The effect of pediatric traumatic brain injury on behavioral outcomes: a systematic review
Dr Jianghong Liu at School of Nursing, University of Pennsylvania, 418 Curie Boulevard, Room 426, Philadelphia, PA 19104-4217, USA. E-mail: firstname.lastname@example.org
Aim To review systematically the empirical evidence on traumatic brain injury (TBI) during childhood and subsequent behavioral problems.
Method An initial literature search with keywords ‘brain injury,’‘children,’ and ‘behavior’ was conducted using Web of Knowledge and PubMed databases. Ancestry was also used. Original research studies published between 1990 and February 2012 focusing on behavioral outcomes of children sustaining TBI from ages 0 to 18 years were included.
Results Fifty studies, varying considerably in methodologies, were included in the review. Findings showed that up to 50% of brain-injured children are at risk for presenting with specific behavioral problems and disorders. These problems may emerge shortly or several years after injury and often persist and even worsen with time. These behavioral impairments appear to be moderated by the family environment.
Interpretation Survivors of childhood TBI are at risk for developing and sustaining behavioral impairments. Stronger research is needed to identify cognitive and environmental factors that contribute to the onset and maintenance of these problems. Healthcare providers should ensure adequate follow-up and assessment of a child’s behavioral, social, and neurocognitive domains. Caregivers should be encouraged to provide positive environments and parenting styles, which may help reduce chronic behavioral problems after brain injury.
Traumatic brain injury
What this paper adds
- • Pediatric patients with TBI are at significant increased risk for various behavioral and novel psychiatric problems, with prevalence rates of 10 to 50%.
- • The paper describes the evidence on possible behavioral problems and contributing factors.
- • It provides suggestions for future research and highlights the need for continued follow-up of childhood patients with TBI.
An average of 634 000 incidents of traumatic brain injury (TBI) occurs among children each year in the United States, with the highest TBI-related emergency room visits occurring in children under the age of 4 years and adolescents 15 years or older.1 Pediatric TBI in young children (<4y) is commonly caused by falls, which may account for approximately 40% of head injuries requiring hospitalization for children less than 4 years old.1 Inflicted injuries acquired secondary to child abuse (e.g. shaken baby syndrome) may account for 20 to 70% of hospitalizations in young children.2–4 In older children, motor-vehicle accidents are the more prevalent causes.1
Although focal brain lesions sustained during childhood generally show favorable recovery,5 the diffuse brain insult that frequently occurs in some combination with focal insults in TBI appears more threatening to the developing brain.6,7 Despite the prevalence and threat of childhood TBI, however, the impact of such damage on the brain remains poorly understood, especially within pediatric populations. Assessing sequelae from brain injury of varying etiologies on the developing brain remains a challenge for researchers. Even within the literature, the physical and cognitive consequences of pediatric TBI have historically received greater attention than behavioral and psychosocial problems. This is surprising as behavioral and psychosocial problems are often more concerning for parents and teachers; they may also more negatively influence patients’ lives than intellectual or physical problems.8,9 Early work by Rutter et al.10 and Brown et al.11 showed that among children with severe head injury, rates of new psychiatric disorders were three times more common than in orthopedic injury comparisons. Left untreated, such adverse outcomes of childhood brain injury may persist into adulthood, predisposing individuals to violent crime12 or mental health disorders.13
It is commonly believed that children with behavioral problems are at greater risk for experiencing a TBI event, owing to their inability to predict the consequences of their actions and resulting tendency to be more accident prone.14 However, this belief is largely based on research using post-injury recall of pre-injury behavior.15 Prospective studies using pre-injury data collected before the TBI event, therefore avoiding potential recall bias, have demonstrated that pre-injury behavior does not actually serve as a significant predictor for a TBI event.14,16,17 It is thus critical that patients, families, and healthcare workers do not assume that post-injury behavioral deficits simply reflect an individual’s pre-injury characteristics or developmental trajectory.18 Rather, behavioral problems are a real and potential consequence of TBI that have detrimental effects on children’s long-term functioning and quality of life.
