Abstract
This study examined the associations among brain volumes, theory of mind (ToM), peer relationships, and psychosocial adjustment in children with traumatic brain injury (TBI). Participants included 8- to 13-year-old children, 82 with TBI and 61 with orthopedic injuries (OIs). Children completed three measures of ToM. Classmates provided ratings of participants’ peer relationships, acceptance, and friendships. Parents rated children’s psychosocial adjustment. MRI was used to determine brain volumes. Brain volumes were associated with ToM, which in turn was associated with peer rejection/victimization. Peer rejection/victimization in the classroom was associated with peer acceptance, friendship, social withdrawal, and general psychopathology. Brain volumes, ToM, peer relationships, and social adjustment show significant links among children with TBI and those with OI. The findings support a multilevel model of social competence in childhood TBI.
In the United States, traumatic brain injury (TBI) accounts for at least 700,000 hospital visits annually among children and adolescents (Faul, Xu, Wald, & Coronado, 2010). Children with TBI demonstrate a range of cognitive, emotional, and behavioral deficits, especially after more severe injuries (Yeates, 2010). However, surprisingly little is known about the social outcomes of children with TBI, despite the importance of social competence in childhood as a predictor of a host of other important outcomes (Rubin, Bukowski, & Parker, 2006).
Children with chronic medical conditions affecting the central nervous system are at risk for poor psychosocial adjustment (Martinez, Carter, & Legato, 2011). Moreover, deficits in social behavior have been observed in animal models of TBI (Semple, Canchola, & Noble-Haeusslein, 2012), and the few existing studies of social outcomes in childhood TBI suggest that children with severe TBI show deficits in social information processing, peer relationships, and social adjustment (Rosema, Crowe, & Anderson, 2012; Yeates et al., 2007). Nonetheless, research on the social problems faced by children with TBI remains relatively sparse, particularly with regard to how different levels of social function are associated after childhood TBI.
To address these lacunae, we recently conducted a multilevel study of social outcomes in children with TBI. The project was guided by a model of social competence in childhood brain disorder that draws on both social neuroscience and developmental psychology (Yeates et al., 2007). The model defines social competence as the ability to achieve personal goals in social interaction while simultaneously maintaining positive relationships with others over time and across situations (Rubin & Rose-Krasnor, 1992). The model distinguishes among three key levels of social competence, namely, social information processing, social interaction, and social adjustment, and assumes significant relationships among those different levels. The model acknowledges that a network of specific brain regions underlies social information processing and social interaction, so that childhood brain disorders such as TBI can influence social competence at all levels.
Our study, which was cross-sectional in design, compared the social outcomes of 8- to 13-year-old children with complicated-mild to severe TBI to those of children with orthopedic injuries (OIs). Participants completed a series of assessments on average 2.5 years postinjury: (a) MRI of the brain; (b) direct measures of executive functions and social cognition, including theory of mind (ToM); (c) classmate ratings of peer relationships, acceptance, and friendship, along with direct observations of interactions with friends and unfamiliar peers; and (d) parent ratings of social adjustment. The project had several aims: (a) to characterize social information processing, social interactions, and social adjustment in children with TBI; (b) to determine the integrity of brain systems vulnerable to TBI and also implicated in social information processing; and (c) to study the linkages among brain abnormalities, social information processing, and social interactions and adjustment.
In recent articles based on the same participants as those described here, we have begun to address the first two aims by focusing on group differences in outcomes at the three levels of our model (Bigler et al., 2013; Dennis et al., 2012; Dennis, Agostino, et al., 2013; Dennis, Simic, Agostino, et al., 2013; Dennis, Simic, Bigler, et al., 2013; Yeates et al., 2013). Collectively, our previous articles have shown that children with TBI display a variety of deficits compared to children with OI: (a) volumetric reductions and more frequent lesions in regions of the brain implicated in social information processing and behavior, (b) deficits in multiple forms of ToM, and (c) higher levels of rejection-victimization and fewer mutual friendships among school classmates. We also have shown that the volumes of specific brain regions are positively associated with better performance on certain ToM tasks and with higher mutual friendship ratings. However, we have not yet conducted a systematic test of the associations among the constructs in our model. Moreover, we have not incorporated parent ratings of psychosocial adjustment in our previous articles.
