Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A roadmap of brain recovery in a mouse model of concussion: insights from neuroimaging

A roadmap of brain recovery in a mouse model of concussion: insights from neuroimaging Concussion or mild traumatic brain injury is the most common form of traumatic brain injury with potentially long- term consequences. Current objective diagnosis and treatment options are limited to clinical assessment, cogni- tive rest, and symptom management, which raises the real danger of concussed patients being released back into activities where subsequent and cumulative injuries may cause disproportionate damages. This study conducted a cross-sectional multi-modal examination investigation of the temporal changes in behavioural and brain changes in a mouse model of concussion using magnetic resonance imaging. Sham and concussed mice were assessed at day 2, day 7, and day 14 post-sham or injury procedures following a single concussion event for motor deficits, psycho - logical symptoms with open field assessment, T2-weighted structural imaging, diffusion tensor imaging (DTI), neurite orientation density dispersion imaging (NODDI), stimulus-evoked and resting-state functional magnetic resonance imaging (fMRI). Overall, a mismatch in the temporal onsets and durations of the behavioural symptoms and struc- tural/functional changes in the brain was seen. Deficits in behaviour persisted until day 7 post-concussion but recov- ered at day 14 post-concussion. DTI and NODDI changes were most extensive at day 7 and persisted in some regions at day 14 post-concussion. A persistent increase in connectivity was seen at day 2 and day 14 on rsfMRI. Stimulus- invoked fMRI detected increased cortical activation at day 7 and 14 post-concussion. Our results demonstrate the capabilities of advanced MRI in detecting the effects of a single concussive impact in the brain, and highlight a mismatch in the onset and temporal evolution of behaviour, structure, and function after a concussion. These results have significant translational impact in developing methods for the detection of human concussion and the time course of brain recovery. Background cognitive, physical, and emotional disturbances mani- Concussion, or mild traumatic brain injury (mTBI), is fest within the first 24 h of the injury, and last for several a physical trauma-induced pathophysiological process weeks [2] or longer [3]. More problematic is that certain affecting the brain, resulting in rapid onset of typically physiological disturbances can persist beyond the typi- transient neurological dysfunction, with or without loss cal 2-week window of clinical recovery, raising concerns of consciousness [1]. Concussions are inherently diverse about the super-additive risks associated with repeated in nature and of unpredictable outcome. The cognitive injury incurred while the brain is still recovering from the sequelae from seemingly minor head injuries incurred effects of the first impact. Indeed, growing evidence sug - during sports can be severe and persistent. Conspicuous gests that a concussed individual is at high risk for fur- ther concussion, and that repeated injuries within a short time window can provoke cumulative brain damage [4]. *Correspondence: f.nasrallah@uq.edu.au Objective methods that can accurately diagnose the The Queensland Brain Institute, The University of Queensland, Building impact of a concussion on the brain, allowing for better 79, Upland Road, Saint Lucia, Brisbane, QLD 4072, Australia © The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/publi cdoma in/ zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. To and Nasrallah acta neuropathol commun (2021) 9:2 Page 2 of 20 understanding of the underlying pathology, and track- found in the WM of concussed athletes. Decreased FA ing post-concussion recovery, are thus required. Mag- and increased ODI were found in the optic tracts of a netic resonance imaging (MRI) is a non-invasive imaging mouse model of closed-head injury [21]. method with a number of modalities optimised to detect Studies on the associated integrity of the functional different aspects of the structural and functional integrity connections in the brain, using resting state functional of the brain. The non-invasive nature of MRI also allows MRI, following a concussion have also been limited. repeated imaging measurements, making it ideal for Resting-state functional Magnetic Resonance Imaging tracking the temporal trajectory of the injury. (rsfMRI) detects synchronous low-frequency fluctua - Diffusion Tensor Imaging (DTI) is an MRI modality tions (< 0.3 Hz) of blood oxygen level dependent (BOLD) that probes the motion of water molecules in biological signals among spatially distinct brain regions in subjects tissue [5, 6]. DTI uses a number of metrics to describe at rest [33]. Regions that have synchronous oscillating water diffusivity in a voxel, namely Fractional Anisotropy fluctuations form brain networks that are functionally (FA) [7] parallel diffusivity (Dp), radial diffusivity (Dr), connected. The default mode network (DMN) has been and mean diffusivity (MD) [8]. DTI has been used to the major functional network disrupted after concussion study human concussions and white matter pathologies, in the absence of structural deficits [34, 35]. Zhou et  al. in particular, though the results have often been contra- reported reduced functional connectivity in the poste- dictory [9–15]; both reduced FA coupled with increased rior cingulate cortex and parietal regions and increased MD [9–12] and increased FA coupled with decreased MD functional connectivity around the medial prefrontal cor- [13–15] have been reported in the white matter follow- tex in mTBI patients on average 22 days post-concussion ing mTBI. A 2014 meta-analysis of 122 studies suggested [34]. The decreased connectivity in the posterior DMN that following a concussion, changes in FA are time- was positively correlated with cognitive deficits (lower dependent—FA increases during the acute phase then ability to rapidly switching between cognitive sets) while decreases in the subacute phases following the impact the increased connectivity in the frontal DMN was cor- [16]. Repeated mTBI in rodents [17, 18], single-impact related negatively with posttraumatic depressive symp- drop-weight model in rats [19] and single-impact piston- toms [34]. Similarly, Johnson et  al. detected a reduction driven closed-head injury in mice [20, 21] have shown in the number and strength of the connections in the decreased FA in the white matter. A couple of other stud- posterior cingulate cortex and lateral parietal cortices, ies in rats have shown opposite results: increased FA in and an increase in the number and strength of connec- the white matter [22, 23]. DTI changes in the grey mat- tions in the medial prefrontal cortex in concussed ath- ter (GM) received less attention though increased FA has letes imaged after symptoms had resolved, on average been found in human patients with persistent post-con- 10  days post-concussion [35]. Other networks found cussive symptoms (PPCS) [24] and in vivo animal models with decreased functional connectivity after concus- of traumatic brain injury (TBI) [25–27]. sion include: the Salience Network (SN) [36, 37], in the Neurite orientation dispersion and density imaging lateralised cognitive control network [37], and in regions (NODDI) is yet another approach that provides more related with motor, sensorimotor, attention, and phono- specificity to diffusion imaging changes [28]. NODDI logical processing [36]. Connectivity strength between provides measures of neurite (axons and dendrites) the left dorso-lateral prefrontal cortex and left lateral density and local structural organisation of neurites parietal cortex was reduced with an increasing number by measuring the level of orientation dispersion of of concussions [35]. More recently, Kaushal et  al. inves- restricted anisotropic diffusivities measured with mul - tigated resting-state functional connectivity in concussed tiple diffusion-weighted directions at higher b-values athletes and found that at 48 h post-injury, there were no using orientation dispersion index (ODI) and the con- changes in functional connectivity, despite psychological tribution fraction of restricted anisotropic diffusion to distress, and oculomotor, balance and memory deficits the total diffusion using neurite density index (NDI) [38]. At 8 days post-injury, a global increase in functional [28, 29]. The few NODDI studies applied to map brain connectivity was seen with improving symptoms, which changes in human concussion have reported different recovered by 15  days post-injury [38]. Similarly, Meier findings. Wu et  al. detected decreased NDI without et  al. reported increased local intrinsic functional con- any DTI changes in the white matter (WM) of concus- nectivity in the right middle and superior frontal gyri at sion patients aged 35  years old approximately 2 weeks 24–48  h post-injury, which were normalised at 7  days post-injury [30]. Similarly, both decreased NDI (asso- and 6  months post-medical-clearance [39]. To the best ciated with decreased FA, and increased Dp and Dr) of the authors’ knowledge no functional imaging study [31] and increased NDI (associated with decreased combining stimulus-evoked functional MRI and rsfMRI ODI, increased FA and decreased MD) [32] were were published in a mouse model of closed-head injury. T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 3 of 20 The conflicting findings in human imaging studies support structure included a reclining body plate to sup- described above suggested the difference might be due to port the mouse’s body and a head plate to support the the patients’ background (sex, age, medical history), the head. The head plate had a hole to allow a brass piston injury severity, and the imaging time post-concussion, to deliver the impact from below; two perpendicular lines which affect where each patient was along the injury/ intersecting the centre of the brass piston were used as recovery time course. Therefore, animal models serve as crosshairs to aim and position the head. The body plate a way to study concussion by allowing precise control of was reclined to allow the frontal and parietal bones on all the variables or their effects on the injury and mani - the mouse’s head to be positioned flat against the head festation on MRI findings. In this study, we attempted to plate. The mouse body was secured in the supine posi - chart the temporal evolution of the microstructural and tion by a non-slip silicone mattress and two Velcro straps functional changes post-concussion in a mouse model across the chest and abdomen. The tail was secured to with the aim of highlighting the functional-structural the body plate with masking tape to prevent slipping. mismatch in their recovery trajectories. Anaesthesia was maintained during the securing and positioning for impact at 2% isoflurane in the same gas Methods mixture via a nose cone; total time under anaesthesia was Study design 10–12  min. Time under anaesthesia was kept consist- This is a cross-sectional study which included a total ent across different groups and the experimenter aligned of 43 3–4  months old (mice age on impact date: all animals, concussed or shams, to the same accuracy 13.2 ± 1.4 weeks) male mice. Mice were divided into four standard. cohorts: Sham (n = 14, n = 6 day 2, n = 3 day 7, n = 5  day Once the animal was secured on the platform, anaes- 14), concussion day 2 (CON 2; n = 9), concussion day 7 thesia was discontinued, and the trigger button imme- (CON 7; n = 10), concussion day 14 (CON 14; n = 10). On diately pressed to induce a concussive impact. Terminal day 0, all mice in the concussion group were exposed to piston velocity used ranged from 5.2 to 5.4 m/s. Mice in a concussive impact using our impactor device (see our sham control group underwent the same procedure with- earlier study [40] for detailed description). The sham out actual impact. animals underwent the exact same procedure but did not receive an impact. After the concussion or sham Behavioural assessment procedure, the loss-of-righting-reflex (LRR) time was Loss‑of‑righting‑reflex (LRR) measured for each of the animals. Thirty minutes after After the impact or sham procedure, the mice were recovery, the Neuro Severity Score (NSS) was measured removed from the restraints and laid in the supine posi- for each of the mice. At day 2, day 7, or day 14, depend- tion on a warmed surface for recovery. Loss-of-righting- ing on the cohort, the NSS was measured again and all reflex (LRR) time was the time (s) from the moment of mice underwent Open Field Assessment and MRI scan. impact to the first sign of the animal, righting itself to a Animals were excluded from this study if obvious brain prone position. injuries or structural abnormalities were observed on T2-weighted structural MRI images. All experiments Neuro Severity Score assessment were approved by the Institutional Animal Ethics Com- Thirty minutes after the first sign of righting reflex, all mittee at the University of Queensland (Animal Ethics mice were assessed based on a modified Neuro Severity Committee approval number QBI/260/17). Score (NSS). Another NSS assessment was performed on Data from this study is available, without reservations, the day of and prior to the MRI scanning session. NSS on request to the corresponding author. is a common assessment 10 tasks scale for neurological deficiency and a detailed description of the tasks can be Concussion procedure found in Flier et  al. [41]. These 10 tasks included: suc - The concussion model used in this study was used in one cessful escape from a 30  cm-diameter walled circle with of our prior publications [40], where more details can be one opening; natural seeking behaviour in an open area; found; the procedural description is mentioned for clar- lack of limping/dragging walking gait or grabbing weak- ity and ease for the readers. ness; presence of a straight walk and lack of an abnormal The animals were housed in the animal holding facil - gait; intact startle response to loud noise; 1 cm beam bal- ity with a 12-h light–dark cycle, with food and water ance; 3, 2, and 1 cm beam walk tasks, and 0.6 cm round available ad libitum. Animals were initially anaesthetised stick balance. An additional 6  mm beam walk task was with 3% isoflurane in 60% Air 40% O gas mixture at 2 L/ added, raising the number of task and the highest possi- min for 2 min. Mice were then transferred to the impac- ble score from 10 to 11. The final task described in Flier tor device and supported on the body plate. The animal et  al. was the round stick balance [41]; preliminary tests To and Nasrallah acta neuropathol commun (2021) 9:2 Page 4 of 20 showed several of our mice were sufficiently skilled and 0.1 × 0.1 × 0.3  mm, Repetition Time (TR) = 7200  ms , motivated to cross the stick similarly to beam walk tasks; Echo Time (TE) = 39 ms , averages = 4, RARE factor = 8, thus we decided to add a 6 mm round stick walk into our and bandwidth = 54.3478 kHz . modified NSS assessment. Resting‑state fMRI scans Open field assessment a 2D gradient-echo echo-planar-imaging (GE-EPI) Open field activity was assessed on the day of and prior sequence with the following parameters was used: to the MRI scanning session using an elevated circular matrix size = 64 × 64, FOV = 19.2 × 19.2  mm, 20 arena (diameter 77  cm) fenced by 32  cm high bound- slices of 0.5  mm thickness and 0.1  mm slice gap; giv- ary. The animals were placed in the centre of the arena ing an effective spatial resolution of 0.3 × 0.3 × 0.6  mm, and spontaneous activities were tracked and recorded TR = 1000  ms , TE = 14  ms , averages = 2, flip using an overhead camera and Tracker software (Bio- angle = 70°, bandwidth = 200  kHz, and 600 volumes Signal Group, NY, USA) for 10  min. The circular plat - were acquired with fat suppression, FOV satura- form was divided into four annuli (with annulus 1 at the tion (covering the head tissue inferior to the subject’s centre and annulus 4 at the periphery of the platform) brain), and navigator pulses turned on. A block-stimu- the fraction of total time the animal spent in each annu- lation design (40  s OFF, 20  s ON) fMRI scan was per- lus was quantified. The Thigmotaxis Index (TI)—an formed approximately 70 min after the initiation of the index for anxiety-like behaviours [42], was calculated as medetomidine bolus; stimulation fMRI scan contained TI = (T − T )/(T + T ) where T and ann4 ann1-3 ann4 ann1-3 ann4 six OFF–ON blocks plus a final 40  s OFF block, for a T represents the time spent in annulus 4 and com- ann1-3 total of 400  s. After the end of the stimulation fMRI bined time spent in annuli 1, 2, and 3, respectively. scan, a 5  min break was given before rsfMRI scan was started; rsfMRI scan duration was 600 frames for a total MRI experiments of 600  s. A single EPI scan with the same parameters Animal handling but opposite phase-encoding direction was acquired Anaesthesia was induced using 3% isoflurane in 60% for distortion correction of the fMRI images. air, 40% O at 1L/min. The isoflurane concentration was maintained at 2–2.5% during the preparation which took around 30 min. Each mouse was positioned on an MRI- Diffusion MRI scans compatible cradle (Bruker Biospin, Germany) with ear Diffusion data were acquired using a diffusion-weighted bars and bite bars to reduce head motion. A peritoneal Imaging (DWI) spin-echo echo planar imaging (SE- catheter was inserted and fixed to the mouse for delivery EPI) sequence with the following parameters matrix of medetomidine (Domitor, Pfizer, USA). For sedation, a size = 96 × 96, FOV = 19.2 × 19.2  mm, 32 slices of bolus of 0.05 mg/kg medetomidine was given intraperito- 0.25 mm thickness and 0.05 mm slice gap; giving output neally and then sedation was maintained with a continu- spatial resolution of 0.2 × 0.2 × 0.3  mm, TR = 4500  ms , ous infusion of 0.1 mg/kg/h. Once the animal was inside TE = 25  ms , averages = 4, 3 b-value shells with b = 600, the MRI scanner, isoflurane was then reduced gradually 1500, and 2000s/mm , 33 diffusion weighted directions and kept at approximately 0.25% throughout the experi- for each shell, and 2 b = 0 images. A pair of reference ment. The total time under anaesthesia for each animal b = 0 SE-EPI scans were acquired with opposite phase- was approximately 2.5 h. At the end of the scanning ses- encoding directions for EPI distortion correction. sion, 1.25  mg/kg atipemazole (medetomidine reversal) (Antisedan, Pfizer, USA) was given intraperitoneally. Behavioural data analysis Structural MRI scans MRI scans were performed All statistical analyses of behavioural data were per- on a 9.4T MRI scanner (Bruker Biospin, Germany) formed in Prism 8 (GraphPad Inc.). Statistically sig- equipped with a cryogenically cooled transmit and nificant threshold was set at p value < 0.05 (two-tailed). receive coil, controlled by a console running Paravision Analysis of Variance (ANOVA) post hoc tests were cor- 6.0.1 (Bruker Biospin, Germany). Structural imaging rected for multiple comparisons by controlling False data was acquired using a 2D T2-weighted (T2w) Turbo Discovery Rate (FDR). Rapid Acquisition with Refocused Echoes (Turbo- RARE) sequence with the following parameters: matrix Loss‑of‑righting‑reflex (LRR) size = 192 × 192, Field of View (FOV) = 19.2 × 19.2 mm, LRR duration was compared between sham animals number of contiguous slices = 52 and slice thick- and all injured animals using Mann–Whitney tests. ness = 0.3  mm; giving an effective spatial resolution of T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 5 of 20 Neuro Severity Score (NSS) affine-transformation-included Jacobian determinant The NSS score was compared between the sham and all was calculated from the subject-specific non-linear concussed mice at 30  min post-concussion using the registration warp; Jacobian determinant for each voxel Mann–Whitney rank test. The NSS of the concussed is a measure that indicate the relative volume change mice from the same cohort assessed at 30  min was required to warp the template voxel to the individual compared against the NSS on the day of the MRI scan voxel. Tensor-based morphometry (TBM) [46] was of that cohort using the Wilcoxon matched pairs signed performed using voxel-wise statistics on the Jacobian rank sum test. NSS of shams and CON cohorts assessed determinants to assess of local structure absolute vol- on imaging day were analysed using Kruskal–Wallis ume differences between groups, as opposed to the One-Way ANOVA with post hoc tests comparing each analysis of local structures normalised by total brain CON cohort with sham cohort. volume. Open field assessment Stimulus‑evoked and resting‑state fMRI data processing Time spent in each annulus of different cohorts were Functional MRI image registration Opposite phase- analysed using repeated measures Two-way ANOVA encoding direction EPI data were used to calculate the with Geisser–Greenhouse correction and post hoc warping field required for EPI distortion correction using tests comparing mean time spent in each annulus of FSL’s TOPUP tool [47, 48]. One distortion corrected EPI each cohort with every other cohort. TIs were analysed data was used as reference for an iterative direct EPI-to- using Brown–Forsythe and Welch One-way ANOVA EPI non-linear registration/template generation image with post hoc tests comparing mean TI of each cohort registration process. All subject’s distortion corrected with mean TI of every other cohort. EPI images were affinely registered to the chosen EPI reference using FSL’s FLIRT [45]; the registered images MRI data processing were then averaged to create the first iteration of a study- All MRI data were exported in DICOM format using specific EPI template. Each subject’s distortion corrected Paravision 6.0.1 before converting to NIFTI data for- EPI images were then non-linearly registered to the new mat using MRIcron (i.e., the dcm2nii tool) [43]. For the study-specific EPI template using FSL’s FNIRT tool [45]. remaining of the processing, the images were given a The non-linearly registered EPI images were averaged to header file with voxel size 20 times larger than the orig - create the second iteration study-specific EPI template. inal size [44]. The non-linear registration process was repeated one more time and the warping field from this iteration used Structural image processing and tensor‑based morphometry to warp the rsfMRI image to a common space. Recent (TBM) studies suggested that for the purpose of image regis- Structural image signal inhomogeneity correction was tration for functional MRI studies, traditional structural applied to the T w structural data using FSL’s FAST image-based step-wise registration of functional image tool (https ://fsl.fmrib .ox.ac.uk/fsl/fslwi ki) [45]. The to same-subject structural image then to common space structural image was then registered rigidly to the Aus- template can be replaced by direct faster EPI-to-EPI non- tralian Mouse Brain Mapping Consortium (AMBMC, linear registration without loss of accuracy [49]. www.imagi ng.org.au/AMBMC ) MRI template resam- pled to 0.1  mm isotropic resolution using FSL’s FLIRT Functional data pre‑processing Each functional data set [45]. The registered images from all animals were aver - underwent slice-timing correction with FSL’s slicetimer aged to create a study-specific “FLIRT template”. This tool, despiked using AFNI’s 3dDespike (https ://afni.nimh. study-specific template was used as the new template nih.gov/) tool, motion corrected using FSL’s MCFLIRT for another iteration of linear registration and “FLIRT tool (FMRIB’s Software Library) [45] with the forward template” creation. This process was repeated for three phase EPI image as the reference image. After motion cor- linear registration iterations before each individual rection, functional images were distortion corrected using structural data was then non-linearly registered to the the warping field obtained from EPI pair. Band-pass filter resulting third iteration study-specific “FLIRT tem - was applied on resting-state fMRI data at 0.01–0.3  Hz plate” using FSL’s FNIRT tool [45] and a study-specific using AFNI’s 3dTproject. Pre-processed stimulus-evoked “FNIRT template” was created from the non-linear reg- and rsfMRI data were warped to a common space using istered images. Individual structural images were non- warp field obtained from the iterative direct EPI-to-EPI linearly registered to the “FNIRT template” and the non-linear registration/template generation image regis- Jacobian determinant was extracted. Each individual tration process described above. To and Nasrallah acta neuropathol commun (2021) 9:2 Page 6 of 20 Functional MRI artefact removal Artefact removal sorted and functionally-relevant components were cho- from registered and pre-processed stimulus-evoked sen based on the following criteria: (1) component spatial and rsfMRI data were performed with group informa- maps clustered on spatially and functionally feasible grey tion guided independent component analysis (GIG-ICA) matter, (2) frequency power spectrum showed power [50], as implemented in the Group ICA of fMRI Toolbox largely in the 0.01–0.1 Hz range (or at least the power in (GIFT) [51]. The number of independent components for this frequency range was larger than those above 0.3 Hz), GIG-ICA was set to 50. GIG-ICA has been shown to be (3) component spatial maps had little or no overlap with superior to not performing artefact removal, or artefact white matter (WM) and cerebral-spinal fluid (CSF), (4) removal using individual-level Independent Component component spatial maps with bilateral or midline-only Analysis (ICA) followed by group ICA—a method recom- areas, or if an IC’s spatial map was unilateral; it was only mended the Human Connectome Project [52]—for arte- included if there was another contralateral IC, (5) ICs facts removal in the context of group ICA [53]. GIG-ICA were rejected as “functionally relevant” if the spatial map also avoids the need for potentially subjective, biased, and had a thickness in the rostral-caudal direction equivalent error-prone human classification of individual level ICA to one EPI slice in the acquired resolution (single-slice [53] or machine learning training of human-classified artefacts). data for automated signal/artefact classification. Artefac - Twenty four ICs were identified and grouped into nine tual components were classified based on criteria simi - anatomical/network groupings; list of component num- lar to prior publications [54, 55]: (1) component spatial bers, component IDs/abbreviations, approximate ana- map having a large overlap with white matter (WM) and tomical structures, and their corresponding anatomical/ cerebral-spinal fluid (CSF) or ring-like or crescent shape network groupings are listed in Table  1. More complete around the edge of the brain or near regions with EPI dis- spatial maps, time courses, and frequency power spectra tortion, (2) frequency power spectrum showing pan fre- for each of the group-level “functionally relevant” ICs are quency distribution, (3) component spatial maps showing shown in Additional file 1: Data 1. alternating positively and negatively correlation bands. Regional homogeneity (ReHo) and intrinsic local functional Independent component analysis (ICA) of functional MRI connectivity analysis Intrinsic local functional con- data Artefact cleaned stimulus-evoked fMRI was ana- nectivity were calculated for GIG-ICA artefact-cleaned lysed using temporal-concatenated spatial group ICA, rsfMRI data for all subjects using the regional homoge- using Infomax algorithm [56, 57], with the number of neity (ReHo) approach [64]. Seven-voxels neighbourhood components set to 18. An independent component with regional homogeneity (ReHo7) maps of each subject’s spatial maps (converted to Z-scores and thresholded at rsfMRI data were created by calculating Kendall’s coeffi - Z > 1) that includes the primary somatosensory area (S1) cient of concordance (KCC) of each voxel’s time course that was shown to be activated in prior fMRI publication with its six face-neighbour voxels using AFNI’s 3dReHo [58], and component time course that matched the stimu- tool. lation paradigm was chosen for further analysis. Artefact-cleaned rsfMRI data were decomposed into Diffusion MRI data processing resting-state functional independent components using Diffusion tensor imaging (DTI) data processing The Independent Vector Analysis (IVA), implemented as opposite phase-encoding direction EPI images of the combined IVA algorithm (IVA-GL) with multivariate diffusion data were used to calculate the warping field Gaussian (IVA-G) [59] source component vectors plus required for EPI distortion correction using FSL’s TOPUP IVA with Laplace source component vectors (IVA-L) tool [47]. This EPI distortion correction warping field was [60]. IVA-GL has been shown to be superior in identify- applied to diffusion data, which was then eddy current ing subject variability and detecting unique biomarkers distortion corrected and motion corrected using FSL’s in fMRI data, compared to more traditional group ICA eddy_correct. Diffusion tensor data was fitted using FSL’s [61–63], which is potentially more suitable more detect- DTIFIT tool using b-value = 1500 s/mm . ing subtle changes in the post-concussion brains. IVA-GL One distortion corrected individual-level FA image simultaneously decomposes artefactual-cleaned individ- was used as reference for an iterative direct EPI-to-EPI ual rsfMRI data into 100 Independent Components (ICs) linear and non-linear registration/template generation spatial maps and time courses per subject then calculates image registration process similar to that of rsfMRI EPI group-averaged spatial maps and time courses. Group data described above (“Stimulus-evoked and resting- averaged ICs spatial maps were scaled and converted state fMRI data processing” section). Warping fields to Z-scores and thresholded at Z > 1 and the ICs were obtain from FA to study-specific FA template FNIRT T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 7 of 20 Table 1 Resting-state functional connectivity supra-networks, network, and  components: list of  component numbers, component IDs/abbreviations, approximate anatomical structures, and  their corresponding anatomical/network groupings and functional supra-networks Supra-network Network group Abbreviations IC no. Anatomical areas Default mode–sensory–memory Anterior Cingulate—Retrosplenial 03—ACA-PL 03 Anterior Cingulate Area + Prelimbic (DMSM) Cortex axis (ACA RSN) area 06—RSN-ACA 06 Retrosplenial area + Anterior Cingu- late Area 52—RSN 52 Retrosplenial area 62—ACA 62 Anterior Cingulate Area 82—RSN 82 Retrosplenial area 89—IL 89 Infralimbic area Hippocampal—Subcortical 05—rHP—RSN—ATN—ACA 05 Subcortical spatial memory pathway memory circuit (HP MEM) (Retrohippocampal Area—Retro- splenial Area—Anterior Thalamic Nuclei—Anterior Cingulate Area) 08—DG 08 Dentate Gyrus 34—ATN 34 Anterior Thalamic Nucleus Visual—Auditory ( VA) 17—AUD 17 Auditory Area 46—VIS 46 Visual Area Primary Somatosensory Area (S1) 04—S1lwlmb 04 Primary Somatosensory Cortex (S1) (Lower Limbs, Trunks) + Posterior Parietal Association Area + Tempo- ral Association Area 11—S1uplmb 11 Primary Somatosensory Cortex (Upper Limbs) 31—S1u 31 Primary Somatosensory Cortex (unas- signed) + Visual Area Thalamus-polymodal association 10—TH-pmc-S1 10 Thalamic nucleus (polymodal cortex ( TH-pmc) association cortex) + Primary Somatosensory Cortex 30—TH-pmc-VA 30 Thalamic nucleus (polymodal association cortex) + Pretectal Region + Visual-Auditor y Area Striatal–motor (STR–MO) CPu 07—CPu 07 Caudate Putamen M1 14—M1 14 Primary Motor Cortex Salience–supplementary soma- Salience Network (SN) 13—S1BF 13 Primary Somatosensory Cortex (Bar- tosensory (SAL–SS) rel Field, Lower Limbs) + Primary Motor Area 37—INS-EP 37 Insula + Endopiriform Nucleus 67—AM 67 Amygdala S2 02—S2-a 02 Supplemental Somatosensory cortex (S2)—anterior (bi lateral), Primary Somatosensory Cortex (Nose, Mouth) 18—S2-p-R 18 Supplemental somatosensory cor- tex—Right 45—S2-p-L 45 Supplemental somatosensory cortex—Left registration were used to warp other DTI metric images proje cts/noddi _toolb ox) [28, 29]; to calculate NODDI to the common space: MD, Dp, and Dr. metrics. In this in  vivo NODDI data fitting, neurites were modelled as impermeable sticks (cylinders with Neurite orientation dispersion and density imaging data zero radius) in a homogeneous background, neurite ori- processing Multi b-value shell data were fitted using entation distribution was modelled as Watson’s distribu- NODDI MATLAB Toolbox (https ://www.nitrc .org/ tion, and the algorithm estimate the hindered diffusiv - To and Nasrallah acta neuropathol commun (2021) 9:2 Page 8 of 20 ity from the free diffusivity, and neurite packing density for further analysis. Another three ROIs were identified using Szafer et  al.’s [65] tortuosity model for randomly to ReHo7 KCC significantly correlated with TI through packed cylinders. NODDI metrics were also warped to voxel-wise correlation analysis: the Caudate Putamen the common space using FA to study-specific FA tem - (CPu), the Amygdala (AM), and the Insula (INS). Simple plate FNIRT registration warp field. Registered DTI and linear regression between DTI/NODDI metrics and KCC NODDI metrics images were smoothed to an estimated quantified from the corresponding ROIs and TI were spatial smoothness of 0.6  mm FWHM using AFNI’s performed on the basis of sham + CON day 2 and CON 3dBlurToFWHM. day 7 + CON day 14, and thresholded at p value < 0.05. T-statistics of two-samples t tests of NSS and TI behav- Statistical analysis of MRI data ioural measures and averaged whole-brain t-statistics Two sample statistical inference of CON cohorts versus sham maps of two samples t tests of Jacobian determinants, cohort DTI/NODDI metrics, and ReHo7’s maps were quantified Voxel-wise Two-samples t tests comparing CON cohorts for each CON cohort comparison with sham cohort and against sham cohort were performed on registered Jaco- the relative behavioural versus MRI changes were plotted bian determinants, DTI and NODDI metrics, S1 inde- for comparison. pendent component spatial maps, and ReHo7 KCC maps using permutation inference for the General Lin- Results ear model [66] as implemented in FSL’s randomise [67]. Behavioural measures The number of permutation was set to 10,000, as recom - Loss‑of‑righting‑reflex (LRR) mended by a prior study [68]. The resulting statistical A significant increase in the LRR time was seen in the map were corrected for multiple comparisons with mass- CON groups 125.6 ± 28.81 (mean ± standard error of based FSL’s Threshold-free Cluster enhancement (TFCE) mean [s]) compared to the sham group 34 ± 2.65 (p [69] and thresholded at p value < 0.05 (two-tailed). value = 0.0005) (Fig. 1a). Reproducibility of DTI and NODDI were tested by ran- domly dividing the sham cohort into two pseudo-groups Neuro Severity Score (NSS) (Fig. 1b) and two-samples t test comparing the pseudo-groups At 30  min post-concussion, all concussed animals had were performed using identical procedure as describe significantly higher NSS 1.68 ± 0.11 compared to sham above. The resulting statistical maps are shown in Addi - animals 0.62 ± 0.13 (p value < 0.0001). NSS of CON day tional file 3: Fig. S3. 2 on the imaging day was significantly higher than that One and two-samples statistical tests were conducted of the sham cohort (1.44 ± 0.24 vs. 0.29 ± 0.12; FDR’s q on individual level IC time courses using GIFT’s Man- value = 0.002). For CON day 7 and CON day 14, there covan toolbox [70] for IC–IC functional network con- was no significant difference in the NSS score at day 7 and nectivity (FNC) averages for each cohort and differences day 14 post-injury compared to that of the sham cohort; between CON and sham cohorts. T test for component– CON day 7 (0.7 ± 0.34; FDR’s q value = 0.248) and CON component FNC were performed using permutation day 14 (0.6 ± 0.22; FDR’s q value = 0.248). There was no inference for the General Linear model [66] as imple- significant difference between the NSS measured at day mented in FSL’s randomise package [67], with the num- 2 (1.44 ± 0.24) or day 7 (1.44 ± 0.24) compared to the ber of permutation set to 50,000 or exhaustive, whichever 30 min NSS in the CON day 2 (1.67 ± 0.17, p value = 0.35) is smaller, and corrected for multiple comparisons with and CON day 7 cohort (1.6 ± 0.3, p value = 0.0625). A sig- False Discovery Rate at q value < 0.05 for one-sample t nificant difference between the NSS measured at day 14 test and q value < 0.1 for two-sample t test. (0.6 ± 0.22) and the 30  min of the CON day 14 cohort (1.9 ± 0.18, p value = 0.0039) was seen. Correlation of behavioural symptoms with local tissue volumes and DTI and NODDI metrics Open field test Registered DTI/NODDI metrics, and ReHo7 KCC maps Two-way ANOVA showed significant annuli (p were correlated with NSS and TI obtained on scanning value < 0.0001) and annuli × cohort effect (p value = 0.03) day, across all subjects. Implementation of voxel-wise sta- but an insignificant cohort effect (p value = 0.2) on tistics were similar to two-sample t tests described above. the fraction of time spent in each annulus. CON day Three regions of interest (ROIs) were identified to have 7 cohort animals spent significantly more time in the DTI and NODDI metrics significantly correlated with TI annuli towards the centre compared to CON day 14 ani- through voxel-wise correlation analysis: the corpus cal- mals and significantly more time in annulus 2 than sham losum (cc), the polymodal association area of the thala- animals (Fig.  1c). There was significant difference of TIs mus (TH-pmc), and the Anterior Cingulate Area (ACA) among different cohorts (p value = 0.022) and CON day T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 9 of 20 Fig. 1 Behavioural assessment of sham and CON cohorts: a LRR time of sham (n = 14) and CON (all 3 CON cohorts, n = 29), two-tailed Mann– Whitney test. Lines showing median and interquartile range. b NSS of experimental cohorts at 30 min post-concussion and on imaging days: sham (n = 14), CON day 2 (n = 9), CON day 7 (n = 10), and CON day 14 (n = 10). Sham versus all CONs at 30 min: two-tailed Mann–Whitney test. Intra-cohort, 30 min versus imaging day: two-tailed Wilcoxon matched-pairs signed rank test. Inter-cohort NSS on imaging day: Kruskal– Wallis One-Way ANOVA with FDR-corrected (q < 0.05, two-tailed) post hoc tests. Data point displayed as mean and standard errors of means. c Proportion of time spent in each of four annuli in Open Field Test. Repeated Measures Two-way ANOVA with Geisser–Greenhouse correction and FDR-corrected (q < 0.05, two-tailed) post hoc tests. d Thigmotaxis Index of each cohort quantified from Open Field Test on imaging day. Brown– Forsythe and Welch One-way ANOVA with FDR-corrected (q < 0.05, two-tailed) post hoc tests. *P value < 0.05, **P value < 0.01, ***P value < 0.001 7 cohort to have significantly lower TI compared to sham in the internal capsule (Fig.  2b, Dp). Increased NDI was (FDR’s q value = 0.03) and CON day 14 cohorts (FDR’s q detected in the grey matter—the visual, auditory, retro- value = 0.0009) (Fig. 1d). splenial, and S1 areas, and the hypothalamus (Fig.  2b, NDI). Decreased ODI was detected in both grey matter— MRI findings the visual, auditory, retrosplenial, hippocampal, and S1 Diffusion MRI findings areas, the striatum, and the anterior cingulate cortex— No significant changes of brain tissues’ DTI and NODDI and white matter—the external capsule (Fig. 2b, ODI). metrics were detected at day 2 post-concussion com- At day 14 post-concussion, DTI and NODDI changes pared to shams (Fig.  2a). At day 7 post-concussion, were still although to a lesser extent compared to day increased FA was detected at day 7 post-concussion com- 7. Increased FA was detected in both grey matter—the pared to shams in both grey matter—the visual, auditory, auditory area and anterior cingulate cortex—and white retrosplenial, hippocampal, and Primary Somatosensory matter—the external capsule (Fig.  2c, FA). Decreased Area (S1), the thalamus, the hypothalamus, the stria- MD, driven by both Dp and Dr decreases was found tum, and the anterior cingulate cortex—and white mat- in the S1, striatum, and internal capsule (Fig.  2c, MD, ter—the external capsule (Fig.  2b, FA). This increased Dp, and Dr). Increased NDI was detected over the same FA in the grey matter was driven mostly by reduced MD regions that had decreased MD, Dp, and Dr (Fig.  2c, and Dr, in the visual, auditory, retrosplenial and S1, the NDI). Decreased ODI were detected in small areas in thalamus, the hypothalamus, and the anterior cingulate the hippocampus, external capsule and corpus callo- cortex (Fig. 2b, MD and Dr). Decreased Dp was detected sum (Fig. 2c, ODI). To and Nasrallah acta neuropathol commun (2021) 9:2 Page 10 of 20 Fig. 2 DTI and NODDI metric changes from day 2 to day 14 post-concussion. Voxel-by-voxel statistical analysis results of Diffusion Tensor Imaging (FA = Fractional Anisotropy, MD = Mean Diffusivity, Dp = Parallel Diffusivity, Dr = Radial Diffusivity) and Neurite Orientation Dispersion and Density Imaging metrics (NDI = Neurite Density Index, ODI = Orientation Dispersion Index) and Tensor-based Morphometry with Jacobian Index (JI) of a CON day 2 (n = 9) versus sham (n = 14), b CON day 7 (n = 10) versus sham (n = 10), and c CON day 14 (n = 10) versus sham (n = 14). Statistical map thresholded at P value < 0.05 (two-tailed), unpaired two sample t test, implemented as permutation tested for the General Linear Model, corrected for multiple comparisons with mass-based FSL’s Threshold-free Cluster enhancement ( TFCE). Statistical maps were overlaid on the averaged and registered DTI and NODDI metrics maps corresponding to the statistical maps (DTI and NODDI results) and structural template ( TBM results). Corresponding grey scale map for each averaged DTI and NODDI metrics maps were provided; units for Dp, Dr, and MD were in mm/s . ACA = Anterior Cingulate Area, AM = Amygdala, AUD = Auditory Area, cc = corpus callosum, ec = external capsule, HP = Hippocampus, HYP = Hypothalamus, ic = internal capsule, INS = Insula, MB = Midbrain, PAL = Palladium, S1 = Primary Somatosensory Cortex, RSN = Retrosplenial Area, STR = Striatum, TH = Thalamus, VIS = Visual Area. Red anatomical orientation marker L = Left, R = Right. An enlarged view of the DTI and NODDI metrics and the identified regions of interest (ROIs) can be found in Additional file 3: Fig. S3 Structural MRI findings of the S1 IC spatial maps was detected at day 7 and 14 TBM detected increased ventricle volumes and local tis- post-concussion (Fig.  3b, c), indicating increased neural sue volumes in the auditory, visual, and primary soma- responses to somatosensory stimulation. tosensory areas, the hippocampus, anterior cingulate cortex, and the amygdala at day 2 post-concussion com- Resting state functional network connectivity architec‑ pared to shams (Fig. 2a, JI). At day 7, a decrease in tissue ture Three main supra-networks based on the con - volume was seen in the midbrain, the hypothalamus, the nectivity architecture: default mode–sensory–memory amygdala, and the insula (Fig.  2b, JI). No changes were (DMSM), the salience–supplementary sensory (SAL– seen at day 14 (Fig. 2c, JI). S2), and the striatal–motor (STR–MO) supra-network (Fig. 3d). Supra-network refers to a collection of networks Functional MRI findings in this study determined to have some common charac- Stimulus‑evoked fMRI changes post‑concussion Group teristics in terms of connectivity to other networks within ICA of stimulus-evoked fMRI showed no difference in the same supra-network while different connectivity char - spatial extents of the S1 IC spatial maps associated with acteristics in regard to networks in other supra-networks. functional activation during forepaw stimulation at day The DMSM supra-network consisted of Default 2 post-concussion (Fig.  3a). An increased spatial extent Mode Network-like ICs along the Anterior T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 11 of 20 Fig. 3 Functional MRI of concussed and sham animals. a–c Post-concussion stimulus-evoked fMRI activity changes of a CON day 2 versus sham, b CON day 7 versus sham, c CON day 14 versus sham. h–j Post-concussion changes in local intrinsic functional connectivity in (H) CON day 2 versus sham, (I) CON day 7 versus sham, and j CON day 14 versus sham. a–c, h–j Statistical map thresholded at p value < 0.05 (two-tailed), unpaired two sample t test, implemented as permutation tested for the General Linear Model, corrected for multiple comparisons with mass-based FSL’s Threshold-free Cluster enhancement ( TFCE). Statistical maps were overlaid on the study-specific averaged EPI images. Red anatomical orientation marker L = Left, R = Right. ACA = Anterior Cingulate Area, AM = Amygdala, AUD = Auditory Area, CPu = Caudate Putamen, GP = Globus Pallidus, HP = Hippocampus, INS = Insula, M2 = Secondary Motor Cortex, MB = Midbrain, PRT = Pretectal area, S1 = Primary Somatosensory Cortex, RSN = Retrosplenial Area, TH-pmc = Thalamus-polymodal association cortex, VIS = Visual Area. d Average functional network connectivity (FNC) matrices among Independent Components (ICs) identified by IVA-GL of (a) sham (n = 14). Colour scaled by z test statistics; non-black cells were defined as component–component connectivity deemed statistically significant. One sample t tests, permutation-tested, and FDR-corrected (q value < 0.05, two-tailed). e–g Post-concussion FNC changes of e CON day 2 versus sham, f CON day 7 versus sham, g CON day 14 versus sham. Colour scaled by z test statistics; non-black cells were defined as component–component connectivity deemed statistically significant. Two sample t tests, permutation-tested, and FDR-corrected (q value < 0.1, two-tailed) Cingulate–Retrosplenial Cortex axis (ACA RSN group: with earlier whole brain group ICA functional network IC 03, 06, 52, 62, 82, 89), Hippocampal–Subcorti- characterisation of the mouse brain [55]. IC 05 spatial cal memory circuit (HP MEM group: IC 05, 08, and maps showed several regions strongly connected to 34), Primary Somatosensory and Association Area one another: the retrosplenial area (RSN), the retro- (S1 group: IC 04, 11, and 31), Visual–Auditory areas hippocampal (rHP) area, the Anterior Thalamic Nuclei (VA group: IC 17 and IC 46), and Thalamic polymodal (ATN); this IC was also positively correlated with the association cortex (TH-pmc group: IC 10 and 30). anterior cingulate and retrosplenial ICs (IC 03, 06, and The strong functional connectivity among CG RSN 52), and the retrohippocampal IC (IC 08). These con - and VA areas were consistent with the earlier char- nections resemble the cortical memory pathway of acterisation of the Default Mode Network (DMN) in HP > RSN > ACA and a subcortical memory pathway the rodents [71, 72] and the overall strong connectiv- of HP > ATN > RSN/ACA, that support head direc- ity within the DMSM supra-network was consistent tion, spatial navigation/memory/coding, and emotion To and Nasrallah acta neuropathol commun (2021) 9:2 Page 12 of 20 [73–75]. Resting-state functional connectivity was with TI. More detailed linear regression revealed that Dp known to have a basis in structural connectivity of and ODI in these ROIs were significantly correlated with direct nerve projections [76]. TI only among CON day 7 and CON day 14 animals but The SAL–SS supra-network consisted of the Primary not among the sham and CON day 2 animals (Fig. 4a–f ), Somatosensory Area, Barrel Field (S1BF, IC 13), the with the exception of ODI in the TH-pmc (Fig.  4e) and Supplementary Somatosensory Areas (S2 group: IC 02, ACA (Fig.  4f). ODI was negatively correlated with TI 18, and 45), and the Salience Network (SN group: IC 37 among sham and CON day 2 cohorts but not among [Insula area] and IC 67 [Amygdala]). The SAL–SS supra- CON day 7 and day 14 cohort. ODI was not correlating network is characterised by positively correlated func- with TI in the ACA. tional connectivity of ICs within the supra-network and Voxel-wise correlation analysis also defined three negatively correlated functional connectivity of its ICs ROIs—CPu, AM, and INS that showed significant cor - with ICs of the DMSM supra-cluster. relation between KCC and TI. More detailed analysis The STR–MO supra-network consisted of the Caudate showed KCC were positively correlated with TI in all Putamen (CPu, IC 07) and the Primary Motor Area (M1, three ROIs among CON day 7 and CON day 14 animals IC 14). The distinguishing feature of this supra-network but not among sham and CON day 2 cohorts (Fig. 4g–i). is its components have a positive correlation within the supra-network and with the other two supra-networks. Discussion One-sample t test results matrices of IC–IC FNC of This work shows a mismatch in the onset and recovery CON cohorts are shown in Additional file 2: Fig. S2A-C. of the structural and functional attributes of the brain following a concussion, evidenced by advanced neuro- Whole brain restingst ‑ ate functional connectivity changes imaging, beyond that of symptom subsidence. Motor post‑concussion Figure  3e–g reflect the two-sample deficits were evident immediately after concussion while t test comparison of the CON cohorts compared to the deficits in learning were only evident at day 7; all symp - sham cohort. At day 2 post-concussion, concussed mice toms resolved at day 14. Imaging findings varied; DTI had increased functional connectivity between the M1 and NODDI changes peaked at day 7 and significantly and S1, upper limbs (Fig. 3e). At day 7 post-concussion, no reduced at day 14 post-concussion while the functional IC–IC FNC connectivity change was detected (Fig. 3f ). At connectivity increased at day 2 and 14 post-concussion. day 14 post-concussion (Fig. 3g), the increased connectiv- Additionally, stimulus-evoked fMRI detected no differ - ity was detected in the CPu—Auditory Area (AUD), Left ences at day 2 but increased cortical activation at day 7 S2—ACA, and Amygdala (AM)—Insula (INS) connec- and 14 post-concussion. tions. Behavioural changes following concussion Local intrinsic functional connectivity changes post‑con ‑ Motor-balance deficits were significant following a sin - cussion At day 2 post-concussion, concussed mice had gle concussive injury, as evidenced by lower NSS at day increased local functional connectivity in several regions 2 post-concussion. Motor-balance symptoms partially in the DMSM supra-network (S1, TH-pmc, and HP), as recovered at day 7 and fully recovered at day 14. This well as the Salience-like network (INS and AM), and the recovery timeline of motor-balance symptoms in our CPu (Fig.  3h). The increased local connectivity at day 2 model of rotational, acceleration/deceleration concussive subsided at day 