Here, we present a protocol to establish a closed-head injury animal model replicating the neuroimage outcome of uncomplicated mild traumatic brain injury with the preserved brain structure in the acute phase and long-term brain atrophy. Longitudinal magnetic resonance imaging is the primary method used for evidence.

Mild traumatic brain injury (mTBI), known as concussion, accounts for more than 85% of brain injuries globally. Specifically, uncomplicated mTBI showing negative findings in routine clinical imaging in the acute phase hinders early and appropriate care in these patients. It has been acknowledged that different impact parameters may affect and even accelerate the progress of subsequent neuropsychological symptoms following mTBI. However, the association of impact parameters during concussion to the outcome has not been extensively examined. In the current study, an animal model with closed-head injury (CHI) modified from the weight-drop injury paradigm was described and demonstrated in detail. Adult male Sprague-Dawley rats (n = 20) were randomly assigned to CHI groups with different impact parameters (n = 4 per group). Longitudinal MR imaging studies, including T2-weighted imaging and diffusion tensor imaging, and sequential behavioral assessments, such as modified neurological severity score (mNSS) and the beam walk test, were conducted over a 50-day study period. Immunohistochemical staining for astrogliosis was performed on day 50 post-injury. Worse behavioral performance was observed in animals following repetitive CHI compared to the single injury and sham group. By using longitudinal magnetic resonance imaging (MRI), no significant brain contusion was observed at 24 h post-injury. Nevertheless, cortical atrophy and alteration of cortical fractional anisotropy (FA) were demonstrated on day 50 post-injury, suggesting the successful replication of clinical uncomplicated mTBI. Most importantly, changes in neurobehavioral outcomes and image features observed after mTBI were dependent on impact number, inter-injury intervals, and the selected impact site in the animals. This in vivo mTBI model combined with preclinical MRI provides a means to explore brain injury on a whole-brain scale. It also allows the investigation of imaging biomarkers sensitive to mTBI across varying impact parameters and severity levels.

Mild traumatic brain injury (mTBI) is primarily observed in athletes engaged in contact sports, military veterans, and individuals involved in traffic accidents¹. It accounts for greater than 85 % of all reported head injuries². The vast etiology of mTBI and its increasing global incidence underscore the inclusion of mTBI as a tentative environmental risk factor of late-onset neurodegenerative disease³. Uncomplicated mild TBI is characterized by a Glasgow Coma Score (GCS) of 13-15, with no structural abnormalities observed in computer tomography (CT) or magnetic resonance imaging (MRI) scans. Common symptoms experienced by patients with uncomplicated mTBI include headaches, dizziness, nausea or vomiting, and fatigue. However, longitudinal assessment of outcomes following uncomplicated mTBI presents considerable challenges due to the high dropout rate in patients⁴.

The concerns of repetitive mTBI have increased, particularly within the National Football League (NFL) professional athlete community, subsequently raising awareness among non-professional athletes⁵. Brain vulnerability is presumed to increase following the initial mTBI, with subsequent insults potentially exacerbating injury outcomes. Recent findings from the largest donated brain cohort of football players not only implicated prior football participation in chronic traumatic encephalopathy (CTE) severity but also suggested a correlation between different football-related factors and the risk and severity of CTE⁶. Hence, the concern about the influence of the number of concussions and the repetitive regime on injury outcomes is growing. Preclinical research has explored neuropathological changes, neuroinflammatory cascade, and neuropsychological impairment after repetitive mTBI by using various closed-head injury (CHI) models^{7,8,9,10,11,12,13,14}. However, the investigation of impact parameters on the uncomplicated mTBI model, which may closely mimic sport-related repetitive concussive head impacts resulting in functional impairment in the acute phase and brain atrophy in the chronic phase, has not been well examined.

Diffusion tensor imaging (DTI), a technique assessing the diffusion of water molecules, has been commonly utilized in studies investigating the effects of mTBI. Fractional anisotropy (FA), a key metric derived from DTI, quantifies the degree of water diffusivity coherence and provides information regarding the structural organization of axons and nerve fiber bundles. Perturbation of FA values in the white matter (WM) has been proposed following mTBI in various models^{8,10,11,15,16,17}. In addition, axial diffusivity (AD) and radial diffusivity (RD), indicating axonal and myelin integrity, changed after mTBI in preclinical studies^{10,15,16,18,19,20}. However, discrepancies in DTI findings among previous studies are likely due to variations in mTBI severity, differences in impact parameters, diverse mTBI models, and inconsistent post-injury follow-up time points⁹.

The current protocol paper, thus, aims to establish an animal model of mTBI designed to evaluate the cumulative effects of single and repetitive mTBI. We incorporated comprehensive and longitudinal assessments, including evaluations of animal well-being, behavioral outcomes, DTI parameters, and cortical volume, to capture dynamic post-injury changes and explore the effects of different impact parameters. By demonstrating both acute functional impairment and long-term microstructural changes, this model effectively replicates the key features of uncomplicated mTBI that were not fully addressed in previous animal studies. Here, we provided a detailed protocol for developing an uncomplicated mTBI model using a modified closed-head weight-drop method^8,11 and conducting longitudinal assessment following mTBI.

The study was performed in accordance with the recommendations of the National Institutes of Health Guidelines for Animal Research (Guide for the Care and Use of Laboratory Animals) and the Animal Research: Reporting In Vivo Experiments guidelines. All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the National Yang Ming Chiao Tung University. Twenty animals were randomly assigned to 5 groups (n = 4 per group): (i) single impact at the sensorimotor cortex (SMCx/single), (ii) double impacts at SMCx with the 1-h interval (SMCx/2 hits/1 h), (iii) double impacts at SMCx with the 10-min interval (SMCx/2 hits/10 min), (iv) double impacts at the central brain with the 1-h interval (Central/2 hits/1 h), and (v) the sham group with surgery only but not impact directly to the head, for longitudinal outcome assessment (Figure 1). Of note, the inter-injury intervals selected for this study (1-h vs. 10-min intervals) were designed to mimic the repetitive subconcussive impacts^{8,10,11,13,21}, which can be up to a thousand times within a single season, experienced by the athletes engaging in contact sports^22,23.

