). A total of 50 adult male Sprague-Dawley rats (BioLASCO Taiwan Co, Taipei City, Taiwan), weighing 350 to 370 g, were used in this study. The rats were randomly divided into 5 experimental groups: single mTBI (1xCHI, n = 5), repetitive mTBI of 2 impacts with a 1-hour interval (2xCHI, n = 5), repetitive mTBI of 3 impacts with a 1-hour interval (3xCHI-1 h, n = 5), repetitive mTBI of 3 impacts with a 6-hour interval (3xCHIi-6 h, n = 5), and a sham group (n = 30). A larger number of animals were included in the sham group to establish a robust baseline dataset for normal brain connectivity and to minimize bias in subsequent comparisons with injury groups. At the end of experimentation, rats were euthanized via CO₂ overdose.

Closed-head injury model

To replicate uncomplicated mTBI, a modified weight-drop method—specifically the CHI model—was employed in this study, as previously described^22,23,25. Briefly, a circular stainless-steel helmet (10 mm diameter) was affixed 3.5 mm posterior to the bregma using dental cement to prevent skull fracture at the impact site. The animal was subjected to the injury via the free-falling weight (600 g) with the assigned impact number and inter-injury interval. In the sham group, the animals underwent identical surgical preparation, but the impactor was suspended without contacting the skull, ensuring no mechanical force was applied. The time to regain righting reflex was monitored²³.

MRI protocol

Rats underwent MRI scanning following a previously established protocol^22,23,25. All animals (sham and injury groups) were imaged at ay 50 post-surgery using a 7 Tesla Bruker sequential PET/MR 7 T system (Bruker BioSpin). A volume coil was used as a transceiver for radiofrequency excitation and signal reception. Physiologic conditions, including heart rate, arterial pulse extension, oxygen saturation, and rectal temperature, were continually monitored and maintained within normal ranges throughout the experiment^22,23,25. T2-weighted anatomical images were acquired using a rapid acquisition with relaxation enhancement (RARE) sequence with the following parameters: TR/TE = 3000/30 ms, field of view = 3.5 × 3.5 cm, matrix = 256 × 256, slice number = 16, and slice thickness = 1 mm. Resting-state functional MRI (rsfMRI) was acquired 90 min following the sedative protocol with isoflurane and dexmedetomidine cocktail²⁶ using a single-shot echo-planar imaging (EPI) sequence with matched geometry. EPI parameters were: TR/TE = 1000/17.5 ms, matrix size = 96 × 96, number of repetitions = 300 for total scan time = 5 min.

Image pre-processing and atlas registration

All anatomical T2-weighted images were re-oriented to a common anatomical origin at the anterior commissure using Statistical Parametric Mapping software (SPM12)²⁷, followed by manual brain extraction. rsfMRI data were initially realigned using SPM12 to correct for head motion, with the realignment parameters set to a quality of 0.9, separation of 0.4 mm, and full width at half maximum (FWHM) of 1.0 mm. The slice-timing correction was applied to EPI to synchronize the timing across all slices within each volume. The rsfMRI images were then first co-registered to the high-resolution T2-weighted anatomical scans. Following bias field correction for intensity inhomogeneity, both the extracted structural brain images and the rsfMRI datasets were linearly registered to a common template using the FSL Linear Image Registration Tool (FLIRT)²⁸. The reference template was based on the Paxinos and Watson rat brain atlas²⁹. To address residual inter-subject anatomical variability not fully compensated by spatial normalization, and to enhance the signal-to-noise ratio, spatial smoothing was performed on the rsfMRI data using a Gaussian smoothing kernel with an FWHM of 1 mm in SPM12.

Functional connectivity analysis

Independent component analysis (ICA) and dual-regression methodology were employed to estimate resting-state functional connectivity using the Group ICA of fMRI Toolbox (GIFT)³⁰, which estimates statistically independent spatial sources from the fMRI dataset^31,32. The number of ICs was set within a range of 30 to 40 among each group of animals. Nine physiologically meaningful regions of interest (ROIs) were identified based on standard anatomical atlases of rat brain: (1) cingulate cortex (Cg), (2) retrosplenial cortex (RSC), (3) orbitofrontal cortex (OFC), (4) sensory barrel field region (S1BF), (5) primary forelimb sensorimotor cortex (S1FL), (6) secondary somatosensory cortex (S2), (7) visual cortex (VC), (8) motor cortex (M1), and (9) caudate putamen (Cpu) (Fig. 1). All maps of ICs were visually inspected and labeled as ROI-specific networks based on the spatial distribution of three major resting-state networks, including the default mode network (DMN), sensory network, and motor network. Statistical comparisons among groups were performed based on the Fisher-transformed z-scores for each ROI-related functional connectivity.