In light of the growing recognition that childhood brain injury may contribute to the development of adverse, lifelong behavioral outcomes, it is critical to review systematically the recent empirical evidence and provide a comprehensive overview of the ongoing progress in this area. Despite the emerging research over the past two decades on the presentation, risk factors for, and correlates of behavioral outcomes after pediatric brain injury, few have taken a systematic approach to evaluate the effects of TBI on behavioral outcomes. In the present review, we therefore examine the relationship between childhood TBI and behavioral problems, as well as possible contributing factors, to help inform healthcare professionals, patients, families, and future research efforts.
Definition of terms
For this review, brain injury was limited to incidents of TBI and excluded injuries such as localized lesions, focal stroke, meningococcal meningitis, cerebral palsy, and intraventricular hemorrhage.
Behaviors were primarily categorized as ‘externalizing’ and ‘internalizing’ problems, a well-known distinction in the field of child psychology and psychiatry.19 The construct of externalizing behavior refers to problems manifested in children’s outward behavior and reflects the child negatively acting on the external environment (e.g. aggression, hyperactivity). Internalizing disorders are directed inward and indicate a child’s psychological and emotional state (e.g. anxiety, depression).20–22 Both symptoms and disorders fitting within these constructs were included as behavioral outcomes. Personality change, a disorder of behavioral change resulting from brain damage, was also included.
Literature search strategy
The initial literature search was conducted using keywords ‘brain injury,’‘behavior,’ and ‘children.’ Searches were also conducted using terms for specific behavioral problems, such as ‘aggression,’‘conduct disorder,’‘ADHD,’‘depression,’‘anxiety,’‘internalizing,’ and ‘externalizing.’ The year 1990 was selected as a cut-off owing to the increasing interest in behavioral and psychosocial outcomes after this time and the greater likelihood that criteria, measures, and outcomes used in included studies would be similar to those currently used. The last search was conducted in February 2012. Additional limits used were human participants, original article, publication in English, and all children (0–18y). This search yielded 662 and 812 original articles in PubMed and Web of Knowledge databases respectively. After an initial screening using title and abstracts to exclude non-relevant and duplicate studies, potentially relevant articles underwent a full text assessment for eligibility. Those meeting eligibility criteria were then assessed for quality. Reference lists were also searched.
Studies were screened for eligibility using the following inclusion criteria: (1) English language; (2) original research published in peer-reviewed journals; (3) study populations of human participants who sustained injury between the ages of 0 to 18 years; (4) identification and confirmation of TBI status through objective sources (e.g. medical records); and (5) behavior measured as a main outcome, as the principal outcomes in this review are prevalence and correlates of behavioral problems occurring after pediatric TBI.
Studies were excluded if they included non-TBI cases (e.g. focal stroke, meningitis) in the brain-injured sample. Furthermore, because this review focuses on manifestations of behavioral problems, papers reporting on neurocognitive deficits or social skills that may have been evaluated by behavioral ratings (e.g. executive functioning) were not selected for inclusion unless they simultaneously examined the relation between such deficits and a main behavioral outcome (e.g. depression).
Quality assessment and data abstraction
Information from the included studies was used to assess the risk of bias and methodological quality. Data extracted by authors and/or research assistants independently included identification of sample characteristics, study design and methodology, instruments, potential cohort overlap with other studies, main findings, and funding sources. Studies were rated using a scale modified from Cappa et al.23 and following recommendations by Satz et al.18 Scores (yes=1; no=0) were given for criteria reflecting a study’s design, sample, and measures. Studies with acceptable methodology and low risk of bias were expected to include a prospective longitudinal design with pre-injury (measured an average of ≤1mo after injury) and late (≥12mo) behavioral outcomes; matched uninjured and/or orthopedic comparisons; samples obtained from a representative population (e.g. consecutive admissions to several large hospitals and not recruited from rehabilitation or treatment centers); groups used in analysis to predict outcome comprising at least 20 participants per variable (e.g. severity group); and clear descriptions of sample characteristics and selection criteria, as well as descriptions of participants who chose not to participate or were lost to follow-up and/or comparisons with the study sample. Assessment of children’s behavior was also expected to use at least two informants (e.g. child, parent, teacher) and age-appropriate, standardized measures for which normative data existed and/or reliability and validity analyses had been conducted. Because this review also aims to investigate predictors of outcomes, analyses for confounders and a clear attempt to adjust for potential covariates (e.g. socio-economic status, injury severity, pre-injury functioning, sex, age, IQ) were also considered. A maximum score of nine was possible; studies with a score of three or less were excluded from further review. If information required for assessment was unavailable from the manuscript, authors were contacted for clarification.