1. Induction of closed-head injury (CHI)

NOTE: Adult male Sprague-Dawley rats aged 10 to 12 weeks and weighing over 250 g are housed under a 12/12 h light/dark cycle with ad libitum access to food and water.

1. Place the rat in a small induction chamber and anesthetize it with a mixture of isoflurane (5%) and medical air (2.5-3 L/min). Remove the rat from the chamber until it is non-responsive to a paw or tail pinch.
2. Place the rat on the heating pad.
   NOTE: Turn on the heating pad during the surgery to maintain the rats' body temperature.
3. Bring the rat onto a stereotaxic frame and secure it with a tooth bar. Administer isoflurane at 2% using the connected nose cone with medical air at a flow rate of 1.5-2 L/min for maintenance during the surgery.
4. Position the ear bars. Ensure the rat is centered and symmetrical on the stereotaxic frame.
   NOTE: All surgical procedures should be performed under aseptic conditions. The instruments were sterilized before surgery using a steam autoclave, and their tips were additionally sterilized with a bead sterilizer during the procedure. To prevent contamination, a surgical drape was placed over the animal. The surgeon wore a cap to cover their hair, a mask to cover their face, and was equipped with a lab coat and surgical gloves during the procedure²⁴.
5. Put on the sensor of the pulse oximeter to the hindpaw of the animal to monitor the respiration rate, heart rate, blood oxygen level, and body temperature of the animal.
6. Inject 1 mL/kg body weight of lidocaine (20 mg/mL) subcutaneously into the rat's neck as an analgesic.
7. Apply hair removal cream to the head of the animal and wait for 3 min. Wipe the cream out with 70% isopropyl alcohol swabs.
8. Clean the shaved area several times using a sterile cotton swab soaked in iodine. Remove the iodine residue using a cotton swab soaked in 70% ethanol.
9. Create a midline incision approximately 2-2.5 cm in length in the shaved skin using sterile surgical blade to access the surface of the skull.
10. Remove the tissue on the bone using a cotton pad to expose the skull. Clean the skull surface using a cotton swab soaked in 0.9% saline, then clean it with a dry cotton pad.
    NOTE: The skull sutures and both bregma and lambda can now be easily identified.
11. Identify the bregma as the reference point to further find out the impact area based on the coordinate.
    NOTE: In this protocol, two sets of coordinates are utilized for CHI induction: (-2.5,-2.0) (2.5 mm lateral 2.0 mm posterior to the bregma) on top of the sensorimotor cortex (SMCx), as well as (0,-3.0) on top of the central brain (central).
12. Identify the chosen coordinates on the skull surface and cement a circular stainless steel helmet (10 mm diameter and 1 mm thickness) over the designated area using dental cement. Remove the heating pad and pulse oximeter.
13. Move the stereotactic device and the rat on the lift table (14 cm length, 8 cm width, and 6.15 cm depth) under the CHI impactor.
14. Elevate the body of the rat using a foam sponge (19 cm in length, 10 cm in width, and 4 cm in depth, with a density of 18 kg/m³).
15. Remove the rat from the ear bars of the stereotaxic frame. Keep the rat still on the tooth bar connected with the nose cone, delivering 2% isoflurane. Ensure that the head and body are aligned level in the rostral-caudal direction.
16. Adjust the lift table to ensure no space between the CHI impactor and the helmet. Turn off isoflurane 5 s right before the impact.
    NOTE: To specify the righting reflex due to brain injury, temporary cessation of isoflurane was performed²⁵.
17. Drop a 600 g brass from a height of 1 m through a stainless steel tube (1 m height with an inner diameter of 20 mm for clearing a column of stainless brass weights) to the secured impactor with a round tip aiming at the metal helmet.
    NOTE: Animals in the sham group did not experience an impact, as the brass drop was released without making contact with the helmet on the rat's head.
18. Lower the lift table. Remove the rat from the stereotaxic frame and place the rat in the supine position on a heating pad.
19. Record the time of the righting reflex, which is the time when the animal attempts to change from the supine position into the prone position^26,27.
    NOTE: The animals subjected to the repetitive CHI were anesthetized again 3 min before the 2^nd impact. For the animals in the SMCx/double/10 min group who did not flip back to the prone position in time, the corresponding time of righting reflex was recorded as 420 s.
20. After the righting reflex recording, anesthetize the rat with isoflurane again using step 1.1.
21. Immobilize the rat with the stereotaxic frame using step 1.2.
    NOTE: After confirming the stability of the helmet on top of the stull, repeat steps 1.13-1.17 again to perform the 2^nd impact.
22. Remove the helmet. Remove all the connective tissue and cement on top of the skull.
23. Cover the skull with the dental cement and allow it to dry. Check that the dental cement is stiff and hard using the tweezer back.
    NOTE: The dental cement was applied atop the skull to eliminate susceptibility artifacts caused by skull-air or skull-blood interfaces between the skull and scalp post-surgery.
24. Close the incision using 4-0 nylon surgical sutures with 4-5 independent knots.
    NOTE: The wound is approximately 2-2.5 cm in length. Ensure that the surgical sutures have no capillary action and are made of silk or nylon material. Do not suture the incision using one single knot to prevent the opening of the wound by the scratching of the animal.
25. Apply topical antibiotics (Dermanest cream) to the surgical site to prevent infections.
26. Inject 1 mL/kg body weight of carprofen (50 mg/mL) subcutaneously into the neck as the post-surgery analgesics.
27. Place the rat in a clean cage on a heating pad till it regains consciousness. Once the rat sits upright, return it to the home cage.
28. Orally administer 5 mL of acetaminophen (24 mg/mL) mixed in 200 mL of water to the animal daily as an analgesic for 3 consecutive days after surgery.