7 post-concussion, with only decreased injury is consistent with earlier studies of repeated or sin- local connectivity in a small area of the hippocampus gle mild impacts of the CHIMERA model, which showed (Fig.  3i). However, at day 14 post-concussion, increased injured animals recovered their motor balance function local functional connectivity were detected in many at approximately day 14 [77, 78]. regions of the brain, including the regions in the DMSM Concussed animals displayed reduced TI and spent supra-network–DMN (ACA, RSN), VA (VIS, AUD), and more time at the centre of the area compared to shams S1 networks, the SAL–SS supra-network–SN (INS and at day 7 post-concussion; this behaviour normalised at AM), and the CPu (Fig. 3j). day 14. This trend of reduced TI and more time spent in Averaged ReHo7 maps of sham and CON cohorts are the centre of the arena was consistent with similar results shown in Additional file 2: Fig. S2D-G. after a single moderate injury CHIMERA impact [79]. On the other hand, increased TI, anxiety-like behaviours, Correlation of MRI findings with behavioural symptoms and less time spent at the centre of the arena during an Voxel-wise correlation analysis defined three ROIs—cc, open field task were observed at day 1 and day 7 after a TH-pmc, and ACA of interest, that showed Dp were single [78] or two consecutive mild CHIMERA impacts positively correlated and ODI were negatively correlated 24 h apart [77] and recovery occurred at day 14 [77]. The T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 13 of 20 Fig. 4 Correlation of diffusion and functional MRI with Open Field behavioural measures. Simple linear regression analysis of Parallel Diffusivity (Dp) (a–c), Orientation Dispersion Index (ODI) (d–f), and Kendall’s Coefficient of Concordance (KCC) from seven voxels neighbourhood Regional Homogeneity analysis of resting-state functional MRI (g–i), quantified from different regions of interest, the corpus callosum (cc—a, d), polymodal association area of the Thalamus ( TH-pmc—b, e), Anterior Cingulate Area (ACA—c, f), Caudate Putamen (CPu—g), Amygdala (AM—h), and Insula (INS—I) with Thigmotaxis Index ( TI). Regression analysis were performed across sham and CON day 2 cohorts combined (blue lines) or CON day 7 and CON day 14 combined (orange line). Regression lines displayed as mean and 95% confidence intervals slopes. ns = not significant, *p value < 0.05, **p value < 0.01 majority of other closed-head injury model studies also reported by human studies showing enlarged cortical showed opposite trends of increased anxiety-like behav- structures in the short-term—from 24 h [85], 7 days [86] iours and decreased time spent in the centre of the arena or up to 5.7  years post-concussion [87]. The majority of at 2 days [80–82] or up to one month post-injury [20, 83]. structural imaging studies in humans reported shrinkage Ertürk et  al. observed no alternations from day 4 up to in brain tissue volume after mild to moderate TBI (see 8 weeks post-injury [84]. Ross et  al. [87] for review), though a recent study com- paring a large number of mild to moderate TBI patients Structural brain changes following concussion (n = 50) to the USA’s Food and Drug Administration- Structural imaging and brain morphometry showed approved NeuroQuant normal control database [88, 89] surprising dynamic post-concussion changes that cor- revealed those patients have enlarged cortical grey mat- related well with motor-balance symptoms (NSS) and ter, cerebellar white matter, and hippocampal volumes anxiety-like behaviours (TI) to a lesser extent. At day 2 [87]. The short-term increased brain tissue volume and post-concussion, TBM analysis detected enlargement in lack of long-term or permanent atrophy in our model can the ventricles and several structures, including the audi- be explained by the lack of diffuse axonal injury; DTI and tory, visual, and S1, the hippocampus, anterior cingulate NODDI detecting no changes characteristic of diffuse cortex, and the amygdala. Similar findings have been axonal injury in our model and the correlation between To and Nasrallah acta neuropathol commun (2021) 9:2 Page 14 of 20 diffuse axonal injury and brain atrophy after TBI were optic tracts of a mouse model of closed-head injury at known in human TBI [90, 91]. Brain structural and mor- 1, 6, 12, and 18  weeks post-injury, even when mem- phological changes in this model occurred as early as 2 ory deficits resolved within the first week post-injury; days post-injury and correlated well with behavioural persistent neuroinflammation, including astroglio - symptoms. While this might seem very quick, this is con- sis and microgliosis were associated with the DTI and sistent with our earlier study in the same model [40] and NODDI changes in this model [21]. On the other hand, other studies which showed brain morphological changes a rat model of a single mild modified controlled corti - could happen as early as 24-h post-intervention [92–94], cal impact showed no changes at day 2 post-injury but ranging from changes in oestrous cycle [92], environ- elevated FA and reduced MD and Dr in the corpus mental enrichment [93], and maze training [94]; though callosum and external capsule at day 7 post-injury— the precise underlying mechanism is unclear. consistent with our results—though the changes nor- malised at day 14 [22]. Increased FA was also found in a rat drop weight model of mild traumatic brain injury Diffusion MRI changes in the grey mater after concussion 7 days post-injury [23]. Diffuse increases in FA in the grey matter of concussed In Mierzwa et  al., demyelination of degenerating animals were observed at day 7 post-injury. Increased FA and intact axons was found at day 3 post-injury in a in the grey matter is known to be associated with per- mouse model of a single mild closed-head injury, and sistent post-concussive symptoms (PPCS) following a was followed by remyelination and excessive myelina- concussion in human patients [24], consistent with our tion (including double-layered myelin sheaths) between results. Increased FA in the grey matter is also consistent day 7 and day 14 post-injury [99]. The process of remy - with other in vivo animal studies [25–27]. Closer analysis elination of previously demyelinated axons in a cupri- showed this increased FA in the grey matter was primar- zone-fed mouse model of demyelination showed that ily driven by a decrease in Dr and MD without as much the remyelination process decreased MD and Dr but decreased Dp. This pattern of elevated FA associated no changes to Dp were observed [100]. The observed with reduced Dr and mostly unchanged Dp in the grey decreased Dr, MD, and ODI in our model at day 14 matter, is consistent with the majority of human PPCS post-concussion suggested the possibility of the mice showing elevated grey matter FA [24], and rat closed- having excessive myelination in the corpus callosum head injury models [27]. A minority of human PPCS as a result of injury repair/recovery. Increased FA and cases [24], rat open-head injury model [25] or repetitive decreased MD have also been observed in non-injury mild blast exposure model [26] report elevated grey mat- neuroplasticity processes (a spatial learning task) ter FA, which is often associated with increased Dp with- involving increased myelination and myelin basic pro- out changes in Dr. Regardless of the patterns of Dp or Dr tein expression [101]. changes, this elevated FA in the grey matter post-TBI has Currently, only a few NODDI studies in human con- been associated with astrogliosis [25, 27] or microstruc- cussion have been published to date and the results tural remodelling in a rat model of open-head injury [95]. have been contradictory. Wu et  al. [30] and Church- Of particular interest is the association of DTI/NODDI ill et  al. [31] detected decreased NDI in the white changes and astrocyte activity, since astrocytes have dual matter of concussed patients imaged no longer than roles in neuronal plasticity and reconstruction after trau- 2  months post-injury. On the other hand, our patterns matic brain injury [96, 97]. of increased NDI and decreased ODI in the white mat- ter were consistent with Churchill et  al. [32] where Diffusion MRI changes in the white mater after concussion increased NDI and decreased ODI were detected in the Increased FA and decreased ODI were detected in white matter and grey matter-white matter boundaries the white matter tracts at both day 7 and day 14 post- of young, healthy athletes with a history of concus- concussion in our model. Decreased FA in the white sion (average of two concussions) imaged, on average, matter is commonly associated with reduced white 24  months post-injury. In our study, increased FA, matter integrity in Alzheimer’s disease [98], and in decreased ODI, and increased NDI were found in the animal models of repeated mild TBI [17, 18]. Other white matter of concussed mice. These findings coupled single-impact rodent closed-head injuries also showed with the relatively young age of the animals used (aver- decreased FA in the white matter: drop-weight model age age at impact or sham procedure 13.2 ± 1.4  weeks) in the rats [19] and piston-driven closed-head injury suggested the age range and injury profile at day 7 and in the mice [20, 21] at 1 and 8 days post-injury [19] day 14 post-concussion of our animals fit with that of or up to a month [20], and 18  weeks post-injury [21]. young, healthy humans imaged on average 24  months Decreased FA and increased ODI were found in the T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 15 of 20 (and at least 9 months) post-concussion. Churchill et al. or anti-correlations were found in the absence of global [32] suggested increased NDI and decreased ODI in signal regression during the pre-processing steps [105]. the white matter of young, healthy and fit athletes post- A number of human studies showed the persistence of concussion and without concussion symptoms was an anti-correlated connections, even without global signal indication that the axons recovering from injury were regression, in ROI-based [106–108], graph theoretical strengthened in the long-term after injury. [109], and independent component analysis [70]. Pat- terns of S2 cortex negative correlation with areas in Brain functional connectivity changes the DMSM were found in mice [55] and rats [110]. The following concussions human SN, which is primarily comprised of the insula, Initially, there was an increase in functional connectivity amygdala, and dorsal anterior cingulate cortex [111], has between the M1 and the S1 upper limbs and increased been shown to also have anti-correlation with the human local intrinsic connectivity in the S1, TH-pmc, and HP, DMN [70, 112], The human Dorsal Attention Network is SN, and the CPu. This increased long-range and local known to be another network that is anti-correlated with functional connectivity was normalised at day 7 post- the human DMN [112]; thus, the S1BF and the S2 identi- concussion, when motor-balance deficits mostly sub - fied in the mouse brains may play an analogous role to sided but psychological symptoms started. At day 14, the human Dorsal Attention Network. all symptoms were normalised but increased long-range and local functional connectivity were found within the Multi-phase brain recovery after concussion SN. Our results of initial increased functional connec- We observed a biphasic pattern of neural recovery and tivity associated with significant motor-balance deficits plasticity post-injury, each with distinct patterns of were consistent with some smaller studies that showed behavioural symptoms and associated MRI findings significant rsfMRI changes associated with post-con - (Fig. 5). cussion symptoms [34, 36, 37, 102]; the trends and net- In the first phase of post-concussion recovery, up to day works most frequently involved in studies were reduced 7 post-injury, brain recovery was mostly related to func- connectivity in the posterior DMN [34, 102], increased tional compensation. In this phase, significant motor-bal - connectivity in the anterior DMN [34, 102], reduced anti- ance symptoms were evident, which were associated with correlation among networks with anti-correlation rela- increased long-range and local functional connectivity tionships [102], and decreased local intrinsic functional with no concurrent changes in stimulus-evoked response connectivity in the SN [36, 37], in the lateralised cognitive detected with task-based fMRI; this can be interpreted as control network [37], and in regions related with motor, the brains’ increased activity to functionally compensate sensorimotor, attention, and phonological processing for the injury prior to injury recovery, healing, or plastic- [36]. On the other hand, Meier et  al. reported short- ity [113]. term elevated local intrinsic functional connectivity in In the second phase of concussion recovery, between regions associated with the DMN that normalised upon day 7 post-injury and day 14, brain recovery was domi- symptom recovery [39]. Other studies in humans have nated by neural plasticity. In this phase, particularly at also found persistent rsfMRI changes beyond symptom day 7 post-concussion, there was significant psycho - recovery [35, 103, 104], with one study reporting no sig- logical symptoms in the mice, evidenced by the open nificant changes to rsfMRI connectivity when symptoms field assessment, and DTI/NODDI detected changes were highest, but ongoing rsfMRI changes after symp- consistent with astrogliosis and neuroinflammation toms had recovered [103]. Our results demonstrated [25]. Interestingly, the interaction of astrocytes with persistent local and long-range functional connectivity injured neurons has been shown to result in hyperex- changes at day 14 despite resolution of symptoms. Our citable neurons [96]. This change in cortical excitability findings of no functional connectivity changes and signif - was observed in a rat model of controlled cortical injury, icant psychological symptoms at day 7 post-concussion, with injured animals showing increased stimulus-evoked were also consistent with human studies reporting no fMRI activation between 1 and 4 weeks post-injury rsfMRI changes despite significant ongoing symptoms as [114]. Furthermore, Verley et  al. showed bilateral hyper- our results [38, 103]. excitability to a unilateral forepaw stimulation that was A number of novel networks were defined in this observed in the first week before functional reorganisa - study as part of the resting-state functional connectiv- tion occurred and consolidated to unilateral hyperex- ity architecture through the patterns of mostly positive citability from week 2 to 4 [114]. Our stimulus-evoked correlation among networks within the same supra-net- fMRI results showed increased fMRI responses at day 7 work and negative or anti-correlation among networks and day 14 post-injury, and also demonstrated bilateral of different supra-network by contrast. These negative response enhancement in the CON day 7 cohort and To and Nasrallah acta neuropathol commun (2021) 9:2 Page 16 of 20 Fig. 5 Deficit and recovery trajectories of behavioural and magnetic resonance imaging (MRI) markers post-concussion. Relative changes of behavioural and MRI metrics of concussed cohorts relative to the sham cohorts (horizontal black dotted line through 0). Behavioural measurements (NSS and TI) scaled as t-statistics of the corresponding CON versus sham comparison. MRI biomarkers (JI, FA, NDI, ODI, KCC, and stim-fMRI) scaled as whole-brain averaged t-statistics of the corresponding CON vesus sham comparison; whole-brain t-statistics were used as a biomarker proxy that incorporated both extents and degrees of change. NSS = Neuro Severity Score, TI = Thigmotaxis Index, JI = Jacobian Index, FA = Fractional Anisotropy, NDI = Neurite Density Index, ODI = Orientation Dispersion Index, KCC = Kendall’s Coefficient of Concordance (resting-state functional Magnetic Resonance Imaging Regional Homogeneity), and stim-fMRI (stimulus-evoked functional Magnetic Resonance Imaging). An alternative version of figure separating behavioural and MRI changes can be found in Supplementary data Fig. S4 unilateral enhancement in the CON day 14 cohort, sug- assess depression, anxiety, irritability, aggression [116], gesting the second phase of recovery being reflective of and hyperarousal [117] symptoms should be included. changes in neuroplasticity. The correlation of MRI biomarkers and the neurological The post-concussion cortical hyperexcitability detect - processes that supposedly underlie the two brain recov- able with stimulus-evoked fMRI may explain the appar- ery phases were hypothesised based on MRI biomarker ently contradictory trajectory of resting-state functional fingerprinting of prior studies, often of different brain connectivity in our model through different timepoints. injury models. Nevertheless, the specific correlation of a The increased functional connectivity at day 2 without histopathological process with a specific trend of diffu - concurrent cortical hyperexcitability possibly reflected sion imaging metric change is a complicated issue. For functional compensation [113] without neural recovery/ example, a specific pathology may result in a predictable plasticity. The increased functional connectivity at day biomarker on DTI or NODDI; however, different patholo - 14 with increasing cortical hyperexcitability might have gies may create the same diffusion metric change [118]. been the result of ongoing neural plasticity in the brains Furthermore, in complex conditions like concussion or [114]. NODDI patterns of increased NDI and decreased traumatic brain injury, pathologies generally do not occur ODI at day 7 and 14 also supported hypothesis of neural alone. Astrogliosis, microgliosis, and axonal injury gener- plasticity occurring during this phase. Furthermore, DTI ally occur together as a consequence of concussive injury changes in the white matter also supports neuroplasticity in mice [40, 77, 78, 115, 119]. As described in our earlier processes that may include remyelination or excessive publication [40] DTI and NODDI metrics only correlated myelination. with microgliosis, similar to a NODDI examination of a model of microglia depletion and repopulation [120] in Limitations regions with just neuroinflammation. A more thorough The study suffers from a number of limitations. The investigation to include biological measures would be behavioural assessments applied were restricted to a small needed to further validate the findings in this work. number; rotarod can be used in addition to NSS to allow There are inherent limitations to extrapolating find - for more fine-grained scoring of motor-balance func - ings in animal models to human pathologies and TBI tion/deficit [77], especially since the NSS range in this models are of no exception. There are differences in model was quite narrow. Learning and memory assess- gross anatomy between rodents and humans (lack of ments, such as Barnes maze [115] should also be consid- gyri/sulci in rodents) and rotational force component ered. In future studies, other behaviour tests that better common in large-brained humans TBIs are difficult to T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 17 of 20 be replicated in small-brained animals [121]. Tauopathy will provide a more complete clinical picture in human is a significant feature in repetitive closed head injury concussions. (for example, contact sport athletes and military per- sonnel) [122], but TBI-induced tauopathy in rodents Supplementary information have mainly been demonstrated in transgenic mice, Supplementary information accompanies this paper at https ://doi. org/10.1186/s4047 8-020-01098 -y. which were genetically modified to develop tauopathy [121]. Other studies in the CHIMERA model showed Additional file 1: Data 1. “Signal” resting-state Group Independent Vector tau phosphorylation to be a transient feature, which Analysis components. normalised by day 7 post-injury [77, 115]. It is notable Additional file 2: Figure S2. Group averaged resting-state MRI functional that in the CHIMERA model, two consecutive mild connectivity. (A–D) Average functional network connectivity (FNC) matri- ces among Independent Components (ICs) identified by IVA-GL of (A) impacts 23  h apart caused increased tau phosphoryla- sham (n = 14), (B) CON day 2 (n = 9), (C) CON day 7 (n = 10), and (D) CON tion lasting until day 2 post-injury while a single mod- day 14 (n = 10) cohorts. Colour scaled by z test statistics; non-black cells erate-severe impact only cause transient increased were defined as component-component connectivity deemed statistically significant. One sample t-tests, permutation-tested, and FDR-corrected tau-phosphorylation at 6  h post-injury. Assuming the (q-value < 0.05, two-tailed). (E–H) Regional Homogeneity analysis’s local brains can completely recover after an injury, our imag- intrinsic functional connectivity represented as averaged seven-voxels- ing study attempted to identify imaging biomarkers neighbourhood Kendall’s Coefficient of Concordance (KCC) maps of (E) sham, (F) CON day 2, (G) CON day 7, and (H) CON day 14 cohorts. for injury recovery and, more importantly, a window Additional file 3: Figure S3. Randomise reproducibility test of DTI and of vulnerability within which consecutive injuries may NODDI metrics. Voxel-by-voxel statistical analysis results of Diffusion Tensor cause disproportionate consequences, for example, Imaging (FA = Fractional Anisotropy, MD = Mean Diffusivity, Dp = Parallel long-term tauopathy. Chronic traumatic encephalopa- Diffusivity, Dr = Radial Diffusivity) and Neurite Orientation Dispersion and Density Imaging metrics (NDI = Neurite Density Index, ODI = Orientation thy has been associated with sub-populations at risk of Dispersion Index) and Tensor-based Morphometry with Jacobian Index repetitive traumatic brain injury, such as contact sport (JI) reproducibility test. Statistical map thresholded at P value < 0.05 (two- athletes and military personnel [123] while animal tailed), unpaired two sample t-test, implemented as permutation tested for the General Linear Model, corrected for multiple comparisons with mass- models of repetitive closed-head injury do not consist- based FSL’s Threshold-free Cluster enhancement ( TFCE). Statistical maps ently demonstrate only transient elevated phospho-tau were overlaid on the averaged and registered DTI and NODDI metrics [124]. Long-term elevated phosphor-tau, 6  months maps corresponding to the statistical maps (DTI and NODDI results) and structural template ( TBM results). Corresponding grey scale map for each post-injury, was shown in one study after as many as 42 averaged DTI and NODDI metrics maps were provided; units for Dp, Dr, and impacts [125]. More research is required at this stage to MD were in mm/s2. ACA = Anterior Cingulate Area, AM = Amygdala, AUD address the question of human pathology can be mod- = Auditory Area, cc = corpus callosum, ec = external capsule, HP = Hip- pocampus, HYP = Hypothalamus, ic = internal capsule, INS = Insula, MB elled in animals and the existence of a hypothetical win- = Midbrain, PAL = Palladium, S1 = Primary Somatosensory Cortex, RSN = dow of vulnerability. Retrosplenial Area, STR = Striatum, TH = Thalamus, VIS = Visual Area. Additional file 4: Figure S4. Deficit and recovery trajectories of behav- Conclusion ioural and Magnetic Resonance Imaging (MRI) markers post-concussion. Relative changes of (A) behavioural and (B) MRI metrics of concussed Our multi-modal assessment of a mouse model of con- cohorts relative to the sham cohorts (horizontal black dotted line through cussion showed varying trends in the temporal profile 0). Behavioural measurements (NSS and TI) scaled as t-statistics of the of different MRI markers. While concussion symptoms corresponding CON vs. sham comparison. MRI biomarkers (JI, FA, NDI, ODI, KCC, and stim-fMRI) scaled as whole-brain averaged t-statistics of the corre- and routine structural imaging found significant changes sponding CON vs. sham comparison; whole-brain t-statistics were used as up to day 7 post-concussion, the changes were normal- a biomarker proxy that incorporated both extents and degrees of change. ised by day 14. By contrast, advanced imaging using NSS = Neuro Severity Score, TI = Thigmotaxis Index, JI = Jacobian Index, FA = Fractional Anisotropy, NDI = Neurite Density Index, ODI = Orienta- DTI, NODDI, resting-state, and stimulus-evoked fMRI tion Dispersion Index, KCC = Kendall’s Coefficient of Concordance (resting- revealed ongoing changes that persisted at day 14 with state functional Magnetic Resonance Imaging Regional Homogeneity), and different onset, reflecting ongoing biological and molec - stim-fMRI (stimulus-evoked functional Magnetic Resonance Imaging). ular changes. To the best of the authors’ knowledge, this study represents the first study to perform multi-modal Acknowledgements advanced MRI and behavioural assessment to moni- This research was supported by Motor Accident Insurance Commission (MAIC), The Queensland Government, Australia (Grant Number: 2014000857) tor recovery after a concussion in a mouse model. The and The University of Queensland (Promoting Women’s fellowship: CRM findings in this study have important implication for 200221-005374). We acknowledge the support from the Queensland NMR translation to human clinical settings. First, recovery Network and the National Imaging Facility (a National Collaborative Research Infrastructure Strategy capability) for the operation of 9.4T MRI and utilisation and discharge by clinical assessment criteria may not of image processing computational resources at the Centre for Advanced indicate complete brain recovery and the brain may still Imaging, the University of Queensland. be vulnerable to disproportionate consequences from subsequent concussions. Second, it is likely that multi- modal assessments using different imaging methods To and Nasrallah acta neuropathol commun (2021) 9:2 Page 18 of 20 Authors’ contribution 16. Eierud C, Craddock RC, Fletcher S, Aulakh M, King-Casas B, Kuehl D XVT: Conceptualisation; Data curation; Formal analysis; Investigation; Meth- et al (2014) Neuroimaging after mild traumatic brain injury: review and odology; Project administration; Resources; Software; Validation; Visualisation; meta-analysis. NeuroImage Clin 4:283–294 Roles/Writing—original draft. FAN: Conceptualisation; Funding acquisition; 17. Wortman RC, Meconi A, Neale KJ, Brady RD, McDonald SJ, Chris- Investigation; Project administration; Supervision; Writing—review & editing. tie BR et al (2018) Diffusion MRI abnormalities in adolescent rats All authors read and approved the final manuscript. given repeated mild traumatic brain injury. Ann Clin Transl Neurol 5:1588–1598 Availability of data and materials 18. Haber M, Hutchinson EB, Sadeghi N, Cheng WH, Namjoshi D, Cripton Data from this study is available, without reservations, on request to the cor- P et al (2017) Defining an analytic framework to evaluate quantitative responding author. MRI markers of traumatic axonal injury: preliminary results in a mouse closed head injury model. eNeuro 4:ENEURO.0164-17.2017 Ethics approval and consent to participate 19. Tu T-W, Lescher JD, Williams RA, Jikaria N, Turtzo LC, Frank JA (2017) All experiments were approved by the Institutional Animal Ethics Committee Abnormal injury response in spontaneous mild ventriculomegaly at the University of Queensland (Animal Ethics Committee approval number Wistar rat brains: a pathological correlation study of diffusion tensor QBI/260/17). and magnetization transfer imaging in mild traumatic brain injury. J Neurotrauma 34:248–256 Competing interests 20. Rodriguez-Grande B, Obenaus A, Ichkova A, Aussudre J, Bessy T, Barse All authors have no competing financial interests to declare. E et al (2018) Gliovascular changes precede white matter damage and long-term disorders in juvenile mild closed head injury. Glia Received: 28 September 2020 Accepted: 1 December 2020 66:1663–1677 21. Gazdzinski LM, Mellerup M, Wang T, Adel SAA, Lerch JP, Sled JG et al (2020) White matter changes caused by mild traumatic brain injury in mice evaluated using neurite orientation dispersion and density imag- ing. J Neurotrauma 11:1–11 22. Hoogenboom WS, Rubin TG, Ye K, Cui MH, Branch KC, Liu J et al (2019) References Diffusion tensor imaging of the evolving response to mild traumatic 1. National Center for Injury Prevention and Control (2003) Report to brain injury in rats. J Exp Neurosci 13:1–22 Congress on mild traumatic brain injury in the United States: steps to 23. Kikinis Z, Muehlmann M, Pasternak O, Peled S, Kulkarni P, Ferris C et al prevent a serious public health problem. Centers for Disease Control (2017) Diffusion imaging of mild traumatic brain injury in the impact and Prevention, Atlanta, GA accelerated rodent model: a pilot study. Brain Inj 31:1376–1381 2. Williams RM, Puetz TW, Giza CC, Broglio SP (2015) Concussion recovery 24. Bouix S, Pasternak O, Rathi Y, Pelavin PE, Zafonte R, Shenton ME (2013) time among high school and collegiate athletes: a systematic review Increased gray matter diffusion anisotropy in patients with persistent and meta-analysis. Sport Med 45:893–903 post-concussive symptoms following mild traumatic brain injury. PLoS 3. Henry LC, Elbin RJ, Collins MW, Marchetti G, Kontos AP (2016) Examin- One 8:e66205 ing recovery trajectories following sport-related concussion using a 25. Budde MD, Janes L, Gold E, Turtzo LC, Frank JA (2011) The contribution multi-modal clinical assessment approach. J Neurosurg 78:232–241 of gliosis to diffusion tensor anisotropy and tractography following 4. Semple BD, Lee S, Sadjadi R, Fritz N, Carlson J, Griep C et al (2015) traumatic brain injury: validation in the rat using Fourier analysis of Repetitive concussions in adolescent athletes—Translating clinical and stained tissue sections. Brain 134:2248–2260 experimental research into perspectives on rehabilitation strategies. 26. Badea A, Kamnaksh A, Anderson RJ, Calabrese E, Long JB, Agoston Front Neurol 6:1–16 DV (2018) Repeated mild blast exposure in young adult rats results in 5. Stejskal EO, Tanner JE (1965) Spin diffusion measurements: spin echoes dynamic and persistent microstructural changes in the brain. NeuroIm- in the presence of a time-dependent field gradient. J Chem Phys age Clin 18:60–73 42:288–292 27. Braeckman K, Descamps B, Pieters L, Vral A, Caeyenberghs K, Vanhove C 6. Basser PJ (1995) Inferring microstructural features and the physiologi- (2019) Dynamic changes in hippocampal diffusion and kurtosis metrics cal state of tissues from diffusion weighted images. NMR Biomed following experimental mTBI correlate with glial reactivity. NeuroImage 8:333–344 Clin 21:101669 7. O’Donnell LJ, Westin C-F (2011) An introduction to diffusion tensor 28. Zhang H, Schneider T, Wheeler-Kingshott CA, Alexander DC (2012) image analysis. Neurosurg Clin N Am 22:185–96, viii NODDI: practical in vivo neurite orientation dispersion and density 8. Alexander AL, Lee JE, Lazar M, Field AS (2007) Diffusion tensor imaging imaging of the human brain. Neuroimage 61:1000–1016 of the brain. Neurotherapeutics 4:316–329 29. Tariq M, Schneider T, Alexander DC, Gandini Wheeler-Kingshott CA, 9. Miles L, Grossman RI, Johnson G, Babb JS, Diller L, Inglese M (2008) Zhang H (2016) Bingham-NODDI: mapping anisotropic orientation Short-term DTI predictors of cognitive dysfunction in mild traumatic dispersion of neurites using diffusion MRI. Neuroimage 133:207–223 brain injury. Brain Inj 22:115–122 30. Wu Y-C, Mustafi SM, Harezlak J, Kodiweera C, Flashman LA, McAllister 10. Aoki Y, Inokuchi R, Gunshin M, Yahagi N, Suwa H (2012) Diffusion tensor TW (2018) Hybrid diffusion imaging in mild traumatic brain injury. J imaging studies of mild traumatic brain injury: a meta-analysis. J Neurol Neurotrauma 35:2377–2390 Neurosurg Psychiatry 83:870–876 31. Churchill NW, Caverzasi E, Graham SJ, Hutchison MG, Schweizer TA 11. Brandstack N, Kurki T, Tenovuo O (2013) Quantitative diffusion-tensor (2019) White matter during concussion recovery: comparing diffusion tractography of long association tracts in patients with traumatic brain tensor imaging (DTI) and neurite orientation dispersion and density injury without associated findings at routine MR imaging. Radiology imaging (NODDI). Hum Brain Mapp 40:1908–1918 267:231–239 32. Churchill NW, Caverzasi E, Graham SJ, Hutchison MG, Schweizer TA (2017) 12. Yuh EL, Cooper SR, Mukherjee P, Yue JK, Lingsma HF, Gordon WA et al White matter microstructure in athletes with a history of concussion: (2014) Diffusion Tensor Imaging for outcome prediction in mild trau- comparing diffusion tensor imaging (DTI) and neurite orientation disper - matic brain Injury: a TRACK-TBI study. J Neurotrauma 31:1457–1477 sion and density imaging (NODDI). Hum Brain Mapp 38:4201–4211 13. Bazarian JJ, Zhong J, Blyth B, Zhu T, Kavcic V, Peterson D (2007) Diffusion 33. Biswal B, Yetkin FZ, Haughton VM, Hyde JS (1995) Functional connectiv- tensor imaging detects clinically important axonal damage after mild ity in the motor cortex of resting human brain using echo-planar MRI. traumatic brain injury: a pilot study. J Neurotrauma 24:1447–1459 Magn Reson Med 34:537–541 14. Mayer AR, Ling J, Mannell MV, Gasparovic C, Phillips JP, Doezema D et al 34. Zhou Y, Milham MP, Lui YW, Miles L, Reaume J, Sodickson DK et al (2012) (2010) A prospective diffusion tensor imaging study in mild traumatic Default-mode network disruption in mild traumatic brain injury. Radiol- brain injury. Neurology 74:643–650 ogy 265:882–892 15. Ling JM, Peña A, Yeo RA, Merideth FL, Klimaj S, Gasparovic C et al (2012) 35. Johnson B, Zhang K, Gay M, Horovitz S, Hallet M, Sebastianelli W Biomarkers of increased diffusion anisotropy in semi-acute mild trau- et al (2012) Alteration of brain default network in subacute phase of matic brain injury: a longitudinal perspective. Brain 135:1281–1292 T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 19 of 20 injury in concussed individuals: resting-state fMRI study. Neuroimage 60. Lee JH, Lee TW, Jolesz FA, Yoo SS (2008) Independent vector analysis (IVA): 59:511–518 multivariate approach for fMRI group study. Neuroimage 40:86–109 36. Zhan J, Gao L, Zhou F, Kuang H, Zhao J, Wang S et al (2015) Decreased 61. Michael AM, Anderson M, Miller RL, Adalı T, Calhoun VD (2014) Preserv- regional homogeneity in patients with acute mild traumatic brain ing subject variability in group fMRI analysis: performance evaluation of injury: a resting-state fMRI Study. J Nerv Ment Dis 203:786–791 GICA vs. IVA. Front Syst Neurosci 8:1–18 37. Peters SK, Dunlop K, Downar J (2016) Cortico-striatal-thalamic loop 62. Laney J, Westlake K, Ma S, Woytowicz E, Calhoun VD, Adali T (2015) circuits of the salience network: a central pathway in psychiatric disease Capturing subject variability in fMRI data: a graph-theoretical analysis of and treatment. Front Syst Neurosci 10:1–23 GICA vs. IVAIVA. J Neurosci Methods 247:32–40 38. Kaushal M, España LY, Nencka AS, Wang Y, Nelson LD, McCrea MA et al 63. Du Y, Lin D, Yu Q, Sui J, Chen J, Rachakonda S et al (2017) Comparison (2019) Resting-state functional connectivity after concussion is associ- of IVA and GIG-ICA in brain functional network estimation using fMRI ated with clinical recovery. Hum Brain Mapp 40:1211–1220 data. Front Neurosci 11:1–18 39. Meier TB, Giraldo-Chica M, España LY, Mayer AR, Harezlak J, Nencka AS 64. Zang Y, Jiang T, Lu Y, He Y, Tian L (2004) Regional homogeneity et al (2020) Resting-state fMRI metrics in acute sport-related concussion approach to fMRI data analysis. Neuroimage 22:394–400 and their association with clinical recovery: a study from the NCAA- 65. Szafer A, Zhong J, Gore JC (1995) Theoretical model for water diffu- DOD CARE consortium. J Neurotrauma 37:152–162 sion in tissues. Magn Reson Med 33:697–712 40. To XV, Benetatos J, Soni N, Liu D, Abraha HM, Yan W et al (2020) Ultra- 66. Anderson MJ, Robinson J (2001) Permutation tests for linear models. high field diffusion tensor imaging identifies discrete patterns of con- Aust New Zeal J Stat 43:75–88 cussive injury in the rodent brain. J Neurotrauma 2020;neu.2019.6944 67. Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE (2014) Per- 41. Flierl MA, Stahel PF, Beauchamp KM, Morgan SJ, Smith WR, Shohami mutation inference for the general linear model. Neuroimage 92:381–397 E (2009) Mouse closed head injury model induced by a weight-drop 68. Dickie DA, Mikhael S, Job DE, Wardlaw JM, Laidlaw DH, Bastin ME device. Nat Protoc 4:1328–1337 (2015) Permutation and parametric tests for effect sizes in voxel- 42. Simon P, Dupuis R, Costentin J (1994) Thigmotaxis as an index of anxiety in based morphometry of gray matter volume in brain structural MRI. mice Influence of dopaminergic transmissions. Behav Brain Res 61:59–64 Magn Reson Imaging 33:1299–1305 43. Rorden C, Brett M (2000) Stereotaxic display of brain lesions. Behav 69. Smith SM, Nichols TE (2009) Threshold-free cluster enhancement: Neurol 12:191–200 addressing problems of smoothing, threshold dependence and 44. Bajic D, Craig MM, Mongerson CRL, Borsook D, Becerra L (2017) Identify- localisation in cluster inference. Neuroimage 44:83–98 ing rodent resting-state brain networks with independent component 70. Allen EA, Erhardt EB, Damaraju E, Gruner W, Segall JM, Silva RF et al analysis. Front Neurosci 11 (2011) A baseline for the multivariate comparison of resting-state 45. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ, networks. Front Syst Neurosci 5:1–23 Johansen-Berg H et al (2004) Advances in functional and structural MR 71. Gozzi A, Schwarz AJ (2016) Large-scale functional connectivity net- image analysis and implementation as FSL. Neuroimage 23:208–219 works in the rodent brain. Neuroimage 127:496–509 46. Ashburner J, Friston KJ (2003) Morphometry. In: Frackowiak RSJ, Friston 72. Hsu L-M, Liang X, Gu H, Brynildsen JK, Stark JA, Ash JA et al (2016) KJ, Frith C, Dolan R, Friston KJ, Price CJ et al (eds) Human brain function. Constituents and functional implications of the rat default mode Academic Press, Cambridge, pp 707–722 network. Proc Natl Acad Sci 113:E4541–E4547 47. Andersson JLR, Skare S, Ashburner J (2003) How to correct susceptibil- 73. Jankowski MM, Ronnqvist KC, Tsanov M, Vann SD, Wright NF, Erichsen ity distortions in spin-echo echo-planar images: application to diffusion JT et al (2013) The anterior thalamus provides a subcortical circuit sup- tensor imaging. Neuroimage 20:870–888 porting memory and spatial navigation. Front Syst Neurosci 7:1–12 48. Hong X, To XV, Teh I, Rui J, Chuang K (2015) Evaluation of EPI distortion 74. O’Mara SM, Aggleton JP (2019) Space and memory (far) beyond the correction methods for quantitative MRI of the brain at high magnetic hippocampus: many subcortical structures also support cognitive field. Magn Reson Imaging 33:1098–1105 mapping and mnemonic processing. Front Neural Circuits 13:1–12 49. Dohmatob E, Varoquaux G, Thirion B (2018) Inter-subject registration of func- 75. Bubb EJ, Kinnavane L, Aggleton JP (2017) Hippocampal–dience- tional images: do we need anatomical images? Front Neurosci 12:1–11 phalic–cingulate networks for memory and emotion: an anatomical 50. Du Y, Fan Y (2013) Group information guided ICA for fMRI data analysis. guide. Brain Neurosci Adv 1:1–20 Neuroimage 69:157–197 76. Grandjean J, Zerbi V, Balsters J, Wenderoth N, Rudina M (2017) The 51. Calhoun VD, Adali T, Pearlson GD, Pekar JJ (2001) A method for making structural basis of large-scale functional connectivity in the mouse. J group inferences from functional MRI data using independent compo- Neurosci 37:0438-17 nent analysis. Hum Brain Mapp 14:140–151 77. Namjoshi DR, Cheng WH, McInnes KA, Martens KM, Carr M, Wilkinson 52. Smith SM, Andersson J, Auerbach EJ, Beckmann CF, Bijsterbosch J, A et al (2014) Merging pathology with biomechanics using CHIMERA Douaud G et al (2013) Resting-state fMRI in the human connectome (closed-head impact model of engineered rotational acceleration): project. Neuroimage 80:144–168 a novel, surgery-free model of traumatic brain injury. Mol Neurode- 53. Du Y, Allen EA, He H, Sui J, Wu L, Calhoun VD (2016) Artifact removal in gener 9:55 the context of group ICA: a comparison of single-subject and group 78. Namjoshi DR, Cheng WH, Bashir A, Wilkinson A, Stukas S, Martens approaches. Hum Brain Mapp 37:1005–1025 KM et al (2017) Defining the biomechanical and biological threshold 54. Gri L, Douaud G, Bijsterbosch J, Evangelisti S, Alfaro-almagro F, Glasser of murine mild traumatic brain injury using CHIMERA (closed head MF et al (2017) Hand classification of fMRI ICA noise components. impact model of engineered rotational acceleration). Exp Neurol Neuroimage 154:188–205 292:80–91 55. Zerbi V, Grandjean J, Rudin M, Wenderoth N (2015) Mapping the mouse 79. Sauerbeck AD, Fanizzi C, Kim JH, Gangolli M, Bayly PV, Wellington CL, brain with rs-fMRI: an optimized pipeline for functional network identi- Brody DL, Kummer TT (2018) modCHIMERA: a novel murine closed- fication. Neuroimage 123:11–21 head model of moderate traumatic brain injury. Sci Rep 8(1):7677 56. Bell AJ, Sejnowski TJ (1995) An information-maximization approach 80. Hehar H, Yu K, Ma I, Mychasiuk R (2016) Paternal age and diet: the con- to blind separation and blind deconvolution. Neural Comput tributions of a father’s experience to susceptibility for post-concussion 1159:1129–1159 symptomology. Neuroscience 332:61–75 57. Lee T, Girolami M, Sejnowski TJ (1999) Independent component analysis 81. Salberg S, Yamakawa G, Christensen J, Kolb B, Mychasiuk R (2017) using an extended infomax algorithm for mixed subgaussian and Assessment of a nutritional supplement containing resveratrol, prebi- supergaussian sources. Neural Comput 11:417–441 otic fiber, and omega-3 fatty acids for the prevention and treatment of 58. Nasrallah FA, Tay H, Chuang K (2014) Detection of functional connectiv- mild traumatic brain injury in rats. Neuroscience 365:146–157 ity in the resting mouse brain. Neuroimage 86:417–424 82. Mychasiuk R, Hehar H, Candy S, Ma I, Esser MJ (2016) The direction of 59. Anderson M, Adali T, Li X-L (2012) Joint blind source separation swith the acceleration and rotational forces associated with mild traumatic multivariate gaussian model: algorithms and performance analysis. IEEE brain injury in rodents effect behavioural and molecular outcomes. J Trans Signal Process 60:1672–1683 Neurosci Methods 257:168–178 To and Nasrallah acta neuropathol commun (2021) 9:2 Page 20 of 20 83. Sherman M, Liu MM, Birnbaum S, Wolf SE, Minei JP, Gatson JW (2016) 105. Murphy K, Birn RM, Handwerker DA, Jones TB, Bandettini PA (2009) The Adult obese mice suffer from chronic secondary brain injury after mild impact of global signal regression on resting state correlations: are anti- TBI. J Neuroinflamm 13:1–10 correlated networks introduced? Neuroimage 44:893–905 84. Ertürk A, Mentz S, Stout EE, Hedehus M, Dominguez SL, Neumaier L et al 106. Chang C, Glover GH (2009) Eec ff ts of model-based physiological noise (2016) Interfering with the chronic immune response rescues chronic correction on default mode network anti-correlations and correlations. degeneration after traumatic brain injury. J Neurosci 36:9962–9975 Neuroimage 47:1448–1459 85. Govindarajan KA, Narayana PA, Hasan KM, Wilde EA, Levin HS, Hunter 107. Fransson P (2005) Spontaneous low-frequency BOLD signal fluctua- JV et al (2016) Cortical thickness in mild traumatic brain injury. J Neuro- tions: an fMRI investigation of the resting-state default mode of brain trauma 33:1809–1817 function hypothesis. Hum Brain Mapp 26:15–29 86. Wang X, Xie H, Cotton AS, Tamburrino MB, Brickman KR, Lewis TJ et al 108. Uddin LQ, Clare Kelly AM, Biswal BB, Xavier Castellanos F, Milham MP (2015) Early cortical thickness change after mild traumatic brain injury (2009) Functional connectivity of default mode network components: following motor vehicle collision. J Neurotrauma 32:455–463 correlation, anticorrelation, and causality. Hum Brain Mapp 30:625–637 87. Ross DE, Seabaugh JD, Seabaugh JM, Alvarez C, Ellis LP, Powell C et al 109. Gopinath K, Krishnamurthy V, Cabanban R, Crosson BA (2015) Hubs of (2020) Patients with chronic mild or moderate traumatic brain injury anticorrelation in high-resolution resting-state functional connectivity have abnormal brain enlargement. Brain Inj 34:11–19 network architecture. Brain Connect 5:267–275 88. Ross DE, Graham TJ, Ochs AL (2013) Review of the evidence supporting 110. Schwarz AJ, Gass N, Sartorius A, Risterucci C, Spedding M, Schenker E the medical and legal use of NeuroQuant in patients with traumatic et al (2013) Anti-correlated cortical networks of intrinsic connectivity in brain injury. Psychol Inj Law 6:75–80 the rat brain. Brain Connect 3:503–511 89. Ross DE, Seabaugh J, Cooper L, Seabaugh J (2018) NeuroQuant and 111. Akiki TJ, Averill CL, Abdallah CG (2017) A network-based neurobiological NeuroGage reveal effects of traumatic brain injury on brain volume. model of PTSD: evidence from structural and functional neuroimaging Brain Inj 32:1437–1441 studies. Curr Psychiatry Rep 19:81 90. Ding K, De La Plata CM, Wang JY, Mumphrey M, Moore C, Harper C et al 112. Raichle ME (2011) The restless brain. Brain Connect 1:3–12 (2008) Cerebral atrophy after traumatic white matter injury: correlation 113. Iraji A, Chen H, Wiseman N, Welch RD, O’Neil BJ, Haacke EM et al (2016) with acute neuroimaging and outcome. J Neurotrauma 25:1433–1440 Compensation through functional hyperconnectivity: a longitudinal 91. Scheid R, Preul C, Gruber O, Wiggins C, von Cramon DY (2003) Diffuse connectome assessment of mild traumatic brain injury. Neural Plast axonal injury associated with chronic traumatic brain injury: evidence 114. Verley DR, Torolira D, Pulido B, Gutman B, Bragin A, Mayer A et al (2018) from T2*-weighted gradient-echo imaging at 3 T. AJNR Am J Neurora- Remote changes in cortical excitability after experimental traumatic brain diol 24:1049–1056 injury and functional reorganization. J Neurotrauma 35:2448–2461 92. Qiu LR, Germann J, Spring S, Alm C, Vousden DA, Palmert MR et al 115. Bashir A, Abebe ZA, McInnes KA, Button EB, Tatarnikov I, Cheng WH et al (2013) Hippocampal volumes differ across the mouse estrous cycle, (2020) Increased severity of the CHIMERA model induces acute vascular can change within 24 hours, and associate with cognitive strategies. injury, sub-acute deficits in memory recall, and chronic white matter Neuroimage 83:593–598 gliosis. Exp Neurol 324:113116 93. Scholz J, Allemang-Grand R, Dazai J, Lerch JP (2015) Environmental 116. Malkesman O, Tucker LB, Ozl J, McCabe JT (2013) Traumatic brain injury— enrichment is associated with rapid volumetric brain changes in adult modeling neuropsychiatric symptoms in rodents. Front Neurol 4:157 mice. Neuroimage 109:190–198 117. Golub Y, Mauch CP, Dahlhoff M, Wotjak CT (2009) Consequences 94. Lerch JP, Yiu AP, Martinez-Canabal A, Pekar T, Bohbot VD, Frankland PW of extinction training on associative and non-associative fear in a et al (2011) Maze training in mice induces MRI-detectable brain shape mouse model of Posttraumatic Stress Disorder (PTSD). Behav Brain Res changes specific to the type of learning. Neuroimage 54:2086–2095 205(2):544–549 95. Li L, Chopp M, Ding G, Qu C, Nejad-Davarani SP, Davoodi-Bojd E et al 118. Jelescu IO, Budde MD (2017) Design and validation of diffusion MRI (2017) Diffusion-derived magnetic resonance imaging measures of models of white matter. Front Phys 5:61 longitudinal microstructural remodeling induced by marrow stromal 119. Namjoshi DR, Cheng WH, Carr M, Martens KM, Zareyan S, Wilkinson A cell therapy after traumatic brain injury. J Neurotrauma 34:182–191 et al (2016) Chronic exposure to androgenic-anabolic steroids exacer- 96. Burda JE, Bernstein AM, Sofroniew MV (2016) Astrocyte roles in trau- bates axonal injury and microgliosis in the CHIMERA mouse model of matic brain injury. Exp Neurol 275:305–315 repetitive concussion. PLoS ONE 11:1–21 97. Zhou Y, Shao A, Yao Y, Tu S, Deng Y, Zhang J (2020) Dual roles of astro- 120. Yi SY, Barnett BR, Torres-Velázquez M, Zhang Y, Hurley SA, Rowley PA cytes in plasticity and reconstruction after traumatic brain injury. Cell et al (2019) Detecting microglial density With quantitative multi-com- Commun Signal 18:1–16 partment diffusion MRI. Front Neurosci 13:1–9 98. Kantarci K, Murray ME, Schwarz CG, Reid RI, Przybelski SA, Lesnick T et al 121. Risling M, Smith D, Stein TD, Thelin EP, Zanier ER, Ankarcrona M et al (2017) White-matter integrity on DTI and the pathologic staging of (2019) Modelling human pathology of traumatic brain injury in animal Alzheimer’s disease. Neurobiol Aging 56:172–179 models. J Intern Med 285:594–607 99. Mierzwa AJ, Marion CM, Sullivan GM, McDaniel DP, Armstrong RC 122. Castellani RJ, Perry G, Tabaton M (2019) Tau biology, tauopathy, (2015) Components of myelin damage and repair in the progression of traumatic brain injury, and diagnostic challenges. J Alzheimer’s Dis white matter pathology after mild traumatic brain injury. J Neuropathol 67:447–467 Exp Neurol 74:218–232 123. Asken BM, Sullan MJ, DeKosky ST, Jaffee MS, Bauer RM (2017) Research 100. Wang N, Zhuang J, Wei H, Dibb R, Qi Y, Liu C (2019) Probing demyelina- gaps and controversies in chronic traumatic encephalopathy: a review. tion and remyelination of the cuprizone mouse model using multimo- JAMA Neurol 74:1255–1262 dality MRI. J Magn Reson Imaging 50:1852–1865 124. Hoogenboom WS, Branch CA, Lipton ML (2019) Animal models of 101. Blumenfeld-Katzir T, Pasternak O, Dagan M, Assaf Y (2011) Diffusion MRI closed-skull, repetitive mild traumatic brain injury. Pharmacol Ther of structural brain plasticity induced by a learning and memory task. 198:109–122 PLoS One 6:e20678 125. Petraglia A, Plog B, Dayawansa S, Dashnaw M, Czerniecka K, Walker C 102. Sours C, Zhuo J, Janowich J, Aarabi B, Shanmuganathan K, Gullapalli et al (2014) The pathophysiology underlying repetitive mild traumatic RP (2013) Default mode network interference in mild traumatic brain brain injury in a novel mouse model of chronic traumatic encephalopa- injury—A pilot resting state study. Brain Res 1537:201–215 thy. Surg Neurol Int 5:184 103. McCuddy WT, España LY, Nelson LD, Birn RM, Mayer AR, Meier TB (2018) Association of acute depressive symptoms and functional connectivity Publisher’s Note of emotional processing regions following sport-related concussion. Springer Nature remains neutral with regard to jurisdictional claims in pub- NeuroImage Clin 19:434–442 lished maps and institutional affiliations. 104. Meier TB, Bellgowan PSF, Mayer AR (2017) Longitudinal assessment of local and global functional connectivity following sports-related concussion. Brain Imaging Behav 11:129–140 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Acta Neuropathologica Communications Springer Journals

A roadmap of brain recovery in a mouse model of concussion: insights from neuroimaging

Loading next page...