Characterization of the Literature
Ninety-two full-text articles were screened and assessed for eligibility, 54 of which were further assessed for quality. The final sample used in this review includes 50 individual studies, but several of these belonged to the same or overlapping cohorts. The authors identified 19 samples from different settings. Information including study design, cohort setting, population demographics, measures, and methodological ratings for the included studies are provided in Table SI (online supplementary information). All studies that categorized severity used the Glasgow Coma Scale24 as a primary severity index. The Glasgow Coma Scale is composed of three parameters (verbal response, motor response, and eye opening) and generally includes the following levels: mild (13–15), moderate (9–12), and severe (≤8). Most studies also used imaging data when available to confirm severity and assess lesion characteristics; other indicators commonly used to describe the injury included duration of loss of consciousness, posttraumatic amnesia, and days hospitalized. One cohort used by several studies25–32 consisted of participants recruited through rehabilitation services. Two other studies also enrolled patients through referrals from the rehabilitation center33 or a treatment study.34 Although children sustaining inflicted head injuries secondary to abuse likely experience other environmental adversities that can contribute to poor behavioral outcomes, studies with cases of inflicted TBI35–37 were included owing to their prevalence in very young children and because they compared outcomes between inflicted and accidental cases when possible.
Most studies (n=31) used purely prospective, longitudinal designs that followed children anywhere from 6 months to several years after injury. Studies that assessed pre-injury behavior did so retrospectively with parental recall, often within the first 3 months after injury. Behavioral outcomes of children sustaining TBI were often compared with those of reference groups, which comprised either healthy, uninjured children (n=15) or children with orthopedic injuries (n=13) not involving the head or central nervous system. Reference groups were matched for age, sex, and sometimes sociodemographic characteristics. Whereas comparison with uninjured reference groups helps control for the normal developmental trajectory of children’s behavior, orthopedic injury groups account for differences between children who are and are not prone to injury, as well as the hospitalization and trauma experience. Only one study used both hospitalized and healthy comparisons.38 Studies without reference groups relied on population norms for measures or examined within-TBI variations (e.g. impact of age or severity, lesion characteristics). A range of instruments were used to assess behavioral outcomes; global measures of psychopathology, such the Child Behavior Checklist,39 were frequently used. Parents or primary caregivers were the most common informants.
Effect of Brain Injury on Behavioral Outcomes
Overall, the literature strongly suggests that after brain injury, children are at increased risk of adverse behavioral outcomes. Studies reporting few to no social or behavioral impairments after TBI25,36,37 assessed young children whose impairments may have lain dormant for several years or that were not yet identifiable using parent or teacher behavioral ratings.37,40 Examinations of behavioral outcomes in older children have frequently reported the emergence of general internalizing and externalizing problems after injury,41–44 with reported increases in behaviors such as aggression,26,27,45 impulsivity and hyperactivity,46,47 withdrawal,37,48 anxiety,49,50 and depression.50,51 Moreover, pediatric TBI has been linked to diagnoses of new-onset psychiatric disorders including personality change,52–56 conduct29 and oppositional defiant disorders,57 attention-deficit–hyperactivity disorder (ADHD),33,58–62 mood disorders,63 anxiety disorders including posttraumatic stress disorder and symptoms,28,63–65 and depressive disorders66 shortly after injury. Behavioral problems commonly co-emerge, such as novel depression and anxiety64,66 or new-onset ADHD and oppositional defiant, conduct, or disruptive behavior disorder.29,38,60,61 Although behavioral problems may resolve within the first year after injury,67 behavioral impairments that do persist until then often stabilize and even worsen in subsequent years.8,40,43,44,68