2. Magnetic resonance imaging (MRI)

NOTE: T2-weighted image and diffusion-tensor imaging are performed using a sequential PET/MR 7T system before CHI, as well as on 1 and 50 days post-injury (Figure 1). A baseline MRI was performed within 1 week before the CHI procedure. For the evaluations on 1 and 50 days post-CHI, the behavioral assessments were conducted in the morning, followed by MRI scans in the afternoon on the same day.

1. Anesthetize the rat in a small induction chamber filled with a mixture of isoflurane (5%) and medical air (2.5-3 L/min).
2. Once the rat is non-responsive to a paw or tail pinch, temporarily suspend the anesthesia when transferring to the animal cradle in a head-first prone position.
3. Position the rat in the head holder connected with a nose cone, delivering 2% isoflurane with medical air at a flow rate of 1.5-2 L/min for maintenance during the image acquisition.
4. Fixate the head with a small piece of tape to avoid movement during scanning.
   NOTE: Apply some toothpaste to the rat's head on the first-day post-CHI to prevent magnetic susceptibility artifacts due to the scalp removal^28,29.
5. Place a pressure pad under the thorax of the rat to monitor respiration. Tape the oximeter clips to the hind limb to monitor heart rate.
6. Insert the rectal probe to measure rectal temperature. Cover the rat with a heating blanket with circulating warm water and tissue wrap during the experiment to maintain body temperature.
   NOTE: Throughout the experiment, monitor physiological conditions, including heart rate, respiratory rate, and rectal temperature. Before scanning, check all the physiological signals of the rat to ensure the quality of the vital sign monitor.
7. Use the laser positioning system of the PET/MR scanner to mark the center of the head for precise alignment.
8. Move the animal into the MRI bore automatically using the motorized animal transport system until the center of the head aligns with the iso-center of the scanner.
   NOTE: The motorized animal transport system integrated into the PET/MR system ensures accurate animal positioning and streamlines the workflow when transitioning between imaging modalities.
9. Obtain MRI sequence.
   1. Perform initial localization and overall adjustment.
   2. Use the middle sagittal slice and align the 8^th slice from the anterior with the decussation of the anterior commissure.
      NOTE: The coronal slices are positioned perpendicular to the horizontal plane defined by the line connecting the anterior commissure and the base of the cerebellum, corresponding to an approximately 15° angle relative to the long axis of the corpus callosum. The key parameters of the T2-RARE scan for the middle sagittal slice are as follows: repetition time (TR) = 2500 ms, echo time (TE) = 44 ms, field of view (FOV) = 3.5 cm, matrix size = 256 256, slice thickness = 1 mm, number of slices = 1, RARE factor = 8, bandwidth = 75 kHz, number of averages = 1, the acquisition time = 1 min 20 s.
   3. Use rapid acquisition with relaxation enhancement (RARE) with fat suppression and a saturation band underneath the brain to obtain T2-weighted images for anatomic reference (Figure 2).
      NOTE: The key parameters of the T2-RARE scan are as follows: repetition time (TR) = 3600 ms, echo time (TE) = 40 ms, field of view (FOV) = 2 cm, matrix size = 256 256, slice thickness = 1 mm, number of slices = 16, RARE factor = 8, bandwidth = 75 kHz, number of averages = 8, the acquisition time = 7 min 40 s.
   4. Use a 4-shot spin-echo EPI to acquire diffusion tensor images (Figure 2).
      NOTE: The key parameters of the DTI scan are as follows: TR = 3000 ms, TE = 28 ms, field of view (FOV) = 2 cm, matrix size = 96 96, thickness = 1 mm, number of slices = 16, duration of the pulse (δ) = 5 ms, time between the two pulses (Δ) = 15 ms, number of B0 = 5, number of directions = 30, b-value = 1000 s/mm³, bandwidth =150 kHz, number of average = 4, the acquisition time = 14 min.
   5. Finish the scanning protocol. Slide the animal cradle out of the magnet. Remove the animal from the cradle.
   6. Transfer the rat to a clean cage with a heating pad underneath to maintain it's body temperature. Return the rat to its home cage once it regains consciousness.
10. Image preprocessing
    NOTE: Use MRtrix3, Statistical Parametric Mapping (SPM) software, and custom MATLAB scripts for data processing and analysis.
    1. Denoise the DTI images using the MRtrix3 command (dwidenoise)³⁰.
    2. Remove Gibb's ringing artifacts from the DTI images using the MRtrix command (mrdegibbs)³⁰.
    3. Coregister DTI images to T2-weighted images of the single subject across longitudinal scans using SPM functions (spm_coreg.m and spm_powell.m).
    4. Perform skull stripping on T2-weighted images by manually contouring the brain area slice-by-slice, followed by removing pixels with intensity below the computed threshold determined by Otsu's method³¹ (custom MATLAB script thr_otsu2.m).
    5. Perform cross-subjects coregistration among animals in the same experimental group using SPM functions (spm_coreg.m and spm_powell.m). Apply the brain mask to the corresponding DTI images.
       NOTE: Skull stripping is performed to decrease computer processing time.
    6. Calculate tensor maps based on DTI (custom MATLAB script, tensormap.m).
    7. Calculate FA maps (custom MATLAB script, calFA.m)
       NOTE: All the custom MATLAB scripts are available via the following database (https://doi-org.remotexs.ntu.edu.sg/10.57770/9ZESXD).
11. Image analysis-FA
    1. Draw regions of interest (ROIs) in the cortex and corpus callosum (CC) for the three consecutive image slices underneath the coordinate of CHI.
       NOTE: All ROIs were manually drawn and visually inspected for gross errors by 2 experienced investigators blinded to the experimental groups.
    2. Extract and average the FA value from ROIs.
       NOTE: For the corpus callosum underneath the cortex, pixels with values of FA < 0.35 were excluded in the selected ROI to eliminate partial volume effects. For the cortex, all the pixels with values of FA < 0.35 in the ROI were recruited for analysis.
12. Image analysis-Volume
    1. Manually draw ROIs covering the cortical regions for 11 consecutive image slices at Bregma -7 to +3 mm.
       NOTE: All ROIs were manually drawn by 2 experienced investigators blinded to the experimental groups.
    2. Sum up the total pixels of the ROIs across slices and transform them into the volume by multiplying the slice thickness (1 mm).
    3. Normalize the cortical volume after CHI with the corresponding volume before CHI of each animal.
       NOTE: Normalize the data to eliminate individual differences in brain volume across animals before presentation.