Fig. 1.

Regions of interest (ROIs) and the corresponding independent component (IC) maps. (A) Spatial distribution of nine selected ROIs. Representative ICs identified from group ICA associated with selected ROIs in (B) default mode network, (C) sensory network, and (E) motor network. The maximum of z-scores is labeled above the color bar for each IC map. Cg: cingulate cortex; RSC: retrosplenial cortex; OFC: orbitofrontal cortex; S1BF: sensory barrel field region; S1FL: primary forelimb sensorimotor cortex; S2: secondary somatosensory cortex; VC: visual cortex; M1: motor cortex; Cpu: caudate putamen.

Behavioral assessments

Neurological dysfunction was evaluated using the modified Neurological Severity Score (mNSS)³³, which integrates assessments across motor, sensory, reflex, and balance domains, with higher total scores indicating greater neurological impairment. Motor deficits were assessed using the beam walk test, which evaluates balance and coordinated motor function³⁴. The latency to complete the round-trip traversal was recorded. Animals that either slipped off the beam or failed to complete the task within the allotted time were assigned the maximum latency of 180 s. The open field test was used to evaluate the spontaneous locomotor activity and anxiety-like behavior in rats²². Parameters measured included mean speed, total distance traveled, time spent in each zone, and the number of entries into the outer zone were tracked and analyzed by using Smart 3.0 (Panlab, Barcelona, Spain)³⁵. For the sham group, all behavioral assessments were performed in a subset of randomly selected rats (n = 5).

Statistical analysis

Statistical analyses were performed using SPSS software (Version 24; IBM Corp., Armonk, NY, USA). Data normality was assessed using the Shapiro–Wilk test. To evaluate group differences in functional connectivity derived from the ICA approach, Welch’s ANOVA followed by Games-Howell post hoc comparisons was applied to the Fisher z-transformed data to account for the imbalance in group sizes between the sham and injury groups. Multiple comparisons among groups were corrected using the Bonferroni method. One-way ANOVA and Kruskal–Wallis tests were used to analyze behavioral parameters across groups. Pearson correlation analysis was conducted to assess the relationship between changes in functional connectivity and behavioral outcomes. A significance level of P<.05 was considered statistically significant for correlation analyses. All data are presented as mean ± standard error of the mean.

Results

The influence of impact numbers on repetitive mTBI

Notably, asymmetrical connectivity patterns of IC spatial maps were observed in the 3 hits/hour group associated with S1FL, S1BF, and M1, compared to all other groups (Fig. 2).

Fig. 2.

Spatial maps of ICs under different impact numbers and intervals. Functional connectivity maps related to nine ROIs in the (A) DMN, (B) sensory network, and (C) motor network are derived from sham and mTBI groups with different impact numbers and intervals. The maximum of z-scores is labeled above the color bar for each IC map. Cg: cingulate cortex; RSC: retrosplenial cortex; OFC: orbitofrontal cortex; S1BF: sensory barrel field region; S1FL: primary forelimb sensorimotor cortex; S2: secondary somatosensory cortex; VC: visual cortex; M1: motor cortex; Cpu: caudate putamen.

In the DMN (Fig. 3A), a significant reduction in connectivity strength was observed related to Cg between the 2 hits and 3 hits/hour groups (P<.01). Similarly, the RSC exhibited significantly lower connectivity in the 3 hits/hour group compared to all other groups (all P<.01) (Fig. 3A). In the sensory network (Fig. 3B), significantly decreased connectivity strength was found associated with both S1FL and S1BF in the 2 and 3 hits/hour groups compared to the sham group (all P<.01). In contrast, the S2 demonstrated a significant increase in connectivity in the 3 hits/hour group compared to the sham (P<.01) (Fig. 3B). In the VC, connectivity was significantly reduced in all CHI groups relative to sham (all P<.01) (Fig. 3B). In the motor network (Fig. 3C), M1 connectivity strength was significantly reduced in all injury groups compared to sham (all P<.01), with a stepwise decline as the impact number increased. Conversely, the Cpu exhibited a significant increase in connectivity in the 3 hits/hour group compared to sham (P<.01), with a progressive increase from 1 to 3 impacts (all P<.01) (Fig. 3C).

Fig. 3.

