A more conservative approach to neuroimaging for clinical diagnosis and management of pediatric mTBI, as compared to adults, is often recommended.²² However, prior to the publication of the CDC Pediatric mTBI Guideline in 2018,²³ no evidence-based guidelines on the diagnosis and management of pediatric patients with mTBI were available that were specific to the United States, relevant to both sport- and non-sport-related injuries, and applicable to younger as well as older age groups. In the area of diagnosis, CDC authors sought to answer a specific question regarding the usage of neuroimaging: “For children (18 years of age and younger) presenting to the emergency department (or other acute care setting) with mTBI, how often does routine head imaging identify important intracranial injury?”²⁴ Based on the confidence levels found across the thirty imaging-modality articles were ultimately included for quantitative synthesis from data extraction based on the inclusion criteria, the CDC mTBI guideline workgroup concluded that healthcare providers should not routinely image a pediatric patient with suspected mTBI for diagnostic purposes.²⁵ Similarly, based on the limited diagnostic and prognostic evidence, presumed low base-rates of positive findings, and high cost, neither routine nor advanced CT or MRI for diagnosis of mTBI and concussion are currently endorsed by the American Academy of Neurology²⁶ and the American Medical Society for Sports Medicine.²⁷ Instead of routine head imaging, the CDC guideline and other guidelines state that healthcare providers should use validated clinical decision rules to identify children at risk for intracranial injury and to determine if imaging is warranted.²³

To this effect, several validated clinical decision rules are available to healthcare providers: the Pediatric Emergency Care Applied Research Network (PECARN)¹⁷ rule; the Canadian Assessment of Tomography for Childhood Head Injury (CATCH) rule;²⁷ and the Children’s Head Injury Algorithm for the Prediction of Important Clinical Events (CHALICE).²⁸ These decision rules evaluate a variety of clinical factors that when assessed together are predictive of more serious injury and should prompt head imaging.²⁹ These factors include: age < 2 years old; vomiting; loss of consciousness; severe mechanism of injury; severe or worsening headache; amnesia; non-frontal scalp hematoma; Glasgow Coma Score (GCS) < 15; and clinical suspicion for skull fracture. A multicenter validation study evaluating 20,137 children seen in an emergency department for head injury found that the PECARN, CATCH, and CHALICE decision rules accurately identified children with clinically significant head injuries.²⁹ Additional studies have also concluded that decision rules that combine the risk factors described above are more effective than CT scans alone in identifying children at low risk for intracranial injury.^{17,27,30–32} When there is concern for abusive head trauma, imaging to determine the likelihood of abuse may be warranted to identify clinically insignificant but forensically important injuries.³³ The Pittsburg Infant Brain Injury Score (PIBIS) is such a clinical prediction rule being evaluated for its ability to aid physicians deciding which high risk cases should undergo head CT.³³

Beyond the use of stereotypical imaging methodologies, the field of brain injury research has seen novel advanced imaging modalities, which may prove useful for pediatric mTBI. However, consideration as to their application, safety, and evidence-based usage must be fully discussed. One factor increasing interest in additional or supplemental diagnostic tools for pediatric mTBI is the lack of definitive indicators as to when children can safely return to sports and school. This may engender uncertainty among healthcare providers managing a pediatric patient with an mTBI. If allowed to return to sports too soon, a child is at increased risk for repeat injury, an exacerbation of current symptoms, and delayed recovery.^34–36 Conversely, children restricted from school and social activities for longer than is physiologically necessary can experience adverse health outcomes.³⁷ For children who experience a prolonged recovery (especially in those with a recovery time > 1 year), little high-quality research (e.g., randomized control trials) is available to guide return to activities.^38–39 Thus, interest is growing in quantifying the physiology of mTBI and identifying specific technologies that can support optimal outcomes and management strategies for healthcare providers caring for children with mTBI.⁴⁰

The use of non-invasive imaging biomarkers to inform prognosis of patients with mTBI at the acute and chronic time points after injury is a focus of increasing research on mTBI. Neuroimaging has been proposed as a possible methodology to discover such markers for identifying brain changes related to pediatric mTBI that may be predictive of recovery time-course.⁴¹ However, the development of imaging-related diagnostics is complex, especially in younger children, and currently no standardized biomarkers are available that healthcare providers can use to diagnose mTBI or predict recovery.^40–42 Current guidelines state that, while imaging biomarkers show promise for informing the pathophysiology of mTBI and neurobiological recovery, they require further investigation and should not be used outside of the research setting.^23,26,43