3. Behavior assessment

NOTE: The behavioral experiments are performed using the beam walk balance test and mNSS before CHI, as well as on 1 and 50 days post-CHI (Figure 1). All the assessment was performed by at least two observers to ensure the accuracy, consistency, and objectivity of the collected data.

1. Beam walk balance test
   1. Turn on the video camera and start the timer.
   2. Place the rats at one end of the balance beam (3 cm in depth, 3 cm in width, 80 cm in length, and 60 cm above the floor).
   3. Stop the timer once the rat completes one round trip, falls over, or freezes over 3 min.
      1. During the experiment, observe the animal condition for the assessment using mNSS^11,32,33,34. Follow these standards for evaluation:
      2. If the rat maintains balance with a steady posture on the beam, assign a score of 0.
      3. If the rat grasps the side of the beam, assign a score of 1.
      4. If the rat falls with one limb off the beam, assign a score of 2.
      5. If the rat falls with two limbs off or spins on the beam (>60 s), assign a score of 3.
      6. If the rat attempts to balance on the beam but falls off (>40 s), assign a score of 4.
      7. If the rat attempts to balance on the beam but falls off (>20 s), assign a score of 5.
      8. If the rat does not attempt to balance or hang on the beam and falls off within 20 s, assign a score of 6.
      9. If the rat fails to complete the task, consider it as having taken the maximum time of 3 min and assign a score of 6.
   4. Schedule testing days at specific time points.
      NOTE: Exclude rats not completing the beam walk round trip for two trials before CHI from the subsequent surgery and follow-up behavioral assessment.
2. Modified neurological severity score (mNSS)
   NOTE: mNSS assessment includes motor tests, sensory tests, absence of reflex, abnormal movements, beam balance, and walking on the floor ^32,33, which was quickly performed on a daily basis.
   1. Perform motor tests.
      1. Raise the rat by the base of the tail and observe the reflexes of its limbs for about 15 s to assess the proper flexion and extension.
      2. If normal flexion is observed in the forelimb, assign a score of 0. If no flexion is observed, assign a score of 1.
      3. If normal flexion is observed in the hindlimb, assign a score of 0. If no flexion is observed, assign a score of 1.
      4. If the head moves >10° to the vertical axis within 30 s after raising the rat by the tail, assign a score of 0. If not, assign a score of 1.
         NOTE: A maximum of 3 scores will be assigned in this session of the test.
   2. Perform limb placing tests.
      NOTE: The placing test is performed to evaluate the coordination between sensory (visual, tactile sensation, and proprioception) and motor function.
      1. Slowly lower the rat toward the surface of the table. Observe whether the paws of the rats reached and stretched toward the surface.
      2. If the rats reached the surface with both limbs stretched and forward, assign a score of 0. If there is a delay or no response, assign a score of 1.
      3. Place the rat on the surface and pull the paw against the table edge. Observe whether its paw returns to a normal position on the surface of the table.
      4. If immediate and normal placing responses are observed, assign a score of 0. If delayed placing responses are observed, assign a score of 1. If there is no response, assign a score of 2.
         NOTE: A maximum of 3 scores will be assigned in this session of the test.
   3. Observe reflect absence and abnormal movements.
      1. Observe for head shakes to assess the pinna reflex when touching the auditory meatus with the cotton end of a cotton swab.
      2. If a normal reflex is observed, assign a score of 0. If no reflex is observed, assign a score of 1.
      3. Assess the presence of the corneal reflex by touching the cornea with the cotton end of a cotton swab.
      4. If a normal response is elicited, assign a score of 0. If no eye blink response is elicited, assign a score of 1.
      5. Make a short and strong clap of the hands. Observe the presence of the startle reflex.
      6. If a reflex is observed, assign a score of 0. If no reflex is observed, assign a score of 1.
      7. Observe whether the rat has seizures, myoclonus, or myodystony.
      8. If any of them occur, assign a score of 1.
         NOTE: A maximum of 4 scores will be assigned in this session of the test.
   4. Perform the beam balance test, as previously described (step 3.1).
      NOTE: A maximum of 6 scores will be assigned in this session of the test.
   5. Perform walk on the floor test.
      1. Prepare the open field arena (75 cm length, 50 cm width, and 40 cm depth). Ensure it is clean and free from any previous odor cues.
      2. Place a rat in the center of the open field arena and observe how the rat walks in the arena.
      3. If the rat performs a regular walk, assign a score of 0.
      4. If the rat cannot walk straight, assign a score of 1.
      5. If the rat falls to the paretic side after placing it on the floor, assign a score of 3.
         NOTE: A maximum of 3 scores will be assigned in this session of the test.
   6. Sum up all the scores; the maximum possible score is 18.
      NOTE: The higher score indicates a worse outcome.