 
/lp/springer-journals/a-roadmap-of-brain-recovery-in-a-mouse-model-of-concussion-insights-zXZKAtjtM0

References (125)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
eISSN
2051-5960
DOI
10.1186/s40478-020-01098-y
Publisher site
See Article on Publisher Site

Abstract

Concussion or mild traumatic brain injury is the most common form of traumatic brain injury with potentially long- term consequences. Current objective diagnosis and treatment options are limited to clinical assessment, cogni- tive rest, and symptom management, which raises the real danger of concussed patients being released back into activities where subsequent and cumulative injuries may cause disproportionate damages. This study conducted a cross-sectional multi-modal examination investigation of the temporal changes in behavioural and brain changes in a mouse model of concussion using magnetic resonance imaging. Sham and concussed mice were assessed at day 2, day 7, and day 14 post-sham or injury procedures following a single concussion event for motor deficits, psycho - logical symptoms with open field assessment, T2-weighted structural imaging, diffusion tensor imaging (DTI), neurite orientation density dispersion imaging (NODDI), stimulus-evoked and resting-state functional magnetic resonance imaging (fMRI). Overall, a mismatch in the temporal onsets and durations of the behavioural symptoms and struc- tural/functional changes in the brain was seen. Deficits in behaviour persisted until day 7 post-concussion but recov- ered at day 14 post-concussion. DTI and NODDI changes were most extensive at day 7 and persisted in some regions at day 14 post-concussion. A persistent increase in connectivity was seen at day 2 and day 14 on rsfMRI. Stimulus- invoked fMRI detected increased cortical activation at day 7 and 14 post-concussion. Our results demonstrate the capabilities of advanced MRI in detecting the effects of a single concussive impact in the brain, and highlight a mismatch in the onset and temporal evolution of behaviour, structure, and function after a concussion. These results have significant translational impact in developing methods for the detection of human concussion and the time course of brain recovery. Background cognitive, physical, and emotional disturbances mani- Concussion, or mild traumatic brain injury (mTBI), is fest within the first 24 h of the injury, and last for several a physical trauma-induced pathophysiological process weeks [2] or longer [3]. More problematic is that certain affecting the brain, resulting in rapid onset of typically physiological disturbances can persist beyond the typi- transient neurological dysfunction, with or without loss cal 2-week window of clinical recovery, raising concerns of consciousness [1]. Concussions are inherently diverse about the super-additive risks associated with repeated in nature and of unpredictable outcome. The cognitive injury incurred while the brain is still recovering from the sequelae from seemingly minor head injuries incurred effects of the first impact. Indeed, growing evidence sug - during sports can be severe and persistent. Conspicuous gests that a concussed individual is at high risk for fur- ther concussion, and that repeated injuries within a short time window can provoke cumulative brain damage [4]. *Correspondence: f.nasrallah@uq.edu.au Objective methods that can accurately diagnose the The Queensland Brain Institute, The University of Queensland, Building impact of a concussion on the brain, allowing for better 79, Upland Road, Saint Lucia, Brisbane, QLD 4072, Australia © The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/publi cdoma in/ zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. To and Nasrallah acta neuropathol commun (2021) 9:2 Page 2 of 20 understanding of the underlying pathology, and track- found in the WM of concussed athletes. Decreased FA ing post-concussion recovery, are thus required. Mag- and increased ODI were found in the optic tracts of a netic resonance imaging (MRI) is a non-invasive imaging mouse model of closed-head injury [21]. method with a number of modalities optimised to detect Studies on the associated integrity of the functional different aspects of the structural and functional integrity connections in the brain, using resting state functional of the brain. The non-invasive nature of MRI also allows MRI, following a concussion have also been limited. repeated imaging measurements, making it ideal for Resting-state functional Magnetic Resonance Imaging tracking the temporal trajectory of the injury. (rsfMRI) detects synchronous low-frequency fluctua - Diffusion Tensor Imaging (DTI) is an MRI modality tions (< 0.3 Hz) of blood oxygen level dependent (BOLD) that probes the motion of water molecules in biological signals among spatially distinct brain regions in subjects tissue [5, 6]. DTI uses a number of metrics to describe at rest [33]. Regions that have synchronous oscillating water diffusivity in a voxel, namely Fractional Anisotropy fluctuations form brain networks that are functionally (FA) [7] parallel diffusivity (Dp), radial diffusivity (Dr), connected. The default mode network (DMN) has been and mean diffusivity (MD) [8]. DTI has been used to the major functional network disrupted after concussion study human concussions and white matter pathologies, in the absence of structural deficits [34, 35]. Zhou et  al. in particular, though the results have often been contra- reported reduced functional connectivity in the poste- dictory [9–15]; both reduced FA coupled with increased rior cingulate cortex and parietal regions and increased MD [9–12] and increased FA coupled with decreased MD functional connectivity around the medial prefrontal cor- [13–15] have been reported in the white matter follow- tex in mTBI patients on average 22 days post-concussion ing mTBI. A 2014 meta-analysis of 122 studies suggested [34]. The decreased connectivity in the posterior DMN that following a concussion, changes in FA are time- was positively correlated with cognitive deficits (lower dependent—FA increases during the acute phase then ability to rapidly switching between cognitive sets) while decreases in the subacute phases following the impact the increased connectivity in the frontal DMN was cor- [16]. Repeated mTBI in rodents [17, 18], single-impact related negatively with posttraumatic depressive symp- drop-weight model in rats [19] and single-impact piston- toms [34]. Similarly, Johnson et  al. detected a reduction driven closed-head injury in mice [20, 21] have shown in the number and strength of the connections in the decreased FA in the white matter. A couple of other stud- posterior cingulate cortex and lateral parietal cortices, ies in rats have shown opposite results: increased FA in and an increase in the number and strength of connec- the white matter [22, 23]. DTI changes in the grey mat- tions in the medial prefrontal cortex in concussed ath- ter (GM) received less attention though increased FA has letes imaged after symptoms had resolved, on average been found in human patients with persistent post-con- 10  days post-concussion [35]. Other networks found cussive symptoms (PPCS) [24] and in vivo animal models with decreased functional connectivity after concus- of traumatic brain injury (TBI) [25–27]. sion include: the Salience Network (SN) [36, 37], in the Neurite orientation dispersion and density imaging lateralised cognitive control network [37], and in regions (NODDI) is yet another approach that provides more related with motor, sensorimotor, attention, and phono- specificity to diffusion imaging changes [28]. NODDI logical processing [36]. Connectivity strength between provides measures of neurite (axons and dendrites) the left dorso-lateral prefrontal cortex and left lateral density and local structural organisation of neurites parietal cortex was reduced with an increasing number by measuring the level of orientation dispersion of of concussions [35]. More recently, Kaushal et  al. inves- restricted anisotropic diffusivities measured with mul - tigated resting-state functional connectivity in concussed tiple diffusion-weighted directions at higher b-values athletes and found that at 48 h post-injury, there were no using orientation dispersion index (ODI) and the con- changes in functional connectivity, despite psychological tribution fraction of restricted anisotropic diffusion to distress, and oculomotor, balance and memory deficits the total diffusion using neurite density index (NDI) [38]. At 8 days post-injury, a global increase in functional [28, 29]. The few NODDI studies applied to map brain connectivity was seen with improving symptoms, which changes in human concussion have reported different recovered by 15  days post-injury [38]. Similarly, Meier findings. Wu et  al. detected decreased NDI without et  al. reported increased local intrinsic functional con- any DTI changes in the white matter (WM) of concus- nectivity in the right middle and superior frontal gyri at sion patients aged 35  years old approximately 2 weeks 24–48  h post-injury, which were normalised at 7  days post-injury [30]. Similarly, both decreased NDI (asso- and 6  months post-medical-clearance [39]. To the best ciated with decreased FA, and increased Dp and Dr) of the authors’ knowledge no functional imaging study [31] and increased NDI (associated with decreased combining stimulus-evoked functional MRI and rsfMRI ODI, increased FA and decreased MD) [32] were were published in a mouse model of closed-head injury. T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 3 of 20 The conflicting findings in human imaging studies support structure included a reclining body plate to sup- described above suggested the difference might be due to port the mouse’s body and a head plate to support the the patients’ background (sex, age, medical history), the head. The head plate had a hole to allow a brass piston injury severity, and the imaging time post-concussion, to deliver the impact from below; two perpendicular lines which affect where each patient was along the injury/ intersecting the centre of the brass piston were used as recovery time course. Therefore, animal models serve as crosshairs to aim and position the head. The body plate a way to study concussion by allowing precise control of was reclined to allow the frontal and parietal bones on all the variables or their effects on the injury and mani - the mouse’s head to be positioned flat against the head festation on MRI findings. In this study, we attempted to plate. The mouse body was secured in the supine posi - chart the temporal evolution of the microstructural and tion by a non-slip silicone mattress and two Velcro straps functional changes post-concussion in a mouse model across the chest and abdomen. The tail was secured to with the aim of highlighting the functional-structural the body plate with masking tape to prevent slipping. mismatch in their recovery trajectories. Anaesthesia was maintained during the securing and positioning for impact at 2% isoflurane in the same gas Methods mixture via a nose cone; total time under anaesthesia was Study design 10–12  min. Time under anaesthesia was kept consist- This is a cross-sectional study which included a total ent across different groups and the experimenter aligned of 43 3–4  months old (mice age on impact date: all animals, concussed or shams, to the same accuracy 13.2 ± 1.4 weeks) male mice. Mice were divided into four standard. cohorts: Sham (n = 14, n = 6 day 2, n = 3 day 7, n = 5  day Once the animal was secured on the platform, anaes- 14), concussion day 2 (CON 2; n = 9), concussion day 7 thesia was discontinued, and the trigger button imme- (CON 7; n = 10), concussion day 14 (CON 14; n = 10). On diately pressed to induce a concussive impact. Terminal day 0, all mice in the concussion group were exposed to piston velocity used ranged from 5.2 to 5.4 m/s. Mice in a concussive impact using our impactor device (see our sham control group underwent the same procedure with- earlier study [40] for detailed description). The sham out actual impact. animals underwent the exact same procedure but did not receive an impact. After the concussion or sham Behavioural assessment procedure, the loss-of-righting-reflex (LRR) time was Loss‑of‑righting‑reflex (LRR) measured for each of the animals. Thirty minutes after After the impact or sham procedure, the mice were recovery, the Neuro Severity Score (NSS) was measured removed from the restraints and laid in the supine posi- for each of the mice. At day 2, day 7, or day 14, depend- tion on a warmed surface for recovery. Loss-of-righting- ing on the cohort, the NSS was measured again and all reflex (LRR) time was the time (s) from the moment of mice underwent Open Field Assessment and MRI scan. impact to the first sign of the animal, righting itself to a Animals were excluded from this study if obvious brain prone position. injuries or structural abnormalities were observed on T2-weighted structural MRI images. All experiments Neuro Severity Score assessment were approved by the Institutional Animal Ethics Com- Thirty minutes after the first sign of righting reflex, all mittee at the University of Queensland (Animal Ethics mice were assessed based on a modified Neuro Severity Committee approval number QBI/260/17). Score (NSS). Another NSS assessment was performed on Data from this study is available, without reservations, the day of and prior to the MRI scanning session. NSS on request to the corresponding author. is a common assessment 10 tasks scale for neurological deficiency and a detailed description of the tasks can be Concussion procedure found in Flier et  al. [41]. These 10 tasks included: suc - The concussion model used in this study was used in one cessful escape from a 30  cm-diameter walled circle with of our prior publications [40], where more details can be one opening; natural seeking behaviour in an open area; found; the procedural description is mentioned for clar- lack of limping/dragging walking gait or grabbing weak- ity and ease for the readers. ness; presence of a straight walk and lack of an abnormal The animals were housed in the animal holding facil - gait; intact startle response to loud noise; 1 cm beam bal- ity with a 12-h light–dark cycle, with food and water ance; 3, 2, and 1 cm beam walk tasks, and 0.6 cm round available ad libitum. Animals were initially anaesthetised stick balance. An additional 6  mm beam walk task was with 3% isoflurane in 60% Air 40% O gas mixture at 2 L/ added, raising the number of task and the highest possi- min for 2 min. Mice were then transferred to the impac- ble score from 10 to 11. The final task described in Flier tor device and supported on the body plate. The animal et  al. was the round stick balance [41]; preliminary tests To and Nasrallah acta neuropathol commun (2021) 9:2 Page 4 of 20 showed several of our mice were sufficiently skilled and 0.1 × 0.1 × 0.3  mm, Repetition Time (TR) = 7200  ms , motivated to cross the stick similarly to beam walk tasks; Echo Time (TE) = 39 ms , averages = 4, RARE factor = 8, thus we decided to add a 6 mm round stick walk into our and bandwidth = 54.3478 kHz . modified NSS assessment. Resting‑state fMRI scans Open field assessment a 2D gradient-echo echo-planar-imaging (GE-EPI) Open field activity was assessed on the day of and prior sequence with the following parameters was used: to the MRI scanning session using an elevated circular matrix size = 64 × 64, FOV = 19.2 × 19.2  mm, 20 arena (diameter 77  cm) fenced by 32  cm high bound- slices of 0.5  mm thickness and 0.1  mm slice gap; giv- ary. The animals were placed in the centre of the arena ing an effective spatial resolution of 0.3 × 0.3 × 0.6  mm, and spontaneous activities were tracked and recorded TR = 1000  ms , TE = 14  ms , averages = 2, flip using an overhead camera and Tracker software (Bio- angle = 70°, bandwidth = 200  kHz, and 600 volumes Signal Group, NY, USA) for 10  min. The circular plat - were acquired with fat suppression, FOV satura- form was divided into four annuli (with annulus 1 at the tion (covering the head tissue inferior to the subject’s centre and annulus 4 at the periphery of the platform) brain), and navigator pulses turned on. A block-stimu- the fraction of total time the animal spent in each annu- lation design (40  s OFF, 20  s ON) fMRI scan was per- lus was quantified. The Thigmotaxis Index (TI)—an formed approximately 70 min after the initiation of the index for anxiety-like behaviours [42], was calculated as medetomidine bolus; stimulation fMRI scan contained TI = (T − T )/(T + T ) where T and ann4 ann1-3 ann4 ann1-3 ann4 six OFF–ON blocks plus a final 40  s OFF block, for a T represents the time spent in annulus 4 and com- ann1-3 total of 400  s. After the end of the stimulation fMRI bined time spent in annuli 1, 2, and 3, respectively. scan, a 5  min break was given before rsfMRI scan was started; rsfMRI scan duration was 600 frames for a total MRI experiments of 600  s. A single EPI scan with the same parameters Animal handling but opposite phase-encoding direction was acquired Anaesthesia was induced using 3% isoflurane in 60% for distortion correction of the fMRI images. air, 40% O at 1L/min. The isoflurane concentration was maintained at 2–2.5% during the preparation which took around 30 min. Each mouse was positioned on an MRI- Diffusion MRI scans compatible cradle (Bruker Biospin, Germany) with ear Diffusion data were acquired using a diffusion-weighted bars and bite bars to reduce head motion. A peritoneal Imaging (DWI) spin-echo echo planar imaging (SE- catheter was inserted and fixed to the mouse for delivery EPI) sequence with the following parameters matrix of medetomidine (Domitor, Pfizer, USA). For sedation, a size = 96 × 96, FOV = 19.2 × 19.2  mm, 32 slices of bolus of 0.05 mg/kg medetomidine was given intraperito- 0.25 mm thickness and 0.05 mm slice gap; giving output neally and then sedation was maintained with a continu- spatial resolution of 0.2 × 0.2 × 0.3  mm, TR = 4500  ms , ous infusion of 0.1 mg/kg/h. Once the animal was inside TE = 25  ms , averages = 4, 3 b-value shells with b = 600, the MRI scanner, isoflurane was then reduced gradually 1500, and 2000s/mm , 33 diffusion weighted directions and kept at approximately 0.25% throughout the experi- for each shell, and 2 b = 0 images. A pair of reference ment. The total time under anaesthesia for each animal b = 0 SE-EPI scans were acquired with opposite phase- was approximately 2.5 h. At the end of the scanning ses- encoding directions for EPI distortion correction. sion, 1.25  mg/kg atipemazole (medetomidine reversal) (Antisedan, Pfizer, USA) was given intraperitoneally. Behavioural data analysis Structural MRI scans MRI scans were performed All statistical analyses of behavioural data were per- on a 9.4T MRI scanner (Bruker Biospin, Germany) formed in Prism 8 (GraphPad Inc.). Statistically sig- equipped with a cryogenically cooled transmit and nificant threshold was set at p value < 0.05 (two-tailed). receive coil, controlled by a console running Paravision Analysis of Variance (ANOVA) post hoc tests were cor- 6.0.1 (Bruker Biospin, Germany). Structural imaging rected for multiple comparisons by controlling False data was acquired using a 2D T2-weighted (T2w) Turbo Discovery Rate (FDR). Rapid Acquisition with Refocused Echoes (Turbo- RARE) sequence with the following parameters: matrix Loss‑of‑righting‑reflex (LRR) size = 192 × 192, Field of View (FOV) = 19.2 × 19.2 mm, LRR duration was compared between sham animals number of contiguous slices = 52 and slice thick- and all injured animals using Mann–Whitney tests. ness = 0.3  mm; giving an effective spatial resolution of T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 5 of 20 Neuro Severity Score (NSS) affine-transformation-included Jacobian determinant The NSS score was compared between the sham and all was calculated from the subject-specific non-linear concussed mice at 30  min post-concussion using the registration warp; Jacobian determinant for each voxel Mann–Whitney rank test. The NSS of the concussed is a measure that indicate the relative volume change mice from the same cohort assessed at 30  min was required to warp the template voxel to the individual compared against the NSS on the day of the MRI scan voxel. Tensor-based morphometry (TBM) [46] was of that cohort using the Wilcoxon matched pairs signed performed using voxel-wise statistics on the Jacobian rank sum test. NSS of shams and CON cohorts assessed determinants to assess of local structure absolute vol- on imaging day were analysed using Kruskal–Wallis ume differences between groups, as opposed to the One-Way ANOVA with post hoc tests comparing each analysis of local structures normalised by total brain CON cohort with sham cohort. volume. Open field assessment Stimulus‑evoked and resting‑state fMRI data processing Time spent in each annulus of different cohorts were Functional MRI image registration Opposite phase- analysed using repeated measures Two-way ANOVA encoding direction EPI data were used to calculate the with Geisser–Greenhouse correction and post hoc warping field required for EPI distortion correction using tests comparing mean time spent in each annulus of FSL’s TOPUP tool [47, 48]. One distortion corrected EPI each cohort with every other cohort. TIs were analysed data was used as reference for an iterative direct EPI-to- using Brown–Forsythe and Welch One-way ANOVA EPI non-linear registration/template generation image with post hoc tests comparing mean TI of each cohort registration process. All subject’s distortion corrected with mean TI of every other cohort. EPI images were affinely registered to the chosen EPI reference using FSL’s FLIRT [45]; the registered images MRI data processing were then averaged to create the first iteration of a study- All MRI data were exported in DICOM format using specific EPI template. Each subject’s distortion corrected Paravision 6.0.1 before converting to NIFTI data for- EPI images were then non-linearly registered to the new mat using MRIcron (i.e., the dcm2nii tool) [43]. For the study-specific EPI template using FSL’s FNIRT tool [45]. remaining of the processing, the images were given a The non-linearly registered EPI images were averaged to header file with voxel size 20 times larger than the orig - create the second iteration study-specific EPI template. inal size [44]. The non-linear registration process was repeated one more time and the warping field from this iteration used Structural image processing and tensor‑based morphometry to warp the rsfMRI image to a common space. Recent (TBM) studies suggested that for the purpose of image regis- Structural image signal inhomogeneity correction was tration for functional MRI studies, traditional structural applied to the T w structural data using FSL’s FAST image-based step-wise registration of functional image tool (https ://fsl.fmrib .ox.ac.uk/fsl/fslwi ki) [45]. The to same-subject structural image then to common space structural image was then registered rigidly to the Aus- template can be replaced by direct faster EPI-to-EPI non- tralian Mouse Brain Mapping Consortium (AMBMC, linear registration without loss of accuracy [49]. www.imagi ng.org.au/AMBMC ) MRI template resam- pled to 0.1  mm isotropic resolution using FSL’s FLIRT Functional data pre‑processing Each functional data set [45]. The registered images from all animals were aver - underwent slice-timing correction with FSL’s slicetimer aged to create a study-specific “FLIRT template”. This tool, despiked using AFNI’s 3dDespike (https ://afni.nimh. study-specific template was used as the new template nih.gov/) tool, motion corrected using FSL’s MCFLIRT for another iteration of linear registration and “FLIRT tool (FMRIB’s Software Library) [45] with the forward template” creation. This process was repeated for three phase EPI image as the reference image. After motion cor- linear registration iterations before each individual rection, functional images were distortion corrected using structural data was then non-linearly registered to the the warping field obtained from EPI pair. Band-pass filter resulting third iteration study-specific “FLIRT tem - was applied on resting-state fMRI data at 0.01–0.3  Hz plate” using FSL’s FNIRT tool [45] and a study-specific using AFNI’s 3dTproject. Pre-processed stimulus-evoked “FNIRT template” was created from the non-linear reg- and rsfMRI data were warped to a common space using istered images. Individual structural images were non- warp field obtained from the iterative direct EPI-to-EPI linearly registered to the “FNIRT template” and the non-linear registration/template generation image regis- Jacobian determinant was extracted. Each individual tration process described above. To and Nasrallah acta neuropathol commun (2021) 9:2 Page 6 of 20 Functional MRI artefact removal Artefact removal sorted and functionally-relevant components were cho- from registered and pre-processed stimulus-evoked sen based on the following criteria: (1) component spatial and rsfMRI data were performed with group informa- maps clustered on spatially and functionally feasible grey tion guided independent component analysis (GIG-ICA) matter, (2) frequency power spectrum showed power [50], as implemented in the Group ICA of fMRI Toolbox largely in the 0.01–0.1 Hz range (or at least the power in (GIFT) [51]. The number of independent components for this frequency range was larger than those above 0.3 