Attention problems are among the most commonly reported disorders after brain injury.67 Roughly 30 to 50% of children are reported to develop symptoms of ADHD soon after TBI.33,59,62,67 With approximately 20% of children hospitalized for TBI meeting criteria for pre-injury ADHD, attention problems are often present before TBI and may be predictive for sustaining head and other injuries.58,59,67 However, secondary ADHD, which can be diagnosed only in the absence of pre-existing ADHD, is also common after brain injury. Diagnoses of secondary ADHD are three times more common in children sustaining head injuries than in their uninjured peers.62 Pediatric patients with TBI who develop secondary ADHD appear to have deficits in inhibitory control that are similar to uninjured children with ADHD, suggesting similar underlying cognitive mechanisms.33 Although some studies have not reported ADHD symptomatology in younger children with TBI,25,37 attention problems may more noticeably emerge in middle to late childhood.38 Interestingly, lesions in the orbitofrontal cortex are significantly associated with secondary ADHD emergence in patients with TBI 6 months after injury, but not onset within longer periods of time in the same cohort.60,61
Aggression and conduct disorders
Similar to ADHD, conduct problems are prevalent in pediatric patients with TBI both before and after injury, suggesting they may increase the risk of injury.29 Max et al.57 found that in children injured from ages 6 to 14 years (n=50), 20 to 40% experienced an increase in oppositional defiant disorder at various points during the first 2 years after injury. Even mild TBI during preschool may put children at significant risk for oppositional defiant disorder/conduct disorder later in childhood (e.g. ages 7–13y).38 Compared with uninjured peers, children with TBI may also engage in more parent- and teacher-rated aggressive behaviors, including temper tantrums and destructiveness.69 Generally, however, there is a lack of strong evidence regarding the emergence of aggression after TBI. For instance, although Cole et al.26 reported increased verbal and physical aggression towards objects and others after TBI, the lack of age-matched comparisons and normative data renders it difficult to discern whether these findings reflect the impact of TBI, the typical developmental trajectory for these children, or the consequence of a general injury and hospitalization experience. Another difficulty is the lack of sensitive, age-appropriate instruments to measure aggressive behaviors. For instance, in comparing aggressive or delinquent behavior in adolescent males with a history of childhood TBI (n=38, mean 8.3y at injury) with healthy peers (n=28), measures of global psychopathology (e.g. Child Behavior Checklist) did not yield any differences between groups; in contrast, measures of aggression based on theoretical frameworks (e.g. social learning theory) found males with TBI to exhibit significantly more reactive and proactive aggression (p<0.05) and to show significantly more aggression out of frustration (p<0.01).27
Aggressive behaviors presenting after TBI are correlated with impaired self-regulation in cognitive, emotional, and behavioral domains that result from abnormalities in the prefrontal cortex.69 Increased aggression may also be predicted by impairments in social problem-solving skills caused by brain damage; children and adolescents with a history of moderate to severe TBI are reported to be more likely to use aggressive and avoidant solutions70 and to provide self-centered and impulsive solutions to social conflicts, rather than collaborative strategies that preserve relationships.71
Novel depressive disorders emerging up to 2 years after TBI occur in 10 to 25% of school-aged children.52,66,67 One study noted that shortly after injury, self-reported depression was not significantly different between children sustaining moderate to severe TBI (n=89, 6–12y old at injury) and those sustaining orthopedic injuries (n=55), suggesting that initial depressive symptoms may accompany the general injury and hospitalization experience.51 Importantly, however, the group with TBI was at significantly greater risk for maintaining self-reported depressive symptoms over time (p<0.05).51 Neurocognitive sequelae may underline both short- and long-term depression in children sustaining TBI. For instance, lesions in the left inferior frontal gyrus and right frontal white matter are correlated with novel depression 6 months after injury.66 Further research in the presentation and correlates of long-term depression resulting from TBI is warranted.