4. Immunohistology

1. Perform transcardiac perfusion³⁵.
   NOTE: Transcardiac perfusion is performed after the MRI scan 50 days post-CHI (Figure 1).
   1. Place the rat in a small induction chamber and anesthetize it with isoflurane (5%) until it is non-responsive to a paw or tail pinch.
   2. Administer 50 mg/kg body weight of zoletil (50 mg/mL) and 10 mg/kg body weight of Xylazine (Roumpun, 23.32 mg/mL) via intraperitoneal injection for deep anesthesia.
   3. Place the rat in the supine position.
   4. Make a transverse incision approximately 4-5 cm in length under the thorax using scissors.
   5. Locate the diaphragm and cut it to expose the heart.
   6. Use hemostatic forceps to clamp the pulmonary artery and then make an incision approximately 0.5-1 cm in length in the right atrium.
   7. Connect the needle to the pipeline attached to the infusion pump.
   8. Insert the needle into the left ventricle.
   9. Rinse the animal through transcardiac perfusion (40 mL/min) with 500 mL of 0.9% saline until the blood is cleared.
   10. Perfuse the animal through transcardiac perfusion (40 mL/min) with 500 mL of 4% paraformaldehyde (PFA) for fixation.
   11. Remove the rat's head, and carefully peel the brain tissue from the skull.
   12. Preserve the brain in approximately 20 mL of 4% PFA in the bottle for 48 h for post-fixation.
2. Perform tissue processing and IHC staining^36,37.
   NOTE: Perform immunohistochemical staining on formalin-fixed, paraffin-embedded tissue sections using the immunoperoxidase secondary detection system kit.
   1. Use formalin-fixed and paraffin-embedded tissue sections.
   2. Perform deparaffinization and treat the slides with 3% H₂O₂ to block endogenous peroxidase activity. Perform antigen retrieval using citrate buffer at 90 °C.
   3. Perform immunohistochemical staining using the immunoperoxidase secondary detection system kit.
      NOTE: Staining procedures are carried out following the manufacturer's recommendations.
   4. Use hematoxylin to counterstain specimens.
   5. Mount specimens with an antifade reagent.
   6. Use anti-glial fibrillary acidic protein (GFAP) antibodies for immunohistochemical staining.
   7. Acquire the images of ROIs using a light microscope slide scanner (Figure 6).

5. Statistical analysis of behavior and image outcomes

NOTE: In the current study, statistical analysis was performed in SPSS; however, the statistical analysis can be performed in other statistical toolboxes.

1. Load the data in the wide format in an SPSS *.sav file.
2. Conduct repeated measures analysis of variance (ANOVA) to compare the behavioral (normalized weight, mNSS, and beam walk duration) and image outcomes (FA values in the cortex and CC) over time among groups.
   1. Click Analyze > General Linear Model > Repeated Measures.
   2. Assign a name in the Within-Subject Factor Name box (e.g., time) and put '3' in the Number of Levels box (three levels with different follow-up time points). Assign a name in the Measure Name box (e.g., mNSS) in the Repeated Measures Define Factor(s) dialogue box.
   3. Load the Within-Subject Variables (data acquired at pre-, D1, and D50 post-CHI) that need to be tested and specify the between-subjects factor (e.g., animal groups with different impact parameters) in the Repeated Measures dialogue box.
   4. Select Bonferroni as a Post Hoc Test for Factor(s) (e.g., animal groups) in the post hoc multiple comparisons for Observed Means dialogue box.
      NOTE: To correct for multiple comparisons, the type I error was adjusted using Bonferroni corrections (0.05/3) for the comparisons over time. Statistical significance was defined as p < 0.05 (SPSS adjusted).
3. Conduct a one-way ANOVA analysis to compare the righting reflex and change of cortical volume among the groups.
   1. Click Analyze > Compare Means > One-Way ANOVA.
   2. Load the variables (righting reflex and change of cortical volume) in the Dependent List and the groups as the Factor in the One-Way ANOVA dialogue box.
   3. Select Bonferroni as a post hoc test in the One-Way ANOVA: Post Hoc Multiple Comparisons dialogue box.
      NOTE: To correct for multiple comparisons, type I error was adjusted using Bonferroni corrections (0.05/5) for the comparisons among groups. Statistical significance was defined as p < 0.05 (SPSS adjusted).

Figure 2 shows longitudinal MRIs from representative animal with sham and repetitive CHI at the SMCx. No significant skull fracture or brain contusion was found in T2-weighted images on 1 and 50 days post-CHI. No significant edema or deformation of WM was found in FA maps on 1 and 50 days post-CHI. All animals subjected to CHI in this study survived the entire experimental duration of 50 days, demonstrating low mortality (0-5%)⁷ of the CHI model.