Hz), GIG-ICA was set to 50. GIG-ICA has been shown to be (3) component spatial maps had little or no overlap with superior to not performing artefact removal, or artefact white matter (WM) and cerebral-spinal fluid (CSF), (4) removal using individual-level Independent Component component spatial maps with bilateral or midline-only Analysis (ICA) followed by group ICA—a method recom- areas, or if an IC’s spatial map was unilateral; it was only mended the Human Connectome Project [52]—for arte- included if there was another contralateral IC, (5) ICs facts removal in the context of group ICA [53]. GIG-ICA were rejected as “functionally relevant” if the spatial map also avoids the need for potentially subjective, biased, and had a thickness in the rostral-caudal direction equivalent error-prone human classification of individual level ICA to one EPI slice in the acquired resolution (single-slice [53] or machine learning training of human-classified artefacts). data for automated signal/artefact classification. Artefac - Twenty four ICs were identified and grouped into nine tual components were classified based on criteria simi - anatomical/network groupings; list of component num- lar to prior publications [54, 55]: (1) component spatial bers, component IDs/abbreviations, approximate ana- map having a large overlap with white matter (WM) and tomical structures, and their corresponding anatomical/ cerebral-spinal fluid (CSF) or ring-like or crescent shape network groupings are listed in Table  1. More complete around the edge of the brain or near regions with EPI dis- spatial maps, time courses, and frequency power spectra tortion, (2) frequency power spectrum showing pan fre- for each of the group-level “functionally relevant” ICs are quency distribution, (3) component spatial maps showing shown in Additional file 1: Data 1. alternating positively and negatively correlation bands. Regional homogeneity (ReHo) and intrinsic local functional Independent component analysis (ICA) of functional MRI connectivity analysis Intrinsic local functional con- data Artefact cleaned stimulus-evoked fMRI was ana- nectivity were calculated for GIG-ICA artefact-cleaned lysed using temporal-concatenated spatial group ICA, rsfMRI data for all subjects using the regional homoge- using Infomax algorithm [56, 57], with the number of neity (ReHo) approach [64]. Seven-voxels neighbourhood components set to 18. An independent component with regional homogeneity (ReHo7) maps of each subject’s spatial maps (converted to Z-scores and thresholded at rsfMRI data were created by calculating Kendall’s coeffi - Z > 1) that includes the primary somatosensory area (S1) cient of concordance (KCC) of each voxel’s time course that was shown to be activated in prior fMRI publication with its six face-neighbour voxels using AFNI’s 3dReHo [58], and component time course that matched the stimu- tool. lation paradigm was chosen for further analysis. Artefact-cleaned rsfMRI data were decomposed into Diffusion MRI data processing resting-state functional independent components using Diffusion tensor imaging (DTI) data processing The Independent Vector Analysis (IVA), implemented as opposite phase-encoding direction EPI images of the combined IVA algorithm (IVA-GL) with multivariate diffusion data were used to calculate the warping field Gaussian (IVA-G) [59] source component vectors plus required for EPI distortion correction using FSL’s TOPUP IVA with Laplace source component vectors (IVA-L) tool [47]. This EPI distortion correction warping field was [60]. IVA-GL has been shown to be superior in identify- applied to diffusion data, which was then eddy current ing subject variability and detecting unique biomarkers distortion corrected and motion corrected using FSL’s in fMRI data, compared to more traditional group ICA eddy_correct. Diffusion tensor data was fitted using FSL’s [61–63], which is potentially more suitable more detect- DTIFIT tool using b-value = 1500 s/mm . ing subtle changes in the post-concussion brains. IVA-GL One distortion corrected individual-level FA image simultaneously decomposes artefactual-cleaned individ- was used as reference for an iterative direct EPI-to-EPI ual rsfMRI data into 100 Independent Components (ICs) linear and non-linear registration/template generation spatial maps and time courses per subject then calculates image registration process similar to that of rsfMRI EPI group-averaged spatial maps and time courses. Group data described above (“Stimulus-evoked and resting- averaged ICs spatial maps were scaled and converted state fMRI data processing” section). Warping fields to Z-scores and thresholded at Z > 1 and the ICs were obtain from FA to study-specific FA template FNIRT T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 7 of 20 Table 1 Resting-state functional connectivity supra-networks, network, and  components: list of  component numbers, component IDs/abbreviations, approximate anatomical structures, and  their corresponding anatomical/network groupings and functional supra-networks Supra-network Network group Abbreviations IC no. Anatomical areas Default mode–sensory–memory Anterior Cingulate—Retrosplenial 03—ACA-PL 03 Anterior Cingulate Area + Prelimbic (DMSM) Cortex axis (ACA RSN) area 06—RSN-ACA 06 Retrosplenial area + Anterior Cingu- late Area 52—RSN 52 Retrosplenial area 62—ACA 62 Anterior Cingulate Area 82—RSN 82 Retrosplenial area 89—IL 89 Infralimbic area Hippocampal—Subcortical 05—rHP—RSN—ATN—ACA 05 Subcortical spatial memory pathway memory circuit (HP MEM) (Retrohippocampal Area—Retro- splenial Area—Anterior Thalamic Nuclei—Anterior Cingulate Area) 08—DG 08 Dentate Gyrus 34—ATN 34 Anterior Thalamic Nucleus Visual—Auditory ( VA) 17—AUD 17 Auditory Area 46—VIS 46 Visual Area Primary Somatosensory Area (S1) 04—S1lwlmb 04 Primary Somatosensory Cortex (S1) (Lower Limbs, Trunks) + Posterior Parietal Association Area + Tempo- ral Association Area 11—S1uplmb 11 Primary Somatosensory Cortex (Upper Limbs) 31—S1u 31 Primary Somatosensory Cortex (unas- signed) + Visual Area Thalamus-polymodal association 10—TH-pmc-S1 10 Thalamic nucleus (polymodal cortex ( TH-pmc) association cortex) + Primary Somatosensory Cortex 30—TH-pmc-VA 30 Thalamic nucleus (polymodal association cortex) + Pretectal Region + Visual-Auditor y Area Striatal–motor (STR–MO) CPu 07—CPu 07 Caudate Putamen M1 14—M1 14 Primary Motor Cortex Salience–supplementary soma- Salience Network (SN) 13—S1BF 13 Primary Somatosensory Cortex (Bar- tosensory (SAL–SS) rel Field, Lower Limbs) + Primary Motor Area 37—INS-EP 37 Insula + Endopiriform Nucleus 67—AM 67 Amygdala S2 02—S2-a 02 Supplemental Somatosensory cortex (S2)—anterior (bi lateral), Primary Somatosensory Cortex (Nose, Mouth) 18—S2-p-R 18 Supplemental somatosensory cor- tex—Right 45—S2-p-L 45 Supplemental somatosensory cortex—Left registration were used to warp other DTI metric images proje cts/noddi _toolb ox) [28, 29]; to calculate NODDI to the common space: MD, Dp, and Dr. metrics. In this in  vivo NODDI data fitting, neurites were modelled as impermeable sticks (cylinders with Neurite orientation dispersion and density imaging data zero radius) in a homogeneous background, neurite ori- processing Multi b-value shell data were fitted using entation distribution was modelled as Watson’s distribu- NODDI MATLAB Toolbox (https ://www.nitrc .org/ tion, and the algorithm estimate the hindered diffusiv - To and Nasrallah acta neuropathol commun (2021) 9:2 Page 8 of 20 ity from the free diffusivity, and neurite packing density for further analysis. Another three ROIs were identified using Szafer et  al.’s [65] tortuosity model for randomly to ReHo7 KCC significantly correlated with TI through packed cylinders. NODDI metrics were also warped to voxel-wise correlation analysis: the Caudate Putamen the common space using FA to study-specific FA tem - (CPu), the Amygdala (AM), and the Insula (INS). Simple plate FNIRT registration warp field. Registered DTI and linear regression between DTI/NODDI metrics and KCC NODDI metrics images were smoothed to an estimated quantified from the corresponding ROIs and TI were spatial smoothness of 0.6  mm FWHM using AFNI’s performed on the basis of sham + CON day 2 and CON 3dBlurToFWHM. day 7 + CON day 14, and thresholded at p value < 0.05. T-statistics of two-samples t tests of NSS and TI behav- Statistical analysis of MRI data ioural measures and averaged whole-brain t-statistics Two sample statistical inference of CON cohorts versus sham maps of two samples t tests of Jacobian determinants, cohort DTI/NODDI metrics, and ReHo7’s maps were quantified Voxel-wise Two-samples t tests comparing CON cohorts for each CON cohort comparison with sham cohort and against sham cohort were performed on registered Jaco- the relative behavioural versus MRI changes were plotted bian determinants, DTI and NODDI metrics, S1 inde- for comparison. pendent component spatial maps, and ReHo7 KCC maps using permutation inference for the General Lin- Results ear model [66] as implemented in FSL’s randomise [67]. Behavioural measures The number of permutation was set to 10,000, as recom - Loss‑of‑righting‑reflex (LRR) mended by a prior study [68]. The resulting statistical A significant increase in the LRR time was seen in the map were corrected for multiple comparisons with mass- CON groups 125.6 ± 28.81 (mean ± standard error of based FSL’s Threshold-free Cluster enhancement (TFCE) mean [s]) compared to the sham group 34 ± 2.65 (p [69] and thresholded at p value < 0.05 (two-tailed). value = 0.0005) (Fig. 1a). Reproducibility of DTI and NODDI were tested by ran- domly dividing the sham cohort into two pseudo-groups Neuro Severity Score (NSS) (Fig. 1b) and two-samples t test comparing the pseudo-groups At 30  min post-concussion, all concussed animals had were performed using identical procedure as describe significantly higher NSS 1.68 ± 0.11 compared to sham above. The resulting statistical maps are shown in Addi - animals 0.62 ± 0.13 (p value < 0.0001). NSS of CON day tional file 3: Fig. S3. 2 on the imaging day was significantly higher than that One and two-samples statistical tests were conducted of the sham cohort (1.44 ± 0.24 vs. 0.29 ± 0.12; FDR’s q on individual level IC time courses using GIFT’s Man- value = 0.002). For CON day 7 and CON day 14, there covan toolbox [70] for IC–IC functional network con- was no significant difference in the NSS score at day 7 and nectivity (FNC) averages for each cohort and differences day 14 post-injury compared to that of the sham cohort; between CON and sham cohorts. T test for component– CON day 7 (0.7 ± 0.34; FDR’s q value = 0.248) and CON component FNC were performed using permutation day 14 (0.6 ± 0.22; FDR’s q value = 0.248). There was no inference for the General Linear model [66] as imple- significant difference between the NSS measured at day mented in FSL’s randomise package [67], with the num- 2 (1.44 ± 0.24) or day 7 (1.44 ± 0.24) compared to the ber of permutation set to 50,000 or exhaustive, whichever 30 min NSS in the CON day 2 (1.67 ± 0.17, p value = 0.35) is smaller, and corrected for multiple comparisons with and CON day 7 cohort (1.6 ± 0.3, p value = 0.0625). A sig- False Discovery Rate at q value < 0.05 for one-sample t nificant difference between the NSS measured at day 14 test and q value < 0.1 for two-sample t test. (0.6 ± 0.22) and the 30  min of the CON day 14 cohort (1.9 ± 0.18, p value = 0.0039) was seen. Correlation of behavioural symptoms with local tissue volumes and DTI and NODDI metrics Open field test Registered DTI/NODDI metrics, and ReHo7 KCC maps Two-way ANOVA showed significant annuli (p were correlated with NSS and TI obtained on scanning value < 0.0001) and annuli × cohort effect (p value = 0.03) day, across all subjects. Implementation of voxel-wise sta- but an insignificant cohort effect (p value = 0.2) on tistics were similar to two-sample t tests described above. the fraction of time spent in each annulus. CON day Three regions of interest (ROIs) were identified to have 7 cohort animals spent significantly more time in the DTI and NODDI metrics significantly correlated with TI annuli towards the centre compared to CON day 14 ani- through voxel-wise correlation analysis: the corpus cal- mals and significantly more time in annulus 2 than sham losum (cc), the polymodal association area of the thala- animals (Fig.  1c). There was significant difference of TIs mus (TH-pmc), and the Anterior Cingulate Area (ACA) among different cohorts (p value = 0.022) and CON day T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 9 of 20 Fig. 1 Behavioural assessment of sham and CON cohorts: a LRR time of sham (n = 14) and CON (all 3 CON cohorts, n = 29), two-tailed Mann– Whitney test. Lines showing median and interquartile range. b NSS of experimental cohorts at 30 min post-concussion and on imaging days: sham (n = 14), CON day 2 (n = 9), CON day 7 (n = 10), and CON day 14 (n = 10). Sham versus all CONs at 30 min: two-tailed Mann–Whitney test. Intra-cohort, 30 min versus imaging day: two-tailed Wilcoxon matched-pairs signed rank test. Inter-cohort NSS on imaging day: Kruskal– Wallis One-Way ANOVA with FDR-corrected (q < 0.05, two-tailed) post hoc tests. Data point displayed as mean and standard errors of means. c Proportion of time spent in each of four annuli in Open Field Test. Repeated Measures Two-way ANOVA with Geisser–Greenhouse correction and FDR-corrected (q < 0.05, two-tailed) post hoc tests. d Thigmotaxis Index of each cohort quantified from Open Field Test on imaging day. Brown– Forsythe and Welch One-way ANOVA with FDR-corrected (q < 0.05, two-tailed) post hoc tests. *P value < 0.05, **P value < 0.01, ***P value < 0.001 7 cohort to have significantly lower TI compared to sham in the internal capsule (Fig.  2b, Dp). Increased NDI was (FDR’s q value = 0.03) and CON day 14 cohorts (FDR’s q detected in the grey matter—the visual, auditory, retro- value = 0.0009) (Fig. 1d). splenial, and S1 areas, and the hypothalamus (Fig.  2b, NDI). Decreased ODI was detected in both grey matter— MRI findings the visual, auditory, retrosplenial, hippocampal, and S1 Diffusion MRI findings areas, the striatum, and the anterior cingulate cortex— No significant changes of brain tissues’ DTI and NODDI and white matter—the external capsule (Fig. 2b, ODI). metrics were detected at day 2 post-concussion com- At day 14 post-concussion, DTI and NODDI changes pared to shams (Fig.  2a). At day 7 post-concussion, were still although to a lesser extent compared to day increased FA was detected at day 7 post-concussion com- 7. Increased FA was detected in both grey matter—the pared to shams in both grey matter—the visual, auditory, auditory area and anterior cingulate cortex—and white retrosplenial, hippocampal, and Primary Somatosensory matter—the external capsule (Fig.  2c, FA). Decreased Area (S1), the thalamus, the hypothalamus, the stria- MD, driven by both Dp and Dr decreases was found tum, and the anterior cingulate cortex—and white mat- in the S1, striatum, and internal capsule (Fig.  2c, MD, ter—the external capsule (Fig.  2b, FA). This increased Dp, and Dr). Increased NDI was detected over the same FA in the grey matter was driven mostly by reduced MD regions that had decreased MD, Dp, and Dr (Fig.  2c, and Dr, in the visual, auditory, retrosplenial and S1, the NDI). Decreased ODI were detected in small areas in thalamus, the hypothalamus, and the anterior cingulate the hippocampus, external capsule and corpus callo- cortex (Fig. 2b, MD and Dr). Decreased Dp was detected sum (Fig. 2c, ODI). To and Nasrallah acta neuropathol commun (2021) 9:2 Page 10 of 20 Fig. 2 DTI and NODDI metric changes from day 2 to day 14 post-concussion. Voxel-by-voxel statistical analysis results of Diffusion Tensor Imaging (FA = Fractional Anisotropy, MD = Mean Diffusivity, Dp = Parallel Diffusivity, Dr = Radial Diffusivity) and Neurite Orientation Dispersion and Density Imaging metrics (NDI = Neurite Density Index, ODI = Orientation Dispersion Index) and Tensor-based Morphometry with Jacobian Index (JI) of a CON day 2 (n = 9) versus sham (n = 14), b CON day 7 (n = 10) versus sham (n = 10), and c CON day 14 (n = 10) versus sham (n = 14). Statistical map thresholded at P value < 0.05 (two-tailed), unpaired two sample t test, implemented as permutation tested for the General Linear Model, corrected for multiple comparisons with mass-based FSL’s Threshold-free Cluster enhancement ( TFCE). Statistical maps were overlaid on the averaged and registered DTI and NODDI metrics maps corresponding to the statistical maps (DTI and NODDI results) and structural template ( TBM results). Corresponding grey scale map for each averaged DTI and NODDI metrics maps were provided; units for Dp, Dr, and MD were in mm/s . ACA = Anterior Cingulate Area, AM = Amygdala, AUD = Auditory Area, cc = corpus callosum, ec = external capsule, HP = Hippocampus, HYP = Hypothalamus, ic = internal capsule, INS = Insula, MB = Midbrain, PAL = Palladium, S1 = Primary Somatosensory Cortex, RSN = Retrosplenial Area, STR = Striatum, TH = Thalamus, VIS = Visual Area. Red anatomical orientation marker L = Left, R = Right. An enlarged view of the DTI and NODDI metrics and the identified regions of interest (ROIs) can be found in Additional file 3: Fig. S3 Structural MRI findings of the S1 IC spatial maps was detected at day 7 and 14 TBM detected increased ventricle volumes and local tis- post-concussion (Fig.  3b, c), indicating increased neural sue volumes in the auditory, visual, and primary soma- responses to somatosensory stimulation. tosensory areas, the hippocampus, anterior cingulate cortex, and the amygdala at day 2 post-concussion com- Resting state functional network connectivity architec‑ pared to shams (Fig. 2a, JI). At day 7, a decrease in tissue ture Three main supra-networks based on the con - volume was seen in the midbrain, the hypothalamus, the nectivity architecture: default mode–sensory–memory amygdala, and the insula (Fig.  2b, JI). No changes were (DMSM), the salience–supplementary sensory (SAL– seen at day 14 (Fig. 2c, JI). S2), and the striatal–motor (STR–MO) supra-network (Fig. 3d). Supra-network refers to a collection of networks Functional MRI findings in this study determined to have some common charac- Stimulus‑evoked fMRI changes post‑concussion Group teristics in terms of connectivity to other networks within ICA of stimulus-evoked fMRI showed no difference in the same supra-network while different connectivity char - spatial extents of the S1 IC spatial maps associated with acteristics in regard to networks in other supra-networks. functional activation during forepaw stimulation at day The DMSM supra-network consisted of Default 2 post-concussion (Fig.  3a). An increased spatial extent Mode Network-like ICs along the Anterior T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 11 of 20 Fig. 3 Functional MRI of concussed and sham animals. a–c Post-concussion stimulus-evoked fMRI activity changes of a CON day 2 versus sham, b CON day 7 versus sham, c CON day 14 versus sham. h–j Post-concussion changes in local intrinsic functional connectivity in (H) CON day 2 versus sham, (I) CON day 7 versus sham, and j CON day 14 versus sham. a–c, h–j Statistical map thresholded at p value < 0.05 (two-tailed), unpaired two sample t test, implemented as permutation tested for the General Linear Model, corrected for multiple comparisons with mass-based FSL’s Threshold-free Cluster enhancement ( TFCE). Statistical maps were overlaid on the study-specific averaged EPI images. Red anatomical orientation marker L = Left, R = Right. ACA = Anterior Cingulate Area, AM = Amygdala, AUD = Auditory Area, CPu = Caudate Putamen, GP = Globus Pallidus, HP = Hippocampus, INS = Insula, M2 = Secondary Motor Cortex, MB = Midbrain, PRT = Pretectal area, S1 = Primary Somatosensory Cortex, RSN = Retrosplenial Area, TH-pmc = Thalamus-polymodal association cortex, VIS = Visual Area. d Average functional network connectivity (FNC) matrices among Independent Components (ICs) identified by IVA-GL of (a) sham (n = 14). Colour scaled by z test statistics; non-black cells were defined as component–component connectivity deemed statistically significant. One sample t tests, permutation-tested, and FDR-corrected (q value < 0.05, two-tailed). e–g Post-concussion FNC changes of e CON day 2 versus sham, f CON day 7 versus sham, g CON day 14 versus sham. Colour scaled by z test statistics; non-black cells were defined as component–component connectivity deemed statistically significant. Two sample t tests, permutation-tested, and FDR-corrected (q value < 0.1, two-tailed) Cingulate–Retrosplenial Cortex axis (ACA RSN group: with earlier whole brain group ICA functional network IC 03, 06, 52, 62, 82, 89), Hippocampal–Subcorti- characterisation of the mouse brain [55]. IC 05 spatial cal memory circuit (HP MEM group: IC 05, 08, and maps showed several regions strongly connected to 34), Primary Somatosensory and Association Area one another: the retrosplenial area (RSN), the retro- (S1 group: IC 04, 11, and 31), Visual–Auditory areas hippocampal (rHP) area, the Anterior Thalamic Nuclei (VA group: IC 17 and IC 46), and Thalamic polymodal (ATN); this IC was also positively correlated with the association cortex (TH-pmc group: IC 10 and 30). anterior cingulate and retrosplenial ICs (IC 03, 06, and The strong functional connectivity among CG RSN 52), and the retrohippocampal IC (IC 08). These con - and VA areas were consistent with the earlier char- nections resemble the cortical memory pathway of acterisation of the Default Mode Network (DMN) in HP > RSN > ACA and a subcortical memory pathway the rodents [71, 72] and the overall strong connectiv- of HP > ATN > RSN/ACA, that support head direc- ity within the DMSM supra-network was consistent tion, spatial navigation/memory/coding, and emotion To and Nasrallah acta neuropathol commun (2021) 9:2 Page 12 of 20 [73–75]. Resting-state functional connectivity was with TI. More detailed linear regression revealed that Dp known to have a basis in structural connectivity of and ODI in these ROIs were significantly correlated with direct nerve projections [76]. TI only among CON day 7 and CON day 14 animals but The SAL–SS supra-network consisted of the Primary not among the sham and CON day 2 animals (Fig. 4a–f ), Somatosensory Area, Barrel Field (S1BF, IC 13), the with the exception of ODI in the TH-pmc (Fig.  4e) and Supplementary Somatosensory Areas (S2 group: IC 02, ACA (Fig.  4f). ODI was negatively correlated with TI 18, and 45), and the Salience Network (SN group: IC 37 among sham and CON day 2 cohorts but not among [Insula area] and IC 67 [Amygdala]). The SAL–SS supra- CON day 7 and day 14 cohort. ODI was not correlating network is characterised by positively correlated func- with TI in the ACA. tional connectivity of ICs within the supra-network and Voxel-wise correlation analysis also defined three negatively correlated functional connectivity of its ICs ROIs—CPu, AM, and INS that showed significant cor - with ICs of the DMSM supra-cluster. relation between KCC and TI. More detailed analysis The STR–MO supra-network consisted of the Caudate showed KCC were positively correlated with TI in all Putamen (CPu, IC 07) and the Primary Motor Area (M1, three ROIs among CON day 7 and CON day 14 animals IC 14). The distinguishing feature of this supra-network but not among sham and CON day 2 cohorts (Fig. 4g–i). is its components have a positive correlation within the supra-network and with the other two supra-networks. Discussion One-sample t test results matrices of IC–IC FNC of This work shows a mismatch in the onset and recovery CON cohorts are shown in Additional file 2: Fig. S2A-C. of the structural and functional attributes of the brain following a concussion, evidenced by advanced neuro- Whole brain restingst ‑ ate functional connectivity changes imaging, beyond that of symptom subsidence. Motor post‑concussion Figure  3e–g reflect the two-sample deficits were evident immediately after concussion while t test comparison of the CON cohorts compared to the deficits in learning were only evident at day 7; all symp - sham cohort. At day 2 post-concussion, concussed mice toms resolved at day 14. Imaging findings varied; DTI had increased functional connectivity between the M1 and NODDI changes peaked at day 7 and significantly and S1, upper limbs (Fig. 3e). At day 7 post-concussion, no reduced at day 14 post-concussion while the functional IC–IC FNC connectivity change was detected (Fig. 3f ). At connectivity increased at day 2 and 14 post-concussion. day 14 post-concussion (Fig. 3g), the increased connectiv- Additionally, stimulus-evoked fMRI detected no differ - ity was detected in the CPu—Auditory Area (AUD), Left ences at day 2 but increased cortical activation at day 7 S2—ACA, and Amygdala (AM)—Insula (INS) connec- and 14 post-concussion. tions. Behavioural changes following concussion Local intrinsic functional connectivity changes post‑con ‑ Motor-balance deficits were significant following a sin - cussion At day 2 post-concussion, concussed mice had gle concussive injury, as evidenced by lower NSS at day increased local functional connectivity in several regions 2 post-concussion. Motor-balance symptoms partially in the DMSM supra-network (S1, TH-pmc, and HP), as recovered at day 7 and fully recovered at day 14. This well as the Salience-like network (INS and AM), and the recovery timeline of motor-balance symptoms in our CPu (Fig.  3h). The increased local connectivity at day 2 model