Owing to some similar pathophysiology, presentations of novel depression and anxiety disorders often coexist.66 Anxiety disorders and symptoms occurring after TBI include posttraumatic stress disorder,28 phobias,31 obsessive-compulsive symptoms,30 and generalized anxiety disorder.63 Six months after injury, children hospitalized for mild to severe TBI are at significantly greater risk for presenting new-onset mood and/or anxiety disorders than those with orthopedic injuries (46% vs 14%, p=0.001).63 Children may also demonstrate significantly increased overanxious symptoms (e.g. worries about performance, worries about events before they happen, and frequently asking adults for reassurance) after injury, although it is unclear whether these are related to brain damage or the injury experience itself.28
One study found that although children with moderate TBI, severe TBI, and orthopedic injuries all reported similar frequencies of posttraumatic stress symptoms 6 months after injury, significant group differences emerged by 1 year (p<0.05).72 By 1 year after injury, approximately 10% of patients with TBI aged 9 to 14 years met the three criteria for posttraumatic stress disorder: re-experiencing, avoidance, and hyperarousal.73 Patients satisfying these criteria, compared with those with TBI who did not, tended to have lower lesion fractions in the right medial frontal cortex, as well as greater lesion fractions in the left middle temporal gyrus.73 In the same cohort, posttraumatic stress symptoms and disorder at 1 year after injury were also predicted by self-reported anxiety and depression symptoms experienced an average of 48 days after injury.28 Another lesion-based study also reported that left orbitofrontal cortex lesions decreased the risk of hyperarousal symptoms, that increased volume and presence of left temporal lesions increased risk for hyperarousal symptoms and posttraumatic stress symptoms, and that a greater number of diffuse brain lesions increased the risk of avoidant symptoms.32
Orbitofrontal cortex lesions also appear to decrease the risk for new-onset obsessions, suggesting that the orbitofrontal cortex may be a central area for expression of general anxiety symptoms including obsessive symptoms.30 New-onset obsessions are also significantly associated with lesions in the temporal lobe, which contains the amygdalo-hippocampal complex connecting to brain areas implicated in obsessive-compulsive disorder (e.g. anterior cingulate, orbitofrontal cortex), and in the mesial prefrontal lobe, which includes the anterior cingulate cortex, the possible neural basis of obsessive symptoms and disorders (p<0.05).30
Children injured at younger ages appear more vulnerable to anxiety disorders.31,64 Varied methodologies and outcome measures have probably contributed to the mixed findings about the predictive effects of pre-injury anxiety symptoms and severity.31,63,64 Although post-injury stress is a robust predictor for new-onset anxiety disorders63 and social status may correlate with levels of child posttraumatic stress symptoms,72 there is little support in the recent literature for other environmental predictors of anxiety, including socio-economic status, parental education, pre-injury family functioning, history of anxiety disorders, and psychosocial adversity.31,63,64
Personality change due to TBI is an important clinical problem among pediatric patients with TBI. In children, this psychiatric disorder is not characterized by stable personality changes; rather, it is associated with impairing behavioral disturbances and deviations from normal development.54 Personality change occurs in approximately 10 to 20% of patients with mild to severe TBI up to 2 years after injury.55,56 Strongly associated with injury severity, personality change is the most common novel disorder after severe TBI.52–54 Transient personality change may occur in up to 60% of patients with severe TBI and 5% of patients with mild to moderate TBI; labile, aggressive, and disinhibited subtypes are most common.54 Onset for personality change usually occurs within the first 3 months after injury.53,55 Lesions in the superior frontal gyrus are significantly associated with the development of personality change within 1 year after injury (p<0.01),55,56 whereas lesion correlates in the frontal lobe white matter have been identified at 2 years after injury (p<0.02).56 The contribution of frontal lesions to personality change probably reflects the importance of the dorsal frontal system in effortful regulation of affective states.56 Persistent personality change at 2 years after injury, which appears specific to patients with severe TBI, is reported to be significantly associated with IQ (p<0.05) but not psychosocial variables.53
Factors Influencing the Brain Injury/Behavioral Outcome Relationship
Pre-injury behavior and functioning are suggested to be strong predictors for long-term behavioral problems.43,68 Increased aggression after TBI, for instance, is most strongly predicted by pre-injury aggression, followed by pre-injury inattention (independent of pre-injury aggression) and pre-injury anxiety.26 Pre-injury conditions such as anxiety and attention problems can be exacerbated by TBI,31,47 and their presence may also impact on the presentation of post-injury symptoms. For instance, Levin et al.58 noted that although symptoms of ADHD remained higher and more stable in children with pre-injury ADHD, those without a pre-injury diagnosis showed more fluctuations with nonlinear changes in inattentive and hyperactive symptoms.