The degree of impairment in consciousness immediately following brain injury was evaluated by the loss of righting reflex, the intrinsic propensity to self-correct its position, of animals. Compared with sham and single CHI at SMCx, the time to regain the righting reflex increased in animals after repetitive CHI (Figure 3A). The general well-being of the animals following CHI was reflected by the change in normalized body weight and mNSS. No significant weight loss was observed after CHI among groups (Figure 3B). While a higher mNSS score was found at day 50 following a single CHI, a significant increase of mNSS score was observed at day 1 after repetitive CHI and maintained high till day 50 regardless of the severity and impact site (Figure 3C). The elevated mNSS induced by repetitive CHI at the central brain decreased at day 50, significantly lower than the corresponding CHI at SMCx. The balance and coordinated motor function in rats after CHI was assessed by the beam walk test. A significant increase in beam walk duration was observed on day 1 after repetitive CHI and maintained high till day 50, regardless of the severity and impact site (Figure 3D). The elongated beam walk duration induced by repetitive CHI at the central brain decreased at day 50, significantly shorter than the corresponding CHI at SMCx.

A significant decrease in the cortical volume was observed on 50 days post-CHI (Figure 4A). The cortical volumes at day 50 were 99.63% ± 2.15%, 95.98% ± 1.65%, 92.26% ± 2.22%, and 90.28% ± 1.17% from the baseline volume, respectively, in the sham and after single and repetitive CHI with the 1 h and 10 min intervals at SMCx (Figure 4B). The cortical volume at day 50 was 91.54% ± 1.98% from the baseline volume after repetitive CHI with the 1-h interval at the central brain. Compared with the sham group, a significant cortical loss was observed after CHI. Compared with the single CHI group, a significant cortical loss was observed after repetitive CHI. A substantial reduction in cortical volume was observed in the slices at Bregma -4 to +0 and Bregma -5 to +1 after repetitive CHI with the 1-h and 10-min intervals, respectively (Figure 4C). Compared between the CHI animals with different impact sites, a significantly smaller cortical volume was found only in the slice at Bregma 0 after CHI at the central brain. While significant cortical atrophy was reported in the previous¹¹ and current studies, T2-weighted images with high spatial resolution, ideally acquired in 3D, are suggested for precise volumetric analysis. In addition, future studies applying an atlas-based diffeomorphic registration approach³⁸ may better address the regional brain changes associated with mild brain injury.

Cortical FA values during longitudinal MRI scans were calculated to indicate the tentative microstructural changes after CHI. After a single CHI at SMCx, no significant FA changes were observed underneath the impact site. After repetitive CHI at the SMCx, a significant increase of ipsi-lesional cortical FA was observed in the cortex at day 50 compared with baseline and 1 day post-repetitive CHI with the 1-h interval (Figure 5A). In addition, a significant reduction of FA in the ipsi-lesional cortex was shown on 1 day post-repetitive CHI with the 10-min interval, which is significantly lower than that after single and repetitive CHI with the 1-h interval. CHI at SMCx did not induce significant changes in FA in the cortex of the central brain (Figure 5B). After repetitive CHI at the central brain, a significant increase of cortical FA underneath the central brain was observed in the cortex at day 50 compared with baseline and day 1 (Figure 5B).

After single CHI at SMCx, no significant changes in FA were observed in the CC underneath the ipsi-lesional SMCx (Figure 5A). After repetitive CHI at the SMCx, significant decrease of ipsi-lesional FA in the CC was observed in the cortex at day 50 compared with baseline and 1 day post-repetitive CHI with the 1-h interval (Figure 5A). Reduction of FA in the ipsi-lesional CC at day 1 and then recovered at day 50 was observed post-repetitive CHI with the 10-min interval. In the ipsi-lesional CC, following repetitive CHI with the 10-min interval, significantly lower FA value at day 1 was shown compared with repetitive CHI with the 1-h interval; significantly higher FA value at day 50 was shown compared with sham, single and repetitive CHI with the 1-h interval. After repetitive CHI at the central brain, significant increase of FA in the CC underneath the ipsi-lesional SMCx was observed at day 1 compared with CHI at SMCx and at day 50 compared with the sham group (Figure 5A).

Neuroinflammation after CHI was assessed by the expression of GFAP at day 50 post-injury. The results from immunostaining demonstrated that astrocytes accumulated in the ipsilesional SMCx after CHI, regardless of severity and impact site (Figure 6).

Figure 1: Schematic of experimental design. Schematics showing the key steps, including the induction of closed-head injury and the corresponding timeline for each assessment. MRI scans and behavioral assessments before CHI were conducted within 7 days before surgery. The time to regain the righting reflex was assessed as the degree of impairment in consciousness. Longitudinal MRI and behavioral data were collected on 1 and 50 days post- CHI. Rats were sacrificed upon completion of all experiments, followed by immunohistology. Abbreviation: SMCx/single = single impact at the sensorimotor cortex; SMCx/2 hits/1 h = double impacts at SMCx with the 1-h interval; SMCx/2 hits/10 min = double impacts at SMCx with the 10-min interval; Central/2 hits/1 h = double impacts at the central brain with the 1-h interval. Please click here to view a larger version of this figure.

Figure 2: Representative MR images following CHI. The T2-weighted images (upper row) and FA maps (lower row) from the representative animal before and at days 1 and 50 after sham and double CHI at SMCx with the 10-min interval. No focal contusion on T2-weighted images after experimental CHI. Please click here to view a larger version of this figure.

Figure 3: Behavioral deficits after CHI with different impact parameters. (A) The time to regain the righting reflex after the last impact. The time of righting reflex increased after repetitive CHI at the SMCx. (B) No significant difference in the normalized body weight after CHI (normalized to the pre-CHI baseline) among groups. An increase in (C) mNSS and (D) Beam-walking duration was observed after repetitive CHI. While mNSS and beam-walking duration remained high after CHI at SMCx, they recovered after CHI at the central brain on day 50. One-way ANOVA with Bonferroni post hoc test for the time of righting reflex; repeated ANOVA with Bonferroni post hoc test for the normalized weight, mNSS, and Beam-waking duration: *, p < .017 between time points; +, p < .05 vs. sham; #, p < .05 vs. SMCx/single; §, p < .05 vs. SMCx/2 hits/1 h. Please click here to view a larger version of this figure.