Functional connectivity associated with different impact numbers. Z-scores of functional connectivity for ROIs within the (A) DMN, (B) sensory network, and (C) motor network. *P<.0083, Bonferroni-corrected for multiple comparisons among groups. Cg: cingulate cortex; RSC: retrosplenial cortex; OFC: orbitofrontal cortex; S1BF: sensory barrel field region; S1FL: primary forelimb sensorimotor cortex; S2: secondary somatosensory cortex; VC: visual cortex; M1: motor cortex; Cpu: caudate putamen.

Regarding the beam walk test, latency to complete the task was significantly longer in the 2- and 3-hit groups versus sham (all P<.01) (Fig. 4), while the 1-hit group did not differ significantly from sham (P>.05). As shown in Supplementary Fig. 1 A, rats subjected to two impacts exhibited significantly higher mNSS scores than sham (P<.01). In the open-field test (Supplementary Fig. 1B-D), no significant differences were observed among groups in general locomotor activity (mean speed and total distance) or anxiety-like behavior (number of entries into the outer zone) (all P>.05). A mild upward pattern in entries into the outer zone was observed with increasing impact number (P>.05) (Supplementary Fig. 1D). Collectively, increasing the impact number in repetitive mTBI was associated with network-specific alterations in functional connectivity, accompanied by corresponding deficits in motor performance.

Fig. 4.

Beam walking performance associated with different impact numbers. Latency to complete the beam walk test from animals with different impact numbers. *P<.0083, Bonferroni-corrected for multiple comparisons among groups.

The effect of inter-injury intervals on repetitive mTBI

To evaluate the effect of inter-injury intervals on functional connectivity and behavioral outcomes, rats subjected to three impacts with either 1-hour or 6-hour inter-injury intervals were compared with the sham controls. The asymmetric patterns of connectivity observed in the 3 hits/hour group were not found in the 3 hits/6 hours group (Fig. 2).

In the DMN (Fig. 5A), the 3 hits/6-hour group exhibited significantly higher connectivity in both the Cg and RSC compared to the 3 hits/1-hour group (all P<.01), while RSC connectivity remained significantly reduced in both 3-hit groups compared to sham (all P<.01). In the sensory network (Fig. 5B), S1FL and S1BF connectivity were significantly reduced in the 3 hits/hour group compared to both the sham and 3 hits/6-hour groups (all P<.01). In the motor network (Fig. 5C), M1 connectivity was significantly reduced in both 3-hit groups compared to the sham, with further reduction in the 1-hour group compared to the 6-hour group (all P<.01). Conversely, Cpu connectivity significantly increased in the 3 hits/hour group compared to both the sham and 3 hits/6 hours groups (all P<.01) (Fig. 5C).

Fig. 5.

Functional connectivity associated with different inter-injury intervals. Z-scores of functional connectivity for ROIs within the (A) DMN, (B) sensory network, and (C) motor network. *P<.0167, Bonferroni-corrected for multiple comparisons among groups. Cg: cingulate cortex; RSC: retrosplenial cortex; OFC: orbitofrontal cortex; S1BF: sensory barrel field region; S1FL: primary forelimb sensorimotor cortex; S2: secondary somatosensory cortex; VC: visual cortex; M1: motor cortex; Cpu: caudate putamen.

Beam walk latency was significantly increased in the 3 hits/1-hour group compared to sham and 6-hour groups (all P<.01) (Fig. 6), while no difference was observed between sham and 6-hour groups. No significant difference in mNSS was observed between the sham and injury groups (Supplementary Fig. 2 A). General locomotor activity and anxiety-like behavior did not differ significantly among groups (all P>.05) (Supplementary Fig. 2B-D). A subtle increase in movement was observed in rats subjected to 3 hits with a 6-hour interval compared with those with a 1-hour interval (P>.05) (Supplementary Fig. 2B and C). These findings indicate that extending the inter-injury interval to 6 h attenuated network-level connectivity disruptions and preserved motor performance compared to closely spaced impacts.

Fig. 6.

Beam walking performance associated with different inter-injury intervals. Latency to complete the beam walk test from animals with. different inter-injury intervals *P<.0167, Bonferroni-corrected for multiple comparisons among groups.

The functional connectivity of sensorimotor networks significantly correlated with motor functions

Since only the beam walk data showed significant differences under varying impact numbers (Fig. 4) and inter-injury intervals (Fig. 6), we exclusively evaluated the correlation between the functional connectivity of all ROIs and beam walk performance. Among all animals, the decreased connectivity in S1FL (r=-.54, P<.01), S1BF (r=-.57, P<.01), VC (r=-.45, P<.05) and M1 (r=-.51, P<.01) was significantly correlated with increased latency to finish the beam walk test, respectively (Fig. 7). No significant correlation was found related to functional connectivity of other ROIs. Overall, decreased functional connectivity within sensorimotor regions, including S1FL, S1BF, VC, and M1, was significantly associated with impaired motor performance, as reflected by increased beam walk latency.