Age at injury
Although it was previously thought that children recover better from brain damage than adults, owing to the plasticity of the young brain, the recent literature has countered this with an emerging ‘theory of vulnerability’ that suggests the developing brain may be more sensitive to – and show less recovery from – diffuse brain injury.6,74 However, it is unclear whether this ‘theory of vulnerability’ relates as clearly or linearly to behavioral outcomes as it does to neurocognitive and functional ones.75 For instance, despite the strong relation between injury timing and deficits in attentional skills after brain injury,41 timing does not significantly predict behavioral deficits such as secondary ADHD.60–62 Furthermore, children injured in preschool or earlier have demonstrated good behavioral outcomes and do not show impairments compared with comparison children or population norms.25,36,37 It is possible that the inability to detect impairments in these young children reflects the use of observational measures that more readily identify a lack of age-appropriate behavioral functioning in older children.40 Impairments may also take several years to manifest, as other studies have found children injured in preschool have significant behavioral impairments from 7 to 13 years of age.38 Still, however, one study reported that children sustaining closed head injuries at ages 11 to 15 years had significantly more behavioral problems than those injured at ages 6 to 10 years.42 The relationship between injury timing and behavioral outcome may be complex because the effect of injury timing varies with outcomes. For instance, younger children are more vulnerable to anxiety after TBI,31,64 but older age at injury is predictive of depression.66 When these disorders coexist as anxious subsets of depression, age appears no longer predictive.66 Age at injury may also affect potential cognitive mediators and mechanisms behind behavioral functioning and regulation. For example, younger age at injury significantly correlates with social problem-solving71 and information processing.34
Sex and biological factors
Studies of patients with severe TBI have demonstrated female sex to predict significantly the development of posttraumatic stress disorder29 and obsessive-compulsive symptoms,30 which may reflect greater female susceptibility to these in the general population as well.29 One study also reported that male children had higher rates of elevated behavioral problems 6 months after injury; however, this effect of sex disappeared by 12 months.43 Finally, genetics may also predispose children for developing disorders of depression after TBI.66
Type of injury
Few studies have focused on the role of injury mechanisms in predicting behavioral outcomes. An exception occurs in studies of infants, who are frequently subjected to inflicted TBI (e.g. shaken baby syndrome, physical abuse). Children with inflicted TBI perform significantly more poorly on cognitive measures of development than those sustaining accidental TBI,36,37 and survivors of inflicted TBI may show frequent behavioral problems such as hyperactivity and rage reactions by the age of 2 to 3 years.35 However, recent findings do not support that behavioral outcomes in cases of inflicted TBI are significantly poorer than those of accidental cases.36,37 In comparing children sustaining inflicted and accidental TBI, Keenan et al.36 found that despite worse outcomes in inflicted TBI cases, the degree of disability was more strongly associated with injury severity than mechanism. Finally, within cases of inflicted TBI, the type of injury (e.g. whiplash, impact) is not significant.35
Although injury severity is strongly associated with the emergence of certain disorders such as personality change,53,54 the effects of severity are unclear for most behavioral outcomes. Several studies have reported that severe TBI is associated with significantly greater risk in developing externalizing problems,34,37,40 anxiety disorders and symptoms,28,63 and attention,40,59 with some studies even reporting little difference between the outcomes of children with mild or moderate TBI and those of comparisons or population norms.40,41,52,59 However, several studies have failed to find severity a significant predictor of behavioral outcomes altogether.26,31,45,48,49,69,76 Severity may also play a role in whether specific symptoms decrease, persist, or even increase over time. For instance, from baseline to an average of 4 years after injury, emerging behavioral problems decreased in an orthopedic injury comparison group, remained stable in cases of moderate TBI, and increased in cases of severe TBI.43 Other studies have additionally found significant severity × time effects, even within the severity gradient of mild TBI.38