Figure 4: Cortical atrophy on 50 days post-CHI with different impact parameters. (A) Slice alignment on the middle sagittal image. The blue line indicates the horizontal plane connecting the anterior commissure and the base of the cerebellum; the dashed bline line shows the the long axis of the corpus callosum. (B) Illustrative cortical ROIs (red) overlaid on T2-weighted images in the representative image slices for the measurement of cortical volume. (C) The change of cortical volume after CHI was represented as the percentage of the baseline volume among different slices at Bregma -7 to +3 mm. A decrease in cortical volume on 50 days post-CHI was demonstrated and impact parameter dependent. Data are expressed as the means ± std. One-way ANOVA with Bonferroni post hoc test: +, p < .05 vs. sham; #, p < .05 vs. SMCx/single; §, p < .05 vs. SMCx/2 hits/1 h. Please click here to view a larger version of this figure.

Figure 5: Longitudinal changes in FA after CHI with different impact parameters. Automatically segmented ROIs are the cortex (green) and corpus callosum (CC) (red) deep to the impact site at (A) SMCx and (B) central brain. The inset shows the 3D brain image with the slice underneath the impact site. Longitudinal follow-up of FA values acquired before and on 1 and 50 days post-CHI were presented as mean ± std. Alteration in FA after repeated CHI was prominent and impact parameter dependent. Repeated ANOVA with Bonferroni post hoc test: *, p < .05 between time points; +, p < .05 vs. sham; #, p < .05 vs. SMCx/single; §, p < .05 vs. SMCx/2 hits/1 h. Please click here to view a larger version of this figure.

Figure 6: CHI-induced neuroinflammation on 50 days post-injury in the cortex underneath the impact site. Representative images of the cerebral cortex underneath the impact site with GFAP staining. Accumulation of astrocytes (arrows) in the cortex was observed after CHI. Scale bar = 40 µm. Please click here to view a larger version of this figure.

This study aimed to establish an animal model of uncomplicated mild traumatic brain injury (mTBI) to evaluate the cumulative effects of single and repetitive injuries, as well as the outcomes of impacts on different brain regions. The closed-head injury (CHI) model, adapted from the closed-head weight-drop injury paradigm, was designed to mimic concussions commonly experienced by athletes and individuals with helmet protection. This model minimizes focal brain damage while enabling the precise manipulation of key impact parameters, including the number of impacts, inter-injury intervals, and impact regions. The findings here demonstrated that these parameters significantly influenced the progression of behavioral outcomes and fractional anisotropy (FA) values. Notably, substantial cortical atrophy, a hallmark feature of chronic traumatic encephalopathy (CTE), was observed during the chronic phase, regardless of impact load or location. This experimental model provides a robust framework for longitudinal studies of functional and microstructural changes following uncomplicated mTBI, addressing gaps in previous animal models.

To replicate mTBI observed in clinical scenarios such as contact sports or motorcycle accidents, diverse rodent helmet designs have been implemented across various animal models⁷. The impact to the closed skull or head generally results in milder and more diffuse brain injury compared to those targeting the exposed brain surface^15,39. Nevertheless, it has been acknowledged that a large variability in the outcome among animals was observed when utilizing affixed helmets, majorly due to the inconsistency of impact site location⁴⁰. The CHI model in this study was modified from Marmarou's weight-drop model, wherein a metal disk was placed over the skull⁴¹. We further refined the original methodology by employing a thinner disk (1 mm) and integrating a fixed impactor tip to mitigate the risk of skull fracture. Our previous micro-computed tomography (CT) results corroborated the absence of discernible micro-fractures after CHI¹¹. Another advantage of using a fixed impact tip directed toward the cemented disk is that it facilitates precise control over the impact site, allowing us to systemically probe the effect of the impact site on experimental outcomes. Of note, the scalp incision and anesthesia in the current model may induce additional immune and inflammatory responses, particularly in the acute phase. Using awake and scalp-intact animals could help mitigate these effects and enhance the translatability to clinical cases of subconcussive brain injury¹⁰.

The cumulative effect of CHI on the behavioral and image outcome was demonstrated by the significantly higher mNSS scores, longer duration to complete beam walk tasks (Figure 3), smaller cortical volume (Figure 4B), and alteration in FA values (Figure 5A) in animals following rCHI compared to the sham or single injury groups. Moreover, significantly lower FA values in the cortex and CC at day 1 post-injury (Figure 5A) and reduced cortical volume at day 50 (Figure 4B) were shown in animals subjected to repetitive CHI with a 10-min interval compared to those with a 1-h interval, suggesting worse outcomes with shorter inter-injury intervals. When the repetitive injury was performed with 1-h interval, animals with impacts over the SMCx exhibited higher mNSS scores (Figure 3C) and longer beam-walk durations (Figure 3D) compared to the impacts over the central brain, indicating that the CHI outcomes are impact-site dependent. In addition to alteration in FA, a decrease of AD in WM^10,11,19 and an increase of RD in GM^10,16,18 have been proposed post-CHI. Future research incorporating a comprehensive analysis of the full spectrum of DTI parameters could provide deeper insights into how different impact parameters influence the progression and outcomes of CHI. The proposed model may also be applied to adolescent rats and mice. However, further adjustments, including the height and the weight of the drops, as well as the helmet dimension, warrant exploration and validation in advance.