The development of biomarkers is particularly complex as the immediate and longitudinal effects of pediatric mTBI are superimposed on a rapidly changing brain. Neurodevelopmental trajectories vary as a function of age, sex, and developmental factors;⁴⁴ thus, findings relevant for one group of children may not directly translate to another despite outward similarities. Significantly, cortical thickness, subcortical volumes, and functional connectivity vary with age,⁴⁵ with significant differences in grey matter organization and cerebral blood flow.⁴⁶ Neurodevelopmental trajectory in pediatric populations adds an additional level of variability to biological assays relative to more homogeneous adult populations. This not only highlights the need for longitudinal designs and the collection of large normative data populations, but the need to factor in age and other factors. For example, some young children sometimes may be unable to follow instructions and remain stationary during image acquisition periods despite preparation, making quality images more difficult to obtain.⁴⁷ Nonetheless, many researchers are pursuing the discovery and development of non-invasive biomarkers of injury using imaging modalities, particularly to suit and accommodate the unique physiology and vulnerabilities of the developing brain.⁴⁸

fMRI serves to indirectly measure neuronal activity through blood oxygen level dependent (BOLD) signal as a measure of brain function. This modality can be used not only to study how different regions of the brain are activated or deactivated in response to specific tasks compared with a baseline, but also to study intrinsic or resting-state synchronous spontaneous brain activity.^53,54 It has been proposed that increased regional activation may reflect the recruitment of additional compensatory brain systems (e.g., those required to accomplish tasks in a compromised neural system) or may be due to injury-induced brain reorganization.⁵⁵ In comparison, decreased levels of activation may be related to several processes, such as impaired neural functioning or difficulty in allocating appropriate cognitive and attention-related resources to the task.^40,56 Task-based fMRI (i.e., where a subject is asked to perform a task while in the MRI scanner) has been applied in a range of publications to investigate both acute and chronic changes in brain activity in pediatric mTBI, and several groups have correlated deficits in neurovascular coupling with degree of persistent symptoms.⁴⁰ Increased activation in the cerebellum has been found to correlate with symptomatology during an inhibitory control component of a working memory task among pediatric mTBI patients evaluated around 1 month post-injury.⁵⁷ In a chronic population (+1 year post-injury), greater activation within working memory circuitry and expanded spatial extent of activation in mTBI patients compared to controls during a working memory task has also been shown.⁵⁸ To this effect, another study reported decreased activation in the dorsolateral prefrontal cortex, premotor, supplementary motor areas, and left superior parietal lobule during a verbal and non-verbal working memory task in children 9 to 90 days post-injury.⁵⁷ Additionally, decreases in activation were found in various areas (e.g., cerebellum, basal ganglia, and thalamus) during an auditory orienting task sub-acutely (<3 weeks post-injury).^59,60 Others have yet shown bidirectional functional changes, utilizing working memory and navigational tasks, finding increased and decreased activity levels in different areas—though during a wide range of time post-injury (0–3 and 3–6 months).^58,59 Interestingly, children with mTBI did not show significant deficits on traditional neuropsychological “paper and pencil tasks,” but showed greater impairment on symptom report measures and “real world” measures of executive functioning.^57,55 Thus, while variability in study design, post-injury time-points, and age groupings limit conclusions as to the standardized clinical applicability of task-based fMRI for pediatric mTBI, findings show positive support for the presence of detectable, symptom-correlated, and potentially diagnostic changes in brain function.

Resting-state fMRI also is based on the BOLD signal but, unlike task-based fMRI, it comparatively measures innate brain connectivity. Findings from resting-state fMRI studies in children with persistent symptoms after mTBI are not as ranged as those of task-based, but do suggest altered functional connectivity in the default mode, executive function, and ventral attention networks.⁶⁰ Furthermore, similar to task-based findings, alterations in brain dynamics and connectivity within functional networks have been posited to be related to neurocognitive dysfunction and posttraumatic symptoms.⁶¹ Briefly, one preliminary report utilizing resting-state fMRI in adolescent athletes showed alterations within the default mode network, increased connectivity in the right frontal pole in the executive function network, and increased connectivity in the left frontal operculum cortex associated with the ventral attention network in the sub-acute phase of mTBI (~35 days post-injury).⁶³ A comprehensive study done by Iyer et al. recruited a large sample of children diagnosed with persistent symptoms after mTBI to study the relationship between resting functional brain connectivity, symptomology, and behavior.⁶⁴ They found that individual variations in resting-state functional connectivity in the mTBI cohort were associated with various symptoms and behavior along a single negative to positive dimension (e.g., decreases in certain brain networks, problems with cognition, and emotion loaded negatively on the dimension, while high connectivity in other brain networks, poor sleep, and fatigue loaded positively). This data suggests the link between brain connectivity and persistent symptoms might provide a basis for improved prognosis and movement towards personalized therapeutic interventions. Nonetheless, 1) methodologically, supplementation of fMRI studies (such as blood flow and vascular reactivity variability) with may be beneficial given that the BOLD signal represents a multifaceted measure of NVC and additional indices may be necessary to accommodate for hemodynamics/perfusion⁴⁰ and 2) further studies are needed to assess dynamic connectivity, regional homogeneity, and global connectivity changes in pediatric mTBI.