of rotational, acceleration/deceleration concussive subsided at day 7 post-concussion, with only decreased injury is consistent with earlier studies of repeated or sin- local connectivity in a small area of the hippocampus gle mild impacts of the CHIMERA model, which showed (Fig.  3i). However, at day 14 post-concussion, increased injured animals recovered their motor balance function local functional connectivity were detected in many at approximately day 14 [77, 78]. regions of the brain, including the regions in the DMSM Concussed animals displayed reduced TI and spent supra-network–DMN (ACA, RSN), VA (VIS, AUD), and more time at the centre of the area compared to shams S1 networks, the SAL–SS supra-network–SN (INS and at day 7 post-concussion; this behaviour normalised at AM), and the CPu (Fig. 3j). day 14. This trend of reduced TI and more time spent in Averaged ReHo7 maps of sham and CON cohorts are the centre of the arena was consistent with similar results shown in Additional file 2: Fig. S2D-G. after a single moderate injury CHIMERA impact [79]. On the other hand, increased TI, anxiety-like behaviours, Correlation of MRI findings with behavioural symptoms and less time spent at the centre of the arena during an Voxel-wise correlation analysis defined three ROIs—cc, open field task were observed at day 1 and day 7 after a TH-pmc, and ACA of interest, that showed Dp were single [78] or two consecutive mild CHIMERA impacts positively correlated and ODI were negatively correlated 24 h apart [77] and recovery occurred at day 14 [77]. The T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 13 of 20 Fig. 4 Correlation of diffusion and functional MRI with Open Field behavioural measures. Simple linear regression analysis of Parallel Diffusivity (Dp) (a–c), Orientation Dispersion Index (ODI) (d–f), and Kendall’s Coefficient of Concordance (KCC) from seven voxels neighbourhood Regional Homogeneity analysis of resting-state functional MRI (g–i), quantified from different regions of interest, the corpus callosum (cc—a, d), polymodal association area of the Thalamus ( TH-pmc—b, e), Anterior Cingulate Area (ACA—c, f), Caudate Putamen (CPu—g), Amygdala (AM—h), and Insula (INS—I) with Thigmotaxis Index ( TI). Regression analysis were performed across sham and CON day 2 cohorts combined (blue lines) or CON day 7 and CON day 14 combined (orange line). Regression lines displayed as mean and 95% confidence intervals slopes. ns = not significant, *p value < 0.05, **p value < 0.01 majority of other closed-head injury model studies also reported by human studies showing enlarged cortical showed opposite trends of increased anxiety-like behav- structures in the short-term—from 24 h [85], 7 days [86] iours and decreased time spent in the centre of the arena or up to 5.7  years post-concussion [87]. The majority of at 2 days [80–82] or up to one month post-injury [20, 83]. structural imaging studies in humans reported shrinkage Ertürk et  al. observed no alternations from day 4 up to in brain tissue volume after mild to moderate TBI (see 8 weeks post-injury [84]. Ross et  al. [87] for review), though a recent study com- paring a large number of mild to moderate TBI patients Structural brain changes following concussion (n = 50) to the USA’s Food and Drug Administration- Structural imaging and brain morphometry showed approved NeuroQuant normal control database [88, 89] surprising dynamic post-concussion changes that cor- revealed those patients have enlarged cortical grey mat- related well with motor-balance symptoms (NSS) and ter, cerebellar white matter, and hippocampal volumes anxiety-like behaviours (TI) to a lesser extent. At day 2 [87]. The short-term increased brain tissue volume and post-concussion, TBM analysis detected enlargement in lack of long-term or permanent atrophy in our model can the ventricles and several structures, including the audi- be explained by the lack of diffuse axonal injury; DTI and tory, visual, and S1, the hippocampus, anterior cingulate NODDI detecting no changes characteristic of diffuse cortex, and the amygdala. Similar findings have been axonal injury in our model and the correlation between To and Nasrallah acta neuropathol commun (2021) 9:2 Page 14 of 20 diffuse axonal injury and brain atrophy after TBI were optic tracts of a mouse model of closed-head injury at known in human TBI [90, 91]. Brain structural and mor- 1, 6, 12, and 18  weeks post-injury, even when mem- phological changes in this model occurred as early as 2 ory deficits resolved within the first week post-injury; days post-injury and correlated well with behavioural persistent neuroinflammation, including astroglio - symptoms. While this might seem very quick, this is con- sis and microgliosis were associated with the DTI and sistent with our earlier study in the same model [40] and NODDI changes in this model [21]. On the other hand, other studies which showed brain morphological changes a rat model of a single mild modified controlled corti - could happen as early as 24-h post-intervention [92–94], cal impact showed no changes at day 2 post-injury but ranging from changes in oestrous cycle [92], environ- elevated FA and reduced MD and Dr in the corpus mental enrichment [93], and maze training [94]; though callosum and external capsule at day 7 post-injury— the precise underlying mechanism is unclear. consistent with our results—though the changes nor- malised at day 14 [22]. Increased FA was also found in a rat drop weight model of mild traumatic brain injury Diffusion MRI changes in the grey mater after concussion 7 days post-injury [23]. Diffuse increases in FA in the grey matter of concussed In Mierzwa et  al., demyelination of degenerating animals were observed at day 7 post-injury. Increased FA and intact axons was found at day 3 post-injury in a in the grey matter is known to be associated with per- mouse model of a single mild closed-head injury, and sistent post-concussive symptoms (PPCS) following a was followed by remyelination and excessive myelina- concussion in human patients [24], consistent with our tion (including double-layered myelin sheaths) between results. Increased FA in the grey matter is also consistent day 7 and day 14 post-injury [99]. The process of remy - with other in vivo animal studies [25–27]. Closer analysis elination of previously demyelinated axons in a cupri- showed this increased FA in the grey matter was primar- zone-fed mouse model of demyelination showed that ily driven by a decrease in Dr and MD without as much the remyelination process decreased MD and Dr but decreased Dp. This pattern of elevated FA associated no changes to Dp were observed [100]. The observed with reduced Dr and mostly unchanged Dp in the grey decreased Dr, MD, and ODI in our model at day 14 matter, is consistent with the majority of human PPCS post-concussion suggested the possibility of the mice showing elevated grey matter FA [24], and rat closed- having excessive myelination in the corpus callosum head injury models [27]. A minority of human PPCS as a result of injury repair/recovery. Increased FA and cases [24], rat open-head injury model [25] or repetitive decreased MD have also been observed in non-injury mild blast exposure model [26] report elevated grey mat- neuroplasticity processes (a spatial learning task) ter FA, which is often associated with increased Dp with- involving increased myelination and myelin basic pro- out changes in Dr. Regardless of the patterns of Dp or Dr tein expression [101]. changes, this elevated FA in the grey matter post-TBI has Currently, only a few NODDI studies in human con- been associated with astrogliosis [25, 27] or microstruc- cussion have been published to date and the results tural remodelling in a rat model of open-head injury [95]. have been contradictory. Wu et  al. [30] and Church- Of particular interest is the association of DTI/NODDI ill et  al. [31] detected decreased NDI in the white changes and astrocyte activity, since astrocytes have dual matter of concussed patients imaged no longer than roles in neuronal plasticity and reconstruction after trau- 2  months post-injury. On the other hand, our patterns matic brain injury [96, 97]. of increased NDI and decreased ODI in the white mat- ter were consistent with Churchill et  al. [32] where Diffusion MRI changes in the white mater after concussion increased NDI and decreased ODI were detected in the Increased FA and decreased ODI were detected in white matter and grey matter-white matter boundaries the white matter tracts at both day 7 and day 14 post- of young, healthy athletes with a history of concus- concussion in our model. Decreased FA in the white sion (average of two concussions) imaged, on average, matter is commonly associated with reduced white 24  months post-injury. In our study, increased FA, matter integrity in Alzheimer’s disease [98], and in decreased ODI, and increased NDI were found in the animal models of repeated mild TBI [17, 18]. Other white matter of concussed mice. These findings coupled single-impact rodent closed-head injuries also showed with the relatively young age of the animals used (aver- decreased FA in the white matter: drop-weight model age age at impact or sham procedure 13.2 ± 1.4  weeks) in the rats [19] and piston-driven closed-head injury suggested the age range and injury profile at day 7 and in the mice [20, 21] at 1 and 8 days post-injury [19] day 14 post-concussion of our animals fit with that of or up to a month [20], and 18  weeks post-injury [21]. young, healthy humans imaged on average 24  months Decreased FA and increased ODI were found in the T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 15 of 20 (and at least 9 months) post-concussion. Churchill et al. or anti-correlations were found in the absence of global [32] suggested increased NDI and decreased ODI in signal regression during the pre-processing steps [105]. the white matter of young, healthy and fit athletes post- A number of human studies showed the persistence of concussion and without concussion symptoms was an anti-correlated connections, even without global signal indication that the axons recovering from injury were regression, in ROI-based [106–108], graph theoretical strengthened in the long-term after injury. [109], and independent component analysis [70]. Pat- terns of S2 cortex negative correlation with areas in Brain functional connectivity changes the DMSM were found in mice [55] and rats [110]. The following concussions human SN, which is primarily comprised of the insula, Initially, there was an increase in functional connectivity amygdala, and dorsal anterior cingulate cortex [111], has between the M1 and the S1 upper limbs and increased been shown to also have anti-correlation with the human local intrinsic connectivity in the S1, TH-pmc, and HP, DMN [70, 112], The human Dorsal Attention Network is SN, and the CPu. This increased long-range and local known to be another network that is anti-correlated with functional connectivity was normalised at day 7 post- the human DMN [112]; thus, the S1BF and the S2 identi- concussion, when motor-balance deficits mostly sub - fied in the mouse brains may play an analogous role to sided but psychological symptoms started. At day 14, the human Dorsal Attention Network. all symptoms were normalised but increased long-range and local functional connectivity were found within the Multi-phase brain recovery after concussion SN. Our results of initial increased functional connec- We observed a biphasic pattern of neural recovery and tivity associated with significant motor-balance deficits plasticity post-injury, each with distinct patterns of were consistent with some smaller studies that showed behavioural symptoms and associated MRI findings significant rsfMRI changes associated with post-con - (Fig. 5). cussion symptoms [34, 36, 37, 102]; the trends and net- In the first phase of post-concussion recovery, up to day works most frequently involved in studies were reduced 7 post-injury, brain recovery was mostly related to func- connectivity in the posterior DMN [34, 102], increased tional compensation. In this phase, significant motor-bal - connectivity in the anterior DMN [34, 102], reduced anti- ance symptoms were evident, which were associated with correlation among networks with anti-correlation rela- increased long-range and local functional connectivity tionships [102], and decreased local intrinsic functional with no concurrent changes in stimulus-evoked response connectivity in the SN [36, 37], in the lateralised cognitive detected with task-based fMRI; this can be interpreted as control network [37], and in regions related with motor, the brains’ increased activity to functionally compensate sensorimotor, attention, and phonological processing for the injury prior to injury recovery, healing, or plastic- [36]. On the other hand, Meier et  al. reported short- ity [113]. term elevated local intrinsic functional connectivity in In the second phase of concussion recovery, between regions associated with the DMN that normalised upon day 7 post-injury and day 14, brain recovery was domi- symptom recovery [39]. Other studies in humans have nated by neural plasticity. In this phase, particularly at also found persistent rsfMRI changes beyond symptom day 7 post-concussion, there was significant psycho - recovery [35, 103, 104], with one study reporting no sig- logical symptoms in the mice, evidenced by the open nificant changes to rsfMRI connectivity when symptoms field assessment, and DTI/NODDI detected changes were highest, but ongoing rsfMRI changes after symp- consistent with astrogliosis and neuroinflammation toms had recovered [103]. Our results demonstrated [25]. Interestingly, the interaction of astrocytes with persistent local and long-range functional connectivity injured neurons has been shown to result in hyperex- changes at day 14 despite resolution of symptoms. Our citable neurons [96]. This change in cortical excitability findings of no functional connectivity changes and signif - was observed in a rat model of controlled cortical injury, icant psychological symptoms at day 7 post-concussion, with injured animals showing increased stimulus-evoked were also consistent with human studies reporting no fMRI activation between 1 and 4 weeks post-injury rsfMRI changes despite significant ongoing symptoms as [114]. Furthermore, Verley et  al. showed bilateral hyper- our results [38, 103]. excitability to a unilateral forepaw stimulation that was A number of novel networks were defined in this observed in the first week before functional reorganisa - study as part of the resting-state functional connectiv- tion occurred and consolidated to unilateral hyperex- ity architecture through the patterns of mostly positive citability from week 2 to 4 [114]. Our stimulus-evoked correlation among networks within the same supra-net- fMRI results showed increased fMRI responses at day 7 work and negative or anti-correlation among networks and day 14 post-injury, and also demonstrated bilateral of different supra-network by contrast. These negative response enhancement in the CON day 7 cohort and To and Nasrallah acta neuropathol commun (2021) 9:2 Page 16 of 20 Fig. 5 Deficit and recovery trajectories of behavioural and magnetic resonance imaging (MRI) markers post-concussion. Relative changes of behavioural and MRI metrics of concussed cohorts relative to the sham cohorts (horizontal black dotted line through 0). Behavioural measurements (NSS and TI) scaled as t-statistics of the corresponding CON versus sham comparison. MRI biomarkers (JI, FA, NDI, ODI, KCC, and stim-fMRI) scaled as whole-brain averaged t-statistics of the corresponding CON vesus sham comparison; whole-brain t-statistics were used as a biomarker proxy that incorporated both extents and degrees of change. NSS = Neuro Severity Score, TI = Thigmotaxis Index, JI = Jacobian Index, FA = Fractional Anisotropy, NDI = Neurite Density Index, ODI = Orientation Dispersion Index, KCC = Kendall’s Coefficient of Concordance (resting-state functional Magnetic Resonance Imaging Regional Homogeneity), and stim-fMRI (stimulus-evoked functional Magnetic Resonance Imaging). An alternative version of figure separating behavioural and MRI changes can be found in Supplementary data Fig. S4 unilateral enhancement in the CON day 14 cohort, sug- assess depression, anxiety, irritability, aggression [116], gesting the second phase of recovery being reflective of and hyperarousal [117] symptoms should be included. changes in neuroplasticity. The correlation of MRI biomarkers and the neurological The post-concussion cortical hyperexcitability detect - processes that supposedly underlie the two brain recov- able with stimulus-evoked fMRI may explain the appar- ery phases were hypothesised based on MRI biomarker ently contradictory trajectory of resting-state functional fingerprinting of prior studies, often of different brain connectivity in our model through different timepoints. injury models. Nevertheless, the specific correlation of a The increased functional connectivity at day 2 without histopathological process with a specific trend of diffu - concurrent cortical hyperexcitability possibly reflected sion imaging metric change is a complicated issue. For functional compensation [113] without neural recovery/ example, a specific pathology may result in a predictable plasticity. The increased functional connectivity at day biomarker on DTI or NODDI; however, different patholo - 14 with increasing cortical hyperexcitability might have gies may create the same diffusion metric change [118]. been the result of ongoing neural plasticity in the brains Furthermore, in complex conditions like concussion or [114]. NODDI patterns of increased NDI and decreased traumatic brain injury, pathologies generally do not occur ODI at day 7 and 14 also supported hypothesis of neural alone. Astrogliosis, microgliosis, and axonal injury gener- plasticity occurring during this phase. Furthermore, DTI ally occur together as a consequence of concussive injury changes in the white matter also supports neuroplasticity in mice [40, 77, 78, 115, 119]. As described in our earlier processes that may include remyelination or excessive publication [40] DTI and NODDI metrics only correlated myelination. with microgliosis, similar to a NODDI examination of a model of microglia depletion and repopulation [120] in Limitations regions with just neuroinflammation. A more thorough The study suffers from a number of limitations. The investigation to include biological measures would be behavioural assessments applied were restricted to a small needed to further validate the findings in this work. number; rotarod can be used in addition to NSS to allow There are inherent limitations to extrapolating find - for more fine-grained scoring of motor-balance func - ings in animal models to human pathologies and TBI tion/deficit [77], especially since the NSS range in this models are of no exception. There are differences in model was quite narrow. Learning and memory assess- gross anatomy between rodents and humans (lack of ments, such as Barnes maze [115] should also be consid- gyri/sulci in rodents) and rotational force component ered. In future studies, other behaviour tests that better common in large-brained humans TBIs are difficult to T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 17 of 20 be replicated in small-brained animals [121]. Tauopathy will provide a more complete clinical picture in human is a significant feature in repetitive closed head injury concussions. (for example, contact sport athletes and military per- sonnel) [122], but TBI-induced tauopathy in rodents Supplementary information have mainly been demonstrated in transgenic mice, Supplementary information accompanies this paper at https ://doi. org/10.1186/s4047 8-020-01098 -y. which were genetically modified to develop tauopathy [121]. Other studies in the CHIMERA model showed Additional file 1: Data 1. “Signal” resting-state Group Independent Vector tau phosphorylation to be a transient feature, which Analysis components. normalised by day 7 post-injury [77, 115]. It is notable Additional file 2: Figure S2. Group averaged resting-state MRI functional that in the CHIMERA model, two consecutive mild connectivity. (A–D) Average functional network connectivity (FNC) matri- ces among Independent Components (ICs) identified by IVA-GL of (A) impacts 23  h apart caused increased tau phosphoryla- sham (n = 14), (B) CON day 2 (n = 9), (C) CON day 7 (n = 10), and (D) CON tion lasting until day 2 post-injury while a single mod- day 14 (n = 10) cohorts. Colour scaled by z test statistics; non-black cells erate-severe impact only cause transient increased were defined as component-component connectivity deemed statistically significant. One sample t-tests, permutation-tested, and FDR-corrected tau-phosphorylation at 6  h post-injury. Assuming the (q-value < 0.05, two-tailed). (E–H) Regional Homogeneity analysis’s local brains can completely recover after an injury, our imag- intrinsic functional connectivity represented as averaged seven-voxels- ing study attempted to identify imaging biomarkers neighbourhood Kendall’s Coefficient of Concordance (KCC) maps of (E) sham, (F) CON day 2, (G) CON day 7, and (H) CON day 14 cohorts. for injury recovery and, more importantly, a window Additional file 3: Figure S3. Randomise reproducibility test of DTI and of vulnerability within which consecutive injuries may NODDI metrics. Voxel-by-voxel statistical analysis results of Diffusion Tensor cause disproportionate consequences, for example, Imaging (FA = Fractional Anisotropy, MD = Mean Diffusivity, Dp = Parallel long-term tauopathy. Chronic traumatic encephalopa- Diffusivity, Dr = Radial Diffusivity) and Neurite Orientation Dispersion and Density Imaging metrics (NDI = Neurite Density Index, ODI = Orientation thy has been associated with sub-populations at risk of Dispersion Index) and Tensor-based Morphometry with Jacobian Index repetitive traumatic brain injury, such as contact sport (JI) reproducibility test. Statistical map thresholded at P value < 0.05 (two- athletes and military personnel [123] while animal tailed), unpaired two sample t-test, implemented as permutation tested for the General Linear Model, corrected for multiple comparisons with mass- models of repetitive closed-head injury do not consist- based FSL’s Threshold-free Cluster enhancement ( TFCE). Statistical maps ently demonstrate only transient elevated phospho-tau were overlaid on the averaged and registered DTI and NODDI metrics [124]. Long-term elevated phosphor-tau, 6  months maps corresponding to the statistical maps (DTI and NODDI results) and structural template ( TBM results). Corresponding grey scale map for each post-injury, was shown in one study after as many as 42 averaged DTI and NODDI metrics maps were provided; units for Dp, Dr, and impacts [125]. More research is required at this stage to MD were in mm/s2. ACA = Anterior Cingulate Area, AM = Amygdala, AUD address the question of human pathology can be mod- = Auditory Area, cc = corpus callosum, ec = external capsule, HP = Hip- pocampus, HYP = Hypothalamus, ic = internal capsule, INS = Insula, MB elled in animals and the existence of a hypothetical win- = Midbrain, PAL = Palladium, S1 = Primary Somatosensory Cortex, RSN = dow of vulnerability. Retrosplenial Area, STR = Striatum, TH = Thalamus, VIS = Visual Area. Additional file 4: Figure S4. Deficit and recovery trajectories of behav- Conclusion ioural and Magnetic Resonance Imaging (MRI) markers post-concussion. Relative changes of (A) behavioural and (B) MRI metrics of concussed Our multi-modal assessment of a mouse model of con- cohorts relative to the sham cohorts (horizontal black dotted line through cussion showed varying trends in the temporal profile 0). Behavioural measurements (NSS and TI) scaled as t-statistics of the of different MRI markers. While concussion symptoms corresponding CON vs. sham comparison. MRI biomarkers (JI, FA, NDI, ODI, KCC, and stim-fMRI) scaled as whole-brain averaged t-statistics of the corre- and routine structural imaging found significant changes sponding CON vs. sham comparison; whole-brain t-statistics were used as up to day 7 post-concussion, the changes were normal- a biomarker proxy that incorporated both extents and degrees of change. ised by day 14. By contrast, advanced imaging using NSS = Neuro Severity Score, TI = Thigmotaxis Index, JI = Jacobian Index, FA = Fractional Anisotropy, NDI = Neurite Density Index, ODI = Orienta- DTI, NODDI, resting-state, and stimulus-evoked fMRI tion Dispersion Index, KCC = Kendall’s Coefficient of Concordance (resting- revealed ongoing changes that persisted at day 14 with state functional Magnetic Resonance Imaging Regional Homogeneity), and different onset, reflecting ongoing biological and molec - stim-fMRI (stimulus-evoked functional Magnetic Resonance Imaging). ular changes. To the best of the authors’ knowledge, this study represents the first study to perform multi-modal Acknowledgements advanced MRI and behavioural assessment to moni- This research was supported by Motor Accident Insurance Commission (MAIC), The Queensland Government, Australia (Grant Number: 2014000857) tor recovery after a concussion in a mouse model. The and The University of Queensland (Promoting Women’s fellowship: CRM findings in this study have important implication for 200221-005374). We acknowledge the support from the Queensland NMR translation to