The righting reflex, an innate animal behavior characterized by the ability to reorient and stand on their feet spontaneously, serves as a surrogate index for assessing loss of consciousness (LOC) in humans⁴². In order to document the time taken to regain the righting reflex after CHI, inhaled anesthetics must be used instead of injectable anesthetics during CHI induction. Moreover, temporary cessation of isoflurane immediately prior to CHI is necessary²⁵. Monitoring changes in body weight post-TBI is recommended to indicate overall impairment⁴³. No significant change in normalized body weight after CHI indicates the mildness of the brain injury in the model described here. Modified NSS and beam walk duration have been widely used to assess general well-being and vestibulomotor function after brain injury⁴⁴. Given that behavioral assessment and MRI experiments were performed on the same day post-CHI, the behavioral tests were conducted before MRI scans for all the follow-up assessments to prevent the interference of anesthesia with the measured behavior outcomes (Figure 1). In addition, the animals exhibiting poor motor coordination, potentially increasing the mNSS score as well, should be excluded based on the pre-CHI test. Our results, in line with the previous study, showed significantly higher mNSS scores and prolonged beam walk duration following repetitive CHI¹¹. Furthermore, we demonstrated that mNSS scores and beam walk duration depend on the impact site of CHI, particularly at day 50 post-injury.

Longitudinal MRI, facilitating the assessment of macro- and mesoscale brain structures over time, represents a crucial tool for validating the fidelity of the CHI model presented here in replicating characteristics of uncomplicated mTBI. During image acquisition, especially on day 1 post-CHI, the physiological parameters, including temperature, respiratory rate, and heart rate of the animal, should be well monitored. Consequently, the concentration of isoflurane should be carefully adjusted in time to maintain physiological stability. While four-shot EPI was employed in the current study for DTI image acquisition, single-shot EPI can also be used to reduce motion artifacts because of the relatively short scan time. Image processing and analysis of preclinical MRI is crucial as most of the studies still rely on custom-made analysis pipelines developed by individual research teams⁴⁵. If the customized algorithm, like Matlab in the current study, is inaccessible, the volume measurement and signal intensity extraction may be accomplished by using ImageJ, an open-source software, for scientific images based on T2-weighted images and FA maps, respectively. For accurate analysis of MRI images acquired at multiple time points, inner-subjects co-registration should be performed first. Given inter-subject variations in brain volume even at the same postnatal ages, normalization of post-injury brain volume to its baseline volume for each subject is essential to delineate cortical atrophy induced by CHI⁴⁶. For FA analysis, the threshold to separate adjacent gray matter (GM) and WM must be performed to eliminate partial volume effects. It is important to note that FA values are influenced by the magnetic field strength⁴⁷ and the number of diffusion gradients employed in DTI⁴⁸. The setting of the FA threshold in the current study may, thus, not be generally applicable to DTI images acquired by using different protocols or MR scanners.

As mTBI, particularly uncomplicated mTBI, is often invisible by conventional neuroimage in the acute phase, research efforts have focused on identifying effective and advanced image markers to capture and provide prognostic information about the subsequent post-injury symptoms^49,50. The heterogeneity of clinical mTBI cases adds further complexity and inconsistency in the data. In this uncomplicated mTBI model, we observed significant micro- and macro-structural changes in imaging alongside measurable behavioral deficits, providing a platform to longitudinally track potential neuroimaging biomarkers following injury. Notably, the impact parameter-dependent changes in both imaging and functional outcomes in the CHI model suggest the possibility of identifying neuroimaging biomarkers sensitive to injury severity and impact parameters. Consistent with the previous findings showing correlations between specific DTI metrics and astrogliosis⁸, future studies examining the relationship among various image traits, microscopic alterations, and functional outcomes may establish promising non-invasive biomarkers for underlying cellular changes and symptom prognosis following mTBI.

In this study, several limitations had to be considered. First, the sample size for each impact parameter group is relatively small (n = 4 per group), and the range of tested impact parameters is limited. Despite the small sample size, we observed significant differences in behavioral measurement, FA values, and cortical volume among CHI groups. Taken together with the previous studies using different impact parameters^8,11, our results support further study in large samples with a broader range of parameters to test. Second, as with most TBI animal studies^7,9, only male rats were used in the current experiments. Recent research has reported sex differences in changes of DTI metrics in the WM following repetitive CHI in mice, highlighting sex-specific responses following brain injury¹⁰. Future studies incorporating both male and female animals will explore the divergent response to CHI impact parameters between the sexes. Finally, while FA changes were observed post-CHI and among different CHI groups, the diffusion signal preprocessing could be further refined. Incorporating more sophisticated techniques, such as eddy current correction, magnetic field bias correction, etc., together with multishell diffusion image¹⁷ may further enhance the sensitivity of DTI signals to detect microstructural damage induced by mTBI.

With the current protocol, we demonstrated the preserved brain structure alongside significant behavioral deficit in the acute phase after CHI. Subsequent analysis revealed notable cortical brain volume loss and altered FA values in the chronic phase. More importantly, the behavioral and neuroimage outcomes depended on the impact parameters used to induce CHI, including the impact number, the inter-injury interval, and the impact site. Compared with the published mTBI models, which mainly focus on behavioral outcomes or neuroinflammation in the brain, this study employed a comprehensive approach encompassing the systemic and whole-brain assessment after CHI. By examination using longitudinal MRI, the CHI model presented preserved structural integrity in the acute phase but pronounced cortical atrophy in the chronic phase, suggesting the successful replication of uncomplicated mTBI. The significance of the study is that it is possible to explore how the varied impact parameters alter the brain after mTBI and to develop tentative image biomarkers for this clinically silent injury.

The authors have no potential conflicts of interest to disclose.

This work was supported by a research grant from the National Science and Technology Council (NSTC) of Taiwan (NSTC 113-2314-B-A49-047).