DTI is a technique that measures diffusion of water molecules in order to explore the microarchitecture of the brain. DTI is a subset of diffusion-weighted imaging that is often used to map white matter (WM) tracts (tractography) in the brain.⁶⁵ WM analysis techniques have proven invaluable in noninvasively examining maturational changes during normal development, as well as in children with acquired injury.⁶⁵ Consequently, DTI in pediatric mTBI has been repeatedly examined for its applicability and biomarker potential in milder insults, especially as WM tracts are likely disturbed following brain injury.^50,66 Certain evidence even suggests that subtle abnormalities following brain trauma are better captured by observing WM metrics relative to conventional MRI sequences.^62,63 However, even recent data from studies investigating the modality in pediatric mTBI show puzzling discrepancies between studies and outcomes, particularly with regard to time post-injury.^66,68

In the semi-acute time period post-injury, reports of increased fractional anisotropy (FA; a measure of diffusion restriction) have been observed in independent samples of pediatric mTBI patients.^69–70 A study of acute (within 96 hours post-injury) pediatric mTBI showed brain injury was associated with significantly higher levels of FA and axial diffusivity (AD; a measure of water diffusion along the principal axis of diffusion) in several WM regions including the middle temporal gyrus, superior temporal gyrus, anterior corona radiata, and superior longitudinal fasciculus.⁷¹ The mTBI group also had significantly lower levels of mean diffusivity (MD; a measure of the total diffusion within a voxel) and/or radial diffusivity (RD; a measure of diffusion perpendicular to the principal axis) in a few WM regions including the middle frontal gyrus WM and anterior corona radiata. However, these diffusion alterations correlated poorly with acute symptom burden.⁷¹ Some publications further completed post-acute (within 21 days post-injury) scans and/or correlated post-concussion symptoms, including one in which pediatric mTBI patients exhibited increased FA in the left temporal cortex and right thalamus relative to controls during the semi-acute injury phase, with the FA abnormalities associated with decreased performance on attentional measures.⁷⁰ Interestingly, in this study FA remained increased within the left temporal cortex, with a trend seen for the right thalamus at approximately 4 months post-injury despite cognitive assessments partially normalizing.⁷⁰ A study focusing on the cingulum bundles and memory functioning in acute (~3 days post-injury) pediatric mTBI found FA of the left bundle was significantly correlated with 30-min delayed recall in injured group when given an episodic verbal learning and memory task.⁷²

Interestingly, a number of DTI studies in chronic pediatric mTBI from the past decade reported decreased FA in the WM of chronic patients, mostly in the corpus callosum.⁷³ However, other groups have not found this to be the case, as some results indicate there were only group differences in two of the measures analyzed post hoc, MD and RD, with some limited results in AD.⁷³ And others showing increased whole brain FA and decreased MD within 2 months post-injury.⁷⁴ Recently, a study was conducted that utilized a prospective, longitudinal, and controlled cohort design to evaluate WM microstructure and persistent post-concussive symptoms in children following mTBI.⁷⁵ This study imaged subjects at one-month post-injury and 4–6 weeks later. They found FA of the left uncinate fasciculi was lower in symptomatic patients as compared to non-mTBI controls.⁷⁵ Regional FA and MD was associated with symptomology at both time points. However, no other significant differences were observed.⁷⁵ In contrast to this paper and prior data, a 2019 publication by Satchell et al. found no significant differences between age- and gender-matched symptomatic pediatric mTBI athletes with clinical controls at an average of 30 days post-injury.⁶⁶ Thus, despite trends in pediatric mTBI research, DTI has not been adequately shown to be a consistent, reliable measure for changes in pediatric mTBI, and therefore should not be used as a clinical diagnostic tool in individual patients. There is great interest in the potential source(s) of the variability in DTI findings before the methodology’s applicability in pediatric mTBI biomarker development can be further assessed.^76,77