Low socio-economic status is significantly associated with behavioral outcomes40,43,44,51,58,60,71 with a notable exception for anxiety problems shortly after injury.31,63 Recently, studies have begun to focus on more specific contributions of family functioning and environment that can significantly contribute to various behavior outcomes.29,40 For instance, permissive parenting may significantly predict40 as well as moderate the development of long-term behavioral problems like ADHD.77,78 In a sample of children aged 3 to 7 years at injury, maladaptive parenting styles were associated with better behavioral adjustment shortly after injury but worse adjustment at 18-month follow-up, suggesting that permissive and authoritarian parenting may initially suppress problems but can, over time, exacerbate them.78 Recently, following children aged 3 to 7 years with severe TBI, researchers found that even after controlling for socio-economic status, parental distress, and family functioning, parental warmness was associated with lower levels of both internalizing and externalizing behavioral problems and ADHD, whereas negativity was associated with more externalizing and ADHD symptoms.79 Another study found that externalizing problems in children injured before the age of 2 years were predicted by the strength of a defined leader in the family.37 Caution should be taken, however, in drawing conclusions of a causal relation between parenting and behavioral outcome. For instance, as behavioral problems become increasingly difficult to control, parents may become more permissive in their parenting approach.77
Behavioral problems are prevalent after brain injury and may persistent and even worsen over time.40,43,44,49,51,68 Lesion correlates suggest that acute onset of some behavioral problems, such as anxiety and depression, may increase after TBI because of damage to specific areas of the brain.31,32,64,66 Feelings of depression and frustration may also accompany the general experience of trauma and hospitalization;51 importantly, brain-injured children appear more vulnerable than those with orthopedic injuries to maintaining these adverse symptoms.50,51 Whereas disorders such as personality change due to TBI, may be noticeable shortly after injury (e.g. <3 months),53–56 other problems such as increased aggression and hyperactivity may not fully manifest until several years after injury.38
Although the pathophysiology of TBI is beyond the scope of this paper and has been thoroughly discussed elsewhere,80–83 injury may result in brain dysfunction and damage that can mediate the persistence and even secondary onset of behavioral problems. Notably, in TBI inertial forces resulting from angular and rotational acceleration–deceleration motions are applied to the head, and, regardless of impact, these can deform white matter, mechanically disrupt nerve fibers through shearing forces, and lead to diffuse axonal injury.80 This commonly results in diffuse, traumatic axonal injury, an important pathophysiological consequence of TBI. Secondary consequences due to pathology (e.g. hemorrhagic and perfusion mechanisms) and neurotransmitter cascades may also contribute to the resulting diffuse damage,84 which may lead to global dysfunction by disconnecting and altering the functional interaction of tens of thousands of neurons.81 The frequent damage after TBI to neural substrate, especially in frontal, temporal, and parietal cortical structures,85–88 may have important implications in neurocognitive skills that mediate behavioral outcomes. For instance, the development of an individual’s executive functioning relies on frontal–striatal circuits that are particularly vulnerable to diffuse axonal injury.88,89 Executive function refers to higher-order skills that allow goal-directed behavior, such as goal setting, inhibition, attentional control, sustained effort, and mental flexibility. Roughly 20 to 40% of children experiencing TBI between 5 and 15 years of age show significant executive dysfunction within the first year of injury,90 and such impairments may mediate the relationship between reduced protective factors (e.g. resiliency) and increased behavioral problems like depression and anxiety.50 The ability to process emotional information, regardless of modality, also uses an extensive network that includes frontal and temporal areas,91 and behavioral problems in children with acquired brain injuries (e.g. TBI, stroke, meningitis) are shown to be predicted by deficits in processing and recognizing emotions, independent of facial identification abilities, executive function, and general cognitive functioning.92
Clinicians should therefore ensure adequate follow-up with assessments of neurocognitive domains (e.g. emotional recognition, executive function) to identify children at risk of developing behavioral problems and help children receive appropriate interventions and treatments. Furthermore, it is critical to acknowledge the environmental correlates that are useful in designing and informing intervention efforts. Family environment plays an important part in the developmental changes of injured children, especially within the first year of injury.47,61 Parental training and education soon after the TBI event are thus important considerations in long-term rehabilitation efforts. Caregivers should also be encouraged to tell teachers about their child’s TBI status, as teachers frequently do not know about a child’s history of TBI48 but can play crucial roles in fostering resilience and protecting children from adverse behavioral outcomes, such as through improving children’s self-esteem and feelings of success.93