Fig. 7.

Scatter plots and significant correlations between the beam walk performance and the functional connectivity of (A) primary forelimb sensorimotor cortex (S1FL), (B) sensory barrel field region (S1BF), (C) visual cortex (VC), and (D) motor cortex (M1).

Discussion

Our study demonstrated that impact number and inter-injury interval of mTBI are two critical factors affecting long-term brain functional connectivity and behavioral outcomes. Behavioral data showed that repetitive mTBI led to worsened neurological scores and impaired motor function. Repetitive mTBI with increased impact numbers, especially 3 hits/hour, led to widespread disruptions in functional connectivity across the DMN, sensory, and motor networks, with marked reductions in Cg, RSC, S1FL, S1BF, VC, and M1. Conversely, increased connectivity was observed in S2 and Cpu regions, indicating region-specific network alterations associated with higher impact frequency. Rats receiving 2 and 3 injuries exhibited significantly greater deficits compared with sham, especially in the mNSS and beam balance test. These results suggested that increasing the impact number aggravates neurobehavioral impairments in the chronic phase.

Our analysis of resting-state functional connectivity revealed widespread and impact-dependent alterations across brain network, even in the absence of overt cortical lesions. Significant hypoconnectivity was found across several networks as the impact number increased. In DMN, connectivity within Cg and RSC decreased markedly after 3 impacts (Fig. 3A). Given its proximity to the impact site and its role in spatial memory, RSC may be particularly susceptible to brain injury³⁶. These findings are in line with clinical TBI studies indicating DMN vulnerability to repetitive head trauma^37,38. In the sensory network, hypoconnectivity was observed in S1FL, S1BF, and VC, while S2 showed hyperconnectivity after multiple impacts (Fig. 3B). Consistent with prior studies in neurological conditions such as stroke and panic disorder^39,40, where S2 activity is relatively preserved despite S1 impairment, these findings suggest a compensatory reorganization within the somatosensory network^41,42, potentially supported by alternative thalamocortical inputs directly to S2^43,44. Similarly, in the motor network (Fig. 3C), decreased connectivity in M1 aligned with behavioral deficits, whereas hyperconnectivity in Cpu following rmTBI may reflect injury-induced adaptation within the basal ganglia circuits. Following brain injury^45,46,increased functional connectivity has been proposed as a compensatory mechanism to maintain network communication despite underlying structural disruption⁴⁷. Consistent with prior studies reporting elevated caudate connectivity during the subacute phase after controlled cortical impact injury⁴⁸, the hyperconnectivity observed in the Cpu in our study may reflect a compensatory reorganization that helps preserve motor function following repetitive mTBI. Of note, in contrast to the previous rodent studies focusing on focal TBI models with extensive tissue damage, such as the controlled cortical impact model^48,49,our findings demonstrate that rmTBI, even without visible structural damage, can still lead to robust and system-wide network dysfunction^22,23. Emerging preclinical evidence suggests several mechanisms underlying such alterations in functional connectivity. Disruption of white matter integrity, including gliosis and neurodegenerative changes, has been associated with impaired long-range connectivity^50,51,while neuroinflammatory responses, particularly astrocyte and microglial activation, have been linked to both local circuit dysfunction and large-scale network alterations⁵². In the present study, although quantitative immunohistochemical analysis was not performed, preliminary findings revealed increased GFAP expression (Supplementary Fig. 3), particularly in groups with higher impact numbers and shorter inter-injury intervals. In addition, our previous work using the same CHI model demonstrated persistent astroglial activation in remote white matter regions during the chronic phase²⁵. These observations support the hypothesis that inflammation-mediated processes may contribute to the network-level connectivity disruptions observed following rmTBI. At the molecular level, dysregulation of the mechanistic target of rapamycin (mTOR) signaling pathway, particularly the mTORC1 complex, has been reported in the pathophysiology of TBI^53,54. Aberrant mTOR activation has been associated with key secondary injury processes, including neuroinflammation, gliosis, oxidative stress, impaired autophagy, and neuronal apoptosis^53,55,56, all of which may contribute to synaptic dysfunction and subsequent alterations in large-scale brain network organization. Modulation of mTOR signaling may therefore represent a potential strategy to mitigate neuroinflammation^53,55,57 and preserve functional connectivity, warranting further investigation in our CHI model.