human clinical settings. First, recovery Network and the National Imaging Facility (a National Collaborative Research Infrastructure Strategy capability) for the operation of 9.4T MRI and utilisation and discharge by clinical assessment criteria may not of image processing computational resources at the Centre for Advanced indicate complete brain recovery and the brain may still Imaging, the University of Queensland. be vulnerable to disproportionate consequences from subsequent concussions. Second, it is likely that multi- modal assessments using different imaging methods To and Nasrallah acta neuropathol commun (2021) 9:2 Page 18 of 20 Authors’ contribution 16. Eierud C, Craddock RC, Fletcher S, Aulakh M, King-Casas B, Kuehl D XVT: Conceptualisation; Data curation; Formal analysis; Investigation; Meth- et al (2014) Neuroimaging after mild traumatic brain injury: review and odology; Project administration; Resources; Software; Validation; Visualisation; meta-analysis. NeuroImage Clin 4:283–294 Roles/Writing—original draft. FAN: Conceptualisation; Funding acquisition; 17. Wortman RC, Meconi A, Neale KJ, Brady RD, McDonald SJ, Chris- Investigation; Project administration; Supervision; Writing—review & editing. tie BR et al (2018) Diffusion MRI abnormalities in adolescent rats All authors read and approved the final manuscript. given repeated mild traumatic brain injury. Ann Clin Transl Neurol 5:1588–1598 Availability of data and materials 18. Haber M, Hutchinson EB, Sadeghi N, Cheng WH, Namjoshi D, Cripton Data from this study is available, without reservations, on request to the cor- P et al (2017) Defining an analytic framework to evaluate quantitative responding author. MRI markers of traumatic axonal injury: preliminary results in a mouse closed head injury model. eNeuro 4:ENEURO.0164-17.2017 Ethics approval and consent to participate 19. Tu T-W, Lescher JD, Williams RA, Jikaria N, Turtzo LC, Frank JA (2017) All experiments were approved by the Institutional Animal Ethics Committee Abnormal injury response in spontaneous mild ventriculomegaly at the University of Queensland (Animal Ethics Committee approval number Wistar rat brains: a pathological correlation study of diffusion tensor QBI/260/17). and magnetization transfer imaging in mild traumatic brain injury. J Neurotrauma 34:248–256 Competing interests 20. Rodriguez-Grande B, Obenaus A, Ichkova A, Aussudre J, Bessy T, Barse All authors have no competing financial interests to declare. E et al (2018) Gliovascular changes precede white matter damage and long-term disorders in juvenile mild closed head injury. Glia Received: 28 September 2020 Accepted: 1 December 2020 66:1663–1677 21. Gazdzinski LM, Mellerup M, Wang T, Adel SAA, Lerch JP, Sled JG et al (2020) White matter changes caused by mild traumatic brain injury in mice evaluated using neurite orientation dispersion and density imag- ing. J Neurotrauma 11:1–11 22. Hoogenboom WS, Rubin TG, Ye K, Cui MH, Branch KC, Liu J et al (2019) References Diffusion tensor imaging of the evolving response to mild traumatic 1. National Center for Injury Prevention and Control (2003) Report to brain injury in rats. J Exp Neurosci 13:1–22 Congress on mild traumatic brain injury in the United States: steps to 23. Kikinis Z, Muehlmann M, Pasternak O, Peled S, Kulkarni P, Ferris C et al prevent a serious public health problem. Centers for Disease Control (2017) Diffusion imaging of mild traumatic brain injury in the impact and Prevention, Atlanta, GA accelerated rodent model: a pilot study. Brain Inj 31:1376–1381 2. Williams RM, Puetz TW, Giza CC, Broglio SP (2015) Concussion recovery 24. Bouix S, Pasternak O, Rathi Y, Pelavin PE, Zafonte R, Shenton ME (2013) time among high school and collegiate athletes: a systematic review Increased gray matter diffusion anisotropy in patients with persistent and meta-analysis. Sport Med 45:893–903 post-concussive symptoms following mild traumatic brain injury. PLoS 3. Henry LC, Elbin RJ, Collins MW, Marchetti G, Kontos AP (2016) Examin- One 8:e66205 ing recovery trajectories following sport-related concussion using a 25. Budde MD, Janes L, Gold E, Turtzo LC, Frank JA (2011) The contribution multi-modal clinical assessment approach. J Neurosurg 78:232–241 of gliosis to diffusion tensor anisotropy and tractography following 4. Semple BD, Lee S, Sadjadi R, Fritz N, Carlson J, Griep C et al (2015) traumatic brain injury: validation in the rat using Fourier analysis of Repetitive concussions in adolescent athletes—Translating clinical and stained tissue sections. Brain 134:2248–2260 experimental research into perspectives on rehabilitation strategies. 26. Badea A, Kamnaksh A, Anderson RJ, Calabrese E, Long JB, Agoston Front Neurol 6:1–16 DV (2018) Repeated mild blast exposure in young adult rats results in 5. Stejskal EO, Tanner JE (1965) Spin diffusion measurements: spin echoes dynamic and persistent microstructural changes in the brain. NeuroIm- in the presence of a time-dependent field gradient. J Chem Phys age Clin 18:60–73 42:288–292 27. Braeckman K, Descamps B, Pieters L, Vral A, Caeyenberghs K, Vanhove C 6. Basser PJ (1995) Inferring microstructural features and the physiologi- (2019) Dynamic changes in hippocampal diffusion and kurtosis metrics cal state of tissues from diffusion weighted images. NMR Biomed following experimental mTBI correlate with glial reactivity. NeuroImage 8:333–344 Clin 21:101669 7. O’Donnell LJ, Westin C-F (2011) An introduction to diffusion tensor 28. Zhang H, Schneider T, Wheeler-Kingshott CA, Alexander DC (2012) image analysis. Neurosurg Clin N Am 22:185–96, viii NODDI: practical in vivo neurite orientation dispersion and density 8. Alexander AL, Lee JE, Lazar M, Field AS (2007) Diffusion tensor imaging imaging of the human brain. Neuroimage 61:1000–1016 of the brain. Neurotherapeutics 4:316–329 29. Tariq M, Schneider T, Alexander DC, Gandini Wheeler-Kingshott CA, 9. Miles L, Grossman RI, Johnson G, Babb JS, Diller L, Inglese M (2008) Zhang H (2016) Bingham-NODDI: mapping anisotropic orientation Short-term DTI predictors of cognitive dysfunction in mild traumatic dispersion of neurites using diffusion MRI. Neuroimage 133:207–223 brain injury. Brain Inj 22:115–122 30. Wu Y-C, Mustafi SM, Harezlak J, Kodiweera C, Flashman LA, McAllister 10. Aoki Y, Inokuchi R, Gunshin M, Yahagi N, Suwa H (2012) Diffusion tensor TW (2018) Hybrid diffusion imaging in mild traumatic brain injury. J imaging studies of mild traumatic brain injury: a meta-analysis. J Neurol Neurotrauma 35:2377–2390 Neurosurg Psychiatry 83:870–876 31. Churchill NW, Caverzasi E, Graham SJ, Hutchison MG, Schweizer TA 11. Brandstack N, Kurki T, Tenovuo O (2013) Quantitative diffusion-tensor (2019) White matter during concussion recovery: comparing diffusion tractography of long association tracts in patients with traumatic brain tensor imaging (DTI) and neurite orientation dispersion and density injury without associated findings at routine MR imaging. Radiology imaging (NODDI). Hum Brain Mapp 40:1908–1918 267:231–239 32. Churchill NW, Caverzasi E, Graham SJ, Hutchison MG, Schweizer TA (2017) 12. Yuh EL, Cooper SR, Mukherjee P, Yue JK, Lingsma HF, Gordon WA et al White matter microstructure in athletes with a history of concussion: (2014) Diffusion Tensor Imaging for outcome prediction in mild trau- comparing diffusion tensor imaging (DTI) and neurite orientation disper - matic brain Injury: a TRACK-TBI study. J Neurotrauma 31:1457–1477 sion and density imaging (NODDI). Hum Brain Mapp 38:4201–4211 13. Bazarian JJ, Zhong J, Blyth B, Zhu T, Kavcic V, Peterson D (2007) Diffusion 33. Biswal B, Yetkin FZ, Haughton VM, Hyde JS (1995) Functional connectiv- tensor imaging detects clinically important axonal damage after mild ity in the motor cortex of resting human brain using echo-planar MRI. traumatic brain injury: a pilot study. J Neurotrauma 24:1447–1459 Magn Reson Med 34:537–541 14. Mayer AR, Ling J, Mannell MV, Gasparovic C, Phillips JP, Doezema D et al 34. Zhou Y, Milham MP, Lui YW, Miles L, Reaume J, Sodickson DK et al (2012) (2010) A prospective diffusion tensor imaging study in mild traumatic Default-mode network disruption in mild traumatic brain injury. Radiol- brain injury. Neurology 74:643–650 ogy 265:882–892 15. Ling JM, Peña A, Yeo RA, Merideth FL, Klimaj S, Gasparovic C et al (2012) 35. Johnson B, Zhang K, Gay M, Horovitz S, Hallet M, Sebastianelli W Biomarkers of increased diffusion anisotropy in semi-acute mild trau- et al (2012) Alteration of brain default network in subacute phase of matic brain injury: a longitudinal perspective. Brain 135:1281–1292 T o and Nasrallah acta neuropathol commun (2021) 9:2 Page 19 of 20 injury in concussed individuals: resting-state fMRI study. Neuroimage 60. Lee JH, Lee TW, Jolesz FA, Yoo SS (2008) Independent vector analysis (IVA): 59:511–518 multivariate approach for fMRI group study. Neuroimage 40:86–109 36. Zhan J, Gao L, Zhou F, Kuang H, Zhao J, Wang S et al (2015) Decreased 61. Michael AM, Anderson M, Miller RL, Adalı T, Calhoun VD (2014) Preserv- regional homogeneity in patients with acute mild traumatic brain ing subject variability in group fMRI analysis: performance evaluation of injury: a resting-state fMRI Study. J Nerv Ment Dis 203:786–791 GICA vs. IVA. Front Syst Neurosci 8:1–18 37. Peters SK, Dunlop K, Downar J (2016) Cortico-striatal-thalamic loop 62. Laney J, Westlake K, Ma S, Woytowicz E, Calhoun VD, Adali T (2015) circuits of the salience network: a central pathway in psychiatric disease Capturing subject variability in fMRI data: a graph-theoretical analysis of and treatment. Front Syst Neurosci 10:1–23 GICA vs. IVAIVA. J Neurosci Methods 247:32–40 38. Kaushal M, España LY, Nencka AS, Wang Y, Nelson LD, McCrea MA et al 63. Du Y, Lin D, Yu Q, Sui J, Chen J, Rachakonda S et al (2017) Comparison (2019) Resting-state functional connectivity after concussion is associ- of IVA and GIG-ICA in brain functional network estimation using fMRI ated with clinical recovery. Hum Brain Mapp 40:1211–1220 data. Front Neurosci 11:1–18 39. Meier TB, Giraldo-Chica M, España LY, Mayer AR, Harezlak J, Nencka AS 64. Zang Y, Jiang T, Lu Y, He Y, Tian L (2004) Regional homogeneity et al (2020) Resting-state fMRI metrics in acute sport-related concussion approach to fMRI data analysis. Neuroimage 22:394–400 and their association with clinical recovery: a study from the NCAA- 65. Szafer A, Zhong J, Gore JC (1995) Theoretical model for water diffu- DOD CARE consortium. J Neurotrauma 37:152–162 sion in tissues. Magn Reson Med 33:697–712 40. To XV, Benetatos J, Soni N, Liu D, Abraha HM, Yan W et al (2020) Ultra- 66. Anderson MJ, Robinson J (2001) Permutation tests for linear models. high field diffusion tensor imaging identifies discrete patterns of con- Aust New Zeal J Stat 43:75–88 cussive injury in the rodent brain. J Neurotrauma 2020;neu.2019.6944 67. Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE (2014) Per- 41. Flierl MA, Stahel PF, Beauchamp KM, Morgan SJ, Smith WR, Shohami mutation inference for the general linear model. Neuroimage 92:381–397 E (2009) Mouse closed head injury model induced by a weight-drop 68. Dickie DA, Mikhael S, Job DE, Wardlaw JM, Laidlaw DH, Bastin ME device. Nat Protoc 4:1328–1337 (2015) Permutation and parametric tests for effect sizes in voxel- 42. Simon P, Dupuis R, Costentin J (1994) Thigmotaxis as an index of anxiety in based morphometry of gray matter volume in brain structural MRI. mice Influence of dopaminergic transmissions. Behav Brain Res 61:59–64 Magn Reson Imaging 33:1299–1305 43. Rorden C, Brett M (2000) Stereotaxic display of brain lesions. Behav 69. Smith SM, Nichols TE (2009) Threshold-free cluster enhancement: Neurol 12:191–200 addressing problems of smoothing, threshold dependence and 44. Bajic D, Craig MM, Mongerson CRL, Borsook D, Becerra L (2017) Identify- localisation in cluster inference. Neuroimage 44:83–98 ing rodent resting-state brain networks with independent component 70. Allen EA, Erhardt EB, Damaraju E, Gruner W, Segall JM, Silva RF et al analysis. Front Neurosci 11 (2011) A baseline for the multivariate comparison of resting-state 45. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ, networks. Front Syst Neurosci 5:1–23 Johansen-Berg H et al (2004) Advances in functional and structural MR 71. Gozzi A, Schwarz AJ (2016) Large-scale functional connectivity net- image analysis and implementation as FSL. Neuroimage 23:208–219 works in the rodent brain. Neuroimage 127:496–509 46. Ashburner J, Friston KJ (2003) Morphometry. In: Frackowiak RSJ, Friston 72. Hsu L-M, Liang X, Gu H, Brynildsen JK, Stark JA, Ash JA et al (2016) KJ, Frith C, Dolan R, Friston KJ, Price CJ et al (eds) Human brain function. Constituents and functional implications of the rat default mode Academic Press, Cambridge, pp 707–722 network. Proc Natl Acad Sci 113:E4541–E4547 47. Andersson JLR, Skare S, Ashburner J (2003) How to correct susceptibil- 73. Jankowski MM, Ronnqvist KC, Tsanov M, Vann SD, Wright NF, Erichsen ity distortions in spin-echo echo-planar images: application to diffusion JT et al (2013) The anterior thalamus provides a subcortical circuit sup- tensor imaging. Neuroimage 20:870–888 porting memory and spatial navigation. Front Syst Neurosci 7:1–12 48. Hong X, To XV, Teh I, Rui J, Chuang K (2015) Evaluation of EPI distortion 74. O’Mara SM, Aggleton JP (2019) Space and memory (far) beyond the correction methods for quantitative MRI of the brain at high magnetic hippocampus: many subcortical structures also support cognitive field. Magn Reson Imaging 33:1098–1105 mapping and mnemonic processing. Front Neural Circuits 13:1–12 49. Dohmatob E, Varoquaux G, Thirion B (2018) Inter-subject registration of func- 75. Bubb EJ, Kinnavane L, Aggleton JP (2017) Hippocampal–dience- tional images: do we need anatomical images? Front Neurosci 12:1–11 phalic–cingulate networks for memory and emotion: an anatomical 50. Du Y, Fan Y (2013) Group information guided ICA for fMRI data analysis. guide. Brain Neurosci Adv 1:1–20 Neuroimage 69:157–197 76. Grandjean J, Zerbi V, Balsters J, Wenderoth N, Rudina M (2017) The 51. Calhoun VD, Adali T, Pearlson GD, Pekar JJ (2001) A method for making structural basis of large-scale functional connectivity in the mouse. J group inferences from functional MRI data using independent compo- Neurosci 37:0438-17 nent analysis. Hum Brain Mapp 14:140–151 77. Namjoshi DR, Cheng WH, McInnes KA, Martens KM, Carr M, Wilkinson 52. Smith SM, Andersson J, Auerbach EJ, Beckmann CF, Bijsterbosch J, A et al (2014) Merging pathology with biomechanics using CHIMERA Douaud G et al (2013) Resting-state fMRI in the human connectome (closed-head impact model of engineered rotational acceleration): project. Neuroimage 80:144–168 a novel, surgery-free model of traumatic brain injury. Mol Neurode- 53. Du Y, Allen EA, He H, Sui J, Wu L, Calhoun VD (2016) Artifact removal in gener 9:55 the context of group ICA: a comparison of single-subject and group 78. Namjoshi DR, Cheng WH, Bashir A, Wilkinson A, Stukas S, Martens approaches. Hum Brain Mapp 37:1005–1025 KM et al (2017) Defining the biomechanical and biological threshold 54. Gri L, Douaud G, Bijsterbosch J, Evangelisti S, Alfaro-almagro F, Glasser of murine mild traumatic brain injury using CHIMERA (closed head MF et al (2017) Hand classification of fMRI ICA noise components. impact model of engineered rotational acceleration). Exp Neurol Neuroimage 154:188–205 292:80–91 55. Zerbi V, Grandjean J, Rudin M, Wenderoth N (2015) Mapping the mouse 79. Sauerbeck AD, Fanizzi C, Kim JH, Gangolli M, Bayly PV, Wellington CL, brain with rs-fMRI: an optimized pipeline for functional network identi- Brody DL, Kummer TT (2018) modCHIMERA: a novel murine closed- fication. Neuroimage 123:11–21 head model of moderate traumatic brain injury. Sci Rep 8(1):7677 56. Bell AJ, Sejnowski TJ (1995) An information-maximization approach 80. Hehar H, Yu K, Ma I, Mychasiuk R (2016) Paternal age and diet: the con- to blind separation and blind deconvolution. Neural Comput tributions of a father’s experience to susceptibility for post-concussion 1159:1129–1159 symptomology. Neuroscience 332:61–75 57. Lee T, Girolami M, Sejnowski TJ (1999) Independent component analysis 81. Salberg S, Yamakawa G, Christensen J, Kolb B, Mychasiuk R (2017) using an extended infomax algorithm for mixed subgaussian and Assessment of a nutritional supplement containing resveratrol, prebi- supergaussian sources. Neural Comput 11:417–441 otic fiber, and omega-3 fatty acids for the prevention and treatment of 58. Nasrallah FA, Tay H, Chuang K (2014) Detection of functional connectiv- mild traumatic brain injury in rats. Neuroscience 365:146–157 ity in the resting mouse brain. Neuroimage 86:417–424 82. Mychasiuk R, Hehar H, Candy S, Ma I, Esser MJ (2016) The direction of 59. Anderson M, Adali T, Li X-L (2012) Joint blind source separation swith the acceleration and rotational forces associated with mild traumatic multivariate gaussian model: algorithms and performance analysis. IEEE brain injury in rodents effect behavioural and molecular outcomes. J Trans Signal Process 60:1672–1683 Neurosci Methods 257:168–178 To and Nasrallah acta neuropathol commun (2021) 9:2 Page 20 of 20 83. Sherman M, Liu MM, Birnbaum S, Wolf SE, Minei JP, Gatson JW (2016) 105. Murphy K, Birn RM, Handwerker DA, Jones TB, Bandettini PA (2009) The Adult obese mice suffer from chronic secondary brain injury after mild impact of global signal regression on resting state correlations: are anti- TBI. J Neuroinflamm 13:1–10 correlated networks introduced? Neuroimage 44:893–905 84. Ertürk A, Mentz S, Stout EE, Hedehus M, Dominguez SL, Neumaier L et al 106. Chang C, Glover GH (2009) Eec ff ts of model-based physiological noise (2016) Interfering with the chronic immune response rescues chronic correction on default mode network anti-correlations and correlations. degeneration after traumatic brain injury. J Neurosci 36:9962–9975 Neuroimage 47:1448–1459 85. Govindarajan KA, Narayana PA, Hasan KM, Wilde EA, Levin HS, Hunter 107. Fransson P (2005) Spontaneous low-frequency BOLD signal fluctua- JV et al (2016) Cortical thickness in mild traumatic brain injury. J Neuro- tions: an fMRI investigation of the resting-state default mode of brain trauma 33:1809–1817 function hypothesis. Hum Brain Mapp 26:15–29 86. Wang X, Xie H, Cotton AS, Tamburrino MB, Brickman KR, Lewis TJ et al 108. Uddin LQ, Clare Kelly AM, Biswal BB, Xavier Castellanos F, Milham MP (2015) Early cortical thickness change after mild traumatic brain injury (2009) Functional connectivity of default mode network components: following motor vehicle collision. J Neurotrauma 32:455–463 correlation, anticorrelation, and causality. Hum Brain Mapp 30:625–637 87. Ross DE, Seabaugh JD, Seabaugh JM, Alvarez C, Ellis LP, Powell C et al 109. Gopinath K, Krishnamurthy V, Cabanban R, Crosson BA (2015) Hubs of (2020) Patients with chronic mild or moderate traumatic brain injury anticorrelation in high-resolution resting-state functional connectivity have abnormal brain enlargement. Brain Inj 34:11–19 network architecture. Brain Connect 5:267–275 88. Ross DE, Graham TJ, Ochs AL (2013) Review of the evidence supporting 110. Schwarz AJ, Gass N, Sartorius A, Risterucci C, Spedding M, Schenker E the medical and legal use of NeuroQuant in patients with traumatic et al (2013) Anti-correlated cortical networks of intrinsic connectivity in brain injury. Psychol Inj Law 6:75–80 the rat brain. Brain Connect 3:503–511 89. Ross DE, Seabaugh J, Cooper L, Seabaugh J (2018) NeuroQuant and 111. Akiki TJ, Averill CL, Abdallah CG (2017) A network-based neurobiological NeuroGage reveal effects of traumatic brain injury on brain volume. model of PTSD: evidence from structural and functional neuroimaging Brain Inj 32:1437–1441 studies. Curr Psychiatry Rep 19:81 90. Ding K, De La Plata CM, Wang JY, Mumphrey M, Moore C, Harper C et al 112. Raichle ME (2011) The restless brain. Brain Connect 1:3–12 (2008) Cerebral atrophy after traumatic white matter injury: correlation 113. Iraji A, Chen H, Wiseman N, Welch RD, O’Neil BJ, Haacke EM et al (2016) with acute neuroimaging and outcome. J Neurotrauma 25:1433–1440 Compensation through functional hyperconnectivity: a longitudinal 91. Scheid R, Preul C, Gruber O, Wiggins C, von Cramon DY (2003) Diffuse connectome assessment of mild traumatic brain injury. Neural Plast axonal injury associated with chronic traumatic brain injury: evidence 114. Verley DR, Torolira D, Pulido B, Gutman B, Bragin A, Mayer A et al (2018) from T2*-weighted gradient-echo imaging at 3 T. AJNR Am J Neurora- Remote changes in cortical excitability after experimental traumatic brain diol 24:1049–1056 injury and functional reorganization. J Neurotrauma 35:2448–2461 92. Qiu LR, Germann J, Spring S, Alm C, Vousden DA, Palmert MR et al 115. Bashir A, Abebe ZA, McInnes KA, Button EB, Tatarnikov I, Cheng WH et al (2013) Hippocampal volumes differ across the mouse estrous cycle, (2020) Increased severity of the CHIMERA model induces acute vascular can change within 24 hours, and associate with cognitive strategies. injury, sub-acute deficits in memory recall, and chronic white matter Neuroimage 83:593–598 gliosis. Exp Neurol 324:113116 93. Scholz J, Allemang-Grand R, Dazai J, Lerch JP (2015) Environmental 116. Malkesman O, Tucker LB, Ozl J, McCabe JT (2013) Traumatic brain injury— enrichment is associated with rapid volumetric brain changes in adult modeling neuropsychiatric symptoms in rodents. Front Neurol 4:157 mice. Neuroimage 109:190–198 117. Golub Y, Mauch CP, Dahlhoff M, Wotjak CT (2009) Consequences 94. Lerch JP, Yiu AP, Martinez-Canabal A, Pekar T, Bohbot VD, Frankland PW of extinction training on associative and non-associative fear in a et al (2011) Maze training in mice induces MRI-detectable brain shape mouse model of Posttraumatic Stress Disorder (PTSD). Behav Brain Res changes specific to the type of learning. Neuroimage 54:2086–2095 205(2):544–549 95. Li L, Chopp M, Ding G, Qu C, Nejad-Davarani SP, Davoodi-Bojd E et al 118. Jelescu IO, Budde MD (2017) Design and validation of diffusion MRI (2017) Diffusion-derived magnetic resonance imaging measures of models of white matter. Front Phys 5:61 longitudinal microstructural remodeling induced by marrow stromal 119. Namjoshi DR, Cheng WH, Carr M, Martens KM, Zareyan S, Wilkinson A cell therapy after traumatic brain injury. J Neurotrauma 34:182–191 et al (2016) Chronic exposure to androgenic-anabolic steroids exacer- 96. Burda JE, Bernstein AM, Sofroniew MV (2016) Astrocyte roles in trau- bates axonal injury and microgliosis in the CHIMERA mouse model of matic brain injury. Exp Neurol 275:305–315 repetitive concussion. PLoS ONE 11:1–21 97. Zhou Y, Shao A, Yao Y, Tu S, Deng Y, Zhang J (2020) Dual roles of astro- 120. Yi SY, Barnett BR, Torres-Velázquez M, Zhang Y, Hurley SA, Rowley PA cytes in plasticity and reconstruction after traumatic brain injury. Cell et al (2019) Detecting microglial density With quantitative multi-com- Commun Signal 18:1–16 partment diffusion MRI. Front Neurosci 13:1–9 98. Kantarci K, Murray ME, Schwarz CG, Reid RI, Przybelski SA, Lesnick T et al 121. Risling M, Smith D, Stein TD, Thelin EP, Zanier ER, Ankarcrona M et al (2017) White-matter integrity on DTI and the pathologic staging of (2019) Modelling human pathology of traumatic brain injury in animal Alzheimer’s disease. Neurobiol Aging 56:172–179 models. J Intern Med 285:594–607 99. Mierzwa AJ, Marion CM, Sullivan GM, McDaniel DP, Armstrong RC 122. Castellani RJ, Perry G, Tabaton M (2019) Tau biology, tauopathy, (2015) Components of myelin damage and repair in the progression of traumatic brain injury, and diagnostic challenges. J Alzheimer’s Dis white matter pathology after mild traumatic brain injury. J Neuropathol 67:447–467 Exp Neurol 74:218–232 123. Asken BM, Sullan MJ, DeKosky ST, Jaffee MS, Bauer RM (2017) Research 100. Wang N, Zhuang J, Wei H, Dibb R, Qi Y, Liu C (2019) Probing demyelina- gaps and controversies in chronic traumatic encephalopathy: a review. tion and remyelination of the cuprizone mouse model using multimo- JAMA Neurol 74:1255–1262 dality MRI. J Magn Reson Imaging 50:1852–1865 124. Hoogenboom WS, Branch CA, Lipton ML (2019) Animal models of 101. Blumenfeld-Katzir T, Pasternak O, Dagan M, Assaf Y (2011) Diffusion MRI closed-skull, repetitive mild traumatic brain injury. Pharmacol Ther of structural brain plasticity induced by a learning and memory task. 198:109–122 PLoS One 6:e20678 125. Petraglia A, Plog B, Dayawansa S, Dashnaw M, Czerniecka K, Walker C 102. Sours C, Zhuo J, Janowich J, Aarabi B, Shanmuganathan K, Gullapalli et al (2014) The pathophysiology underlying repetitive mild traumatic RP (2013) Default mode network interference in mild traumatic brain brain injury in a novel mouse model of chronic traumatic encephalopa- injury—A pilot resting state study. Brain Res 1537:201–215 thy. Surg Neurol Int 5:184 103. McCuddy WT, España LY, Nelson LD, Birn RM, Mayer AR, Meier TB (2018) Association of acute depressive symptoms and functional connectivity Publisher’s Note of emotional processing regions following sport-related concussion. Springer Nature remains neutral with regard to jurisdictional claims in pub- NeuroImage Clin 19:434–442 lished maps and institutional affiliations. 104. Meier TB, Bellgowan PSF, Mayer AR (2017) Longitudinal assessment of local and global functional connectivity following sports-related concussion. Brain Imaging Behav 11:129–140

Journal

Acta Neuropathologica CommunicationsSpringer Journals

Published: Jan 6, 2021

There are no references for this article.