Continued research is necessary in understanding the impact of pediatric brain injury on long-term behavioral outcomes. Research in this area has important limitations that should be acknowledged. First, studies vary greatly in sample, methodologies, instruments, and outcomes measured. Even within studies, the use of different behavioral outcome measures often yielded different findings about the presence and significance of behavioral problems, such as in cases of using parent- or self-report for depression51 and anxiety,72 behavioral rating instruments or psychiatric interviewing procedures for general internalizing disorders,67 and global psychopathology measures or theoretically based measures for aggression.27 Second, most studies are limited by methodological weaknesses, including a reliance on parent-report measures, retrospective recall for pre-injury behaviors, small samples with reduced statistical power, short-term follow-up, and a lack of matched reference groups. Reliance on parental report, for both children’s current behavior as well as pre-injury status, is particularly problematical because parents’ perceptions of their children’s status may be distorted by the TBI event as well as their own distress caused by it.14 Third, research in this field appears to lack age-appropriate, sensitive, and reliable measures for identifying specific behavioral disturbances within pediatric survivors of TBI. The most commonly used measures are those of global psychopathology, which may not optimally detect behavioral changes after diffuse brain injury. Fourth, commonly used imaging techniques such as computed tomography may also be insensitive to the detection of small lesions, subcortical lesions, and diffuse injuries, and thus fail to detect the full extent of injury to the young brain.59 Newer technologies that have become available, such as susceptibility weighted imaging and diffusing tensor imaging, allow the detection of diffuse axonal injury and changes in white matter tracts from TBI30 and should be used in future studies. Finally, very few studies examining the brain injury/behavior relationship have hypothesized and tested mediation and moderation effects of cognitive and environmental factors. Future research must place a greater emphasis on investigating the interactive relationships between cognitive, environmental, and behavioral domains in the aftermath of injury to the developing brain. These efforts are critical in reducing the threat of pediatric brain injury as a cause for detrimental lifelong outcomes.
Limitations to the present review are also worth acknowledging. First, our selection of search terms may have limited our ability to cover thoroughly all relevant publications and findings. However, we attempted to identify any eligible study we may have missed through keyword or database selection by using ancestry. Second, we were only able to include studies published in English. It is unclear whether the fact that most studies included in the present review used samples from the USA, UK, or Australia reflects this limitation or simply a lack of brain injury and behavior research from other parts of the world, such as Asia. Thus, we are also unable to generalize the present review’s findings. Third, given that our review covers a period of over 20 years, advances in neuroimaging that have allowed more accurate identification of brain damage and severity are not necessarily represented in the current paper.
This review illustrates that children sustaining brain injuries are at risk for newly emerging behavioral problems, although problems are heterogeneous in both etiology and presentation and may not manifest until several years later. Even mild injury may damage a child’s neurocognitive abilities, causing deficits that may persist for several years. Developing skills such as executive function and social problem-solving are also often damaged in TBI, and such neurocognitive vulnerabilities can in turn mediate the persistence of continued or even secondary behavioral problems in children. As children age, they may become increasingly aware of their impairments, which probably become more salient as they are faced with increasing demands and expectations (e.g. higher cognitive functions, academic demands, behavioral and emotional regulation). Worsening behavioral outcomes may thus reflect actual neurocognitive sequelae as well as children’s increasing frustrations of poor school performance and peer relations that may subsequently occur. Given the potential for lifetime behavioral impairments for survivors of pediatric TBI, it is critical that families and patients receive adequate follow-up care and appropriate information about potential behavioral consequences. Future prospective studies, especially those examining neurocognitive and environmental correlates, are necessary to minimize the onset and persistence of these detrimental behavioral problems.
This research was funded in part by National Institutes of Health/National Institute of Environmental Health Sciences, K01-ES015 877; R01-ES018858 to Dr Liu.