To examine the effect of inter-injury interval on functional connectivity, we compared rats receiving 3 hits spaced 1-hour versus 6-hours apart. Rats in the 6-hour group showed better preservation of connectivity across multiple regions, including DMN (Cg, RSC), sensory (S1FL, S1BF, VC), and motor (M1, M2) networks (Fig. 5). These animals also showed relatively improved neurological and sensorimotor outcomes compared to the 1-hour group (Fig. 6). Our findings suggest that extending the inter-injury interval may mitigate cumulative network disruption, potentially allowing for partial functional recovery. Our results align with the prior studies indicating that longer intervals reduce axonal injury and neuroinflammation^7,17. While the optimal interval for recovery is not fully defined, our results demonstrate that even a modest increase in inter-injury interval (from 1 h to 6 h) significantly impacts functional restoration in the chronic phase. This has important translational implications for injury management in high-risk populations such as professional contact-sport athletes⁵⁸.

Beam walk assessment, a well-established measure of vestibulomotor coordination and balance that requires precise integration of sensorimotor pathways, has been shown to be sensitive to impairments following brain injury^59–61. In the current study, the prolonged traversal time observed after rmTBI with clear dependence on impact parameters (Figs. 4 and 6) further demonstrates the sensitivity of this task in detecting cumulative injury effects. As we observed significant changes in both functional connectivity and beam walk performance following brain injury, Pearson’s correlation analyses revealed that reduced connectivity in sensory and motor networks was strongly associated with impaired beam walk performance, as indicated by longer completion duration (Fig. 7). Our results were consistent with previous findings linking sensorimotor function to motor ability⁶². Abnormal functional connectivity in the sensorimotor network has been associated with reduced motor speed in mTBI patients⁶³. In addition, visual deficits, such as blurred or double vision, are common after concussion⁶⁴. Our findings, consistent with prior studies in both humans and animal models^64–66, suggest that disrupted connectivity in sensory and motor circuits may underlie the motor and visual impairments seen after mTBI, further emphasizing the interplay between brain connectivity and functional outcomes in the context of brain injury.

Several limitations must be acknowledged. First, the relatively small sample size in the injury groups (n = 5 per group), together with the imbalance compared with the larger sham group (n = 30), represents a limitation that may influence statistical power and effect size estimation. While post hoc power analysis (GPower 3.1.9) indicated adequate power (1–β > 0.82 across comparisons) for the observed effect sizes (d > 1), the findings should be interpreted with caution. Future studies with larger and more balanced group sizes are warranted to enhance the robustness and generalizability of these results. Second, our study focused on within-network connectivity changes and did not explore inter-network interactions. Future studies should examine connectivity between networks such as DMN, salience, and executive control systems to gain deeper insight. Third, while a subtle alteration in exploratory behavior was observed with increasing impact number in the current study (Supplementary Fig. 1), the absence of dedicated cognitive assessments limits our ability to directly link network changes to cognitive dysfunction. Future studies incorporating targeted tests of anxiety, learning, memory, and executive function are warranted. Lastly, although our preliminary data indicate increased astrocyte accumulation following rmTBI (Supplementary Fig. 3), quantitative analysis of neuroinflammation based on immunohistochemistry (IHC) was not performed. Future studies will incorporate systematic IHC analyses, prioritizing markers of astrogliosis and neuroinflammation, such as GFAP and Iba-1, which will be essential to better elucidate the mechanistic relationship between pathological changes and functional connectivity alterations.

Conclusion

Our findings provide clear evidence that increasing the impact number worsens connectivity disruptions and behavioral deficits in repetitive mTBI. Meanwhile, extending inter-injury intervals to 6 h reduces these impairments and supports partial network recovery. RsfMRI provides a valuable tool to map functional alterations following different conditions of repetitive mTBI. The cumulative effects of repeated injuries and the protective role of longer intervals highlight the need to consider both frequency and timing in the clinical management of repetitive mTBI.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Y.-H.K.: formal analysis, investigation and writing— original draft. C.-F.L.: data curation, visualization, writing—review and editing. B.-Y.H: visualization, writing—review and editing. Y.-C.J.K.: conceptualization, funding acquisition, methodology, data curation, validation, and supervision, visualization, writing—review and editing.All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Funding

This work was supported by a research grant from the National Science and Technology Council (NSTC) of Taiwan (NSTC 114-2314-B-A49-064 and NSTC 114-2320-B-A49-005).

Data availability

The data that support the findings of this study are openly available in Dataverse at https://doi.org/10.57770/PPSERZ.

Declarations

Competing interests

The authors declare no competing interests.