Two techniques for estimating brain hemodynamics in pediatric mTBI are SPECT, a nuclear medicine technique that requires the injection of a radio-nuclide trace, and ASL, a non-invasive way of estimating cerebral blood flow (CBF) using MRI. SPECT may not be as promising for a pediatric population, as it involves radiation. However, studies using SPECT have shown reduced CBF in children with concussion within 12 hours of the head injury.⁷⁸ In contrast, ASL does not rely on an external contrast agent to measure perfusion, which increases its utility in the clinical setting and pediatric populations.⁷⁹ In adult mTBI, alterations in CBF have been found, but the recovery progression does not appear to match what is observed in severe TBI.⁸⁰ Children with mTBI, however, exhibit cerebrovascular reactivity impairments similar to moderate or severe TBI, suggesting age-related CBF biomarkers may be discoverable for determining initial risk and during recovery.^80,81 Conducting pseudo-continuous ASL at 40 days post-injury, Barlow et al. demonstrated that cerebral perfusion was significantly higher in pediatric TBI patients with post-concussion symptoms than controls, and lower in the asymptomatic TBI patients.⁷⁹ They further postulated that children who were thought to have clinically “recovered” may still have ongoing decreases in cerebral perfusion and thus may not be fully “neurologically recovered.”⁸¹ In 2018, Stephens et al. sought to provide more consistency in the field and utilized ASL to study cerebrovascular physiology after sports-related concussion/mTBI with regard to symptomology.⁷⁹ The resulting data showed teenage athletes 2 weeks after concussion had significantly higher regional cerebral blood flow (rCBF) in the left insula and left dorsal ACC; increases in the left dorsal ACC persisted at 6 weeks post-injury. In addition, perfusion in the left dorsal ACC was higher in athletes reporting physical symptoms 6 weeks post-injury compared with asymptomatic athletes.⁷⁹ Overall, these data seem to suggest pediatric mTBI symptomology is related to higher global CBF compared to controls, suggesting CBF perfusion may be a marker of physiological status after concussion. While there still exists variability in the literature in both brain hemodynamics study design and findings, which is compounded by variability in the current mTBI literature, the promise of longitudinal and age-matched studies alongside relating CBF indices to symptoms is a promising indicator of the potential use of brain hemodynamics as biomarkers for pediatric mTBI.

While both PET and MRSI have been used to assess metabolic changes following severe TBI in children, their ability to evaluate metabolic shifts that occur during the neurometabolic cascade of concussion does differ.^82,83 PET examines a wide variety of underlying neural pathophysiologies, such as changes in glucose metabolism or neurotransmitters, but necessitates exposure to radioactive tracers—rendering PET a less desirable modality for children and adolescents.^83,84 In contrast, MRSI measures brain metabolite concentrations reflective of components of the neurovascular unit, including neuronal and glial metabolism even across age groups, without such exposure.⁸⁵ However, pediatric MRSI studies are limited and have shown variable results. Maugans et al. 2012 performed combined neurocognitive and neuroimaging on pediatric mTBI patients <72 hours, 14 days, and 30 days+ post-injury.^{86, 84} They found no longitudinal metabolic changes in thalamus, frontal grey or white matter, nucleus accumbens, or lactate in mTBI, and no differences compared to controls.⁸⁶ Another followed high-school football athletes longitudinally and found a significant metabolic change in the thalamus at both subacute (2 weeks post-injury) and chronic (~1 year post-injury) time points after injury, as well as differences in frontal WM metabolites at the chronic time point.^84,87 However, a following study using MRSI demonstrated metabolic changes in a mTBI group at 3 months post-injury, long after clinical scores had returned to normal and the athletes had been cleared to return to play.⁸⁸ More recently, one group has shown changes in specific frontal lobe metabolites (at ~30 days post-injury), as well as that mTBI patients lacked a correlation between frontal lobe metabolites and brain activation during a working memory fMRI task that was present in controls.⁸⁹ While intriguing, research to date highlights the need to control for temporal and case- specific variables (i.e. time since injury and number of injuries) in future research, as well as the need to fully highlight the potential of metabolic biomarkers for detection of mTBI and identification of the biological corelates of persistent symptoms.