The primary goal of the current study, therefore, was to examine the associations among brain volumes, ToM, peer relationships, and psychosocial adjustment in children with TBI and OI. Our first hypothesis was that volumetric reductions in the social brain network (Adolphs, 2001; Johnson et al., 2005) would be associated with poorer performance on ToM tasks. In contrast to our previous article (Dennis, Simic, Bigler, et al., 2013), which relied on an a priori, region-of-interest (ROI) approach to relate specific brain networks to ToM, we used statistical parametric mapping and voxel-based morphometry (VBM; Ashburner & Friston, 2000) to determine which brain regions are associated with ToM. This approach has less power than the ROI approach, but allowed us to examine volumetric relationships across the entire brain. Our second hypothesis was that poorer ToM would be associated with greater rejection-victimization as reported by the participants’ classmates. We focused on rejection-victimization because it was the dimension of peer relationships that showed the largest group difference in our sample (Yeates et al., 2013). Our final hypothesis was that higher levels of peer rejection-victimization would be associated with lower ratings of peer acceptance and fewer friendship nominations by classmates, as well as with poorer psychosocial adjustment as rated by parents.
Methods
Participants
Participants were recruited from children’s hospitals at three metropolitan sites, including the Hospital for Sick Children in Toronto (Canada), Nationwide Children’s Hospital in Columbus, Ohio (United States), and Rainbow Babies and Children’s Hospital and MetroHealth Medical Center in Cleveland, Ohio (United States). Eligible participants included children hospitalized for either a TBI (n = 82) or OI (n = 61) who were from 8 to 13 years of age at the time of their participation and injured between 12 and 63 months prior to participation. All children were injured after 3 years of age. The TBI group was restricted to children with complicated mild to severe TBI, as defined by a lowest postresuscitation Glasgow Coma Scale (GCS; Teasdale & Jennett, 1974) score of 12 or less, or a GCS score of 13 to 15 in association with trauma-related abnormalities on neuroimaging at the time of hospitalization. The OI group consisted of children who sustained fractures not associated with any loss of consciousness or other risks or indications of brain injury (e.g., skull or facial fractures).
The following exclusion criteria applied to all children: (a) history of more than one serious injury requiring medical treatment, (b) premorbid neurological disorder or mental retardation, (c) any injury determined to be a result of child abuse or assault, (d) a history of severe psychiatric disorder requiring hospitalization prior to the injury, (e) any sensory or motor impairment that prevented valid administration of study measures, (f) placement in full-time special education classroom, (g) not fluent in English, and (h) any medical contraindication to MRI.
Among children eligible to participate and approached about the study, 47% of those with TBI and 26% of those with OI agreed to enroll. Although the participation rate was significantly higher for TBI than for OI participants, participants and nonparticipants in both groups did not differ in age at injury, age at initial contact about the study, sex, race, or census tract measures of socioeconomic status (SES) that included median family income, percentage of minority heads of household, and percentage of households below the poverty line. Participants and nonparticipants also did not differ on measures of injury severity (i.e., mean length of hospitalization, median GCS score for children with TBI).
Tables 1 and 2 presents demographic and injury characteristics for the two injury groups. The groups did not differ in sex, race, ethnicity, age at injury, or age at assessment. The groups differed significantly in full-scale IQ and SES (measured using a standardized composite based on parental education, parent occupational status, and census tract median family income), with the TBI group having the lowest mean IQ and SES, although IQ scores for both groups were in the average to high-average range. The groups also differed in the distribution of mechanism of injury. Group differences in SES were not significant when injury mechanism was taken into account, consistent with epidemiological studies showing that the risk of TBI, particularly those linked to motor vehicles, is highest for children of lower SES and minority status (Brown, 2010; Howard, Joseph, & Natale, 2005; Langlois, Rutland-Brown, & Thomas, 2005). For that reason, we did not treat SES as a covariate in data analyses because the SES differences appeared to be intrinsic to the injury groups (Dennis et al., 2009).
Group Demographics and Injury Mechanisms
Note: OI = orthopedic injury group; TBI = traumatic brain injury.
Groups differ significantly, p < .05.
Participant Characteristics
Note: OI = orthopedic injury group; TBI = traumatic brain injury group; SES = socioeconomic status.
IQ measured using two-subtest version of Wechsler Abbreviated Scale of Intelligence.
Groups differ significantly, p < .05.
Classroom data were available for 87 of 143 (61%) children from the larger study (n = 55 of 82 with TBI, n = 32 of 61 with OI). Completion of classroom data collection was not significantly associated with injury group, sex, race, age at injury, or age at assessment. For neuroimaging analyses, useable data were available for 106 children (TBI = 59, OI = 47). The availability of useable imaging data did not vary by group. Reasons for missing data were poor scan quality as a result of movement or artifact (either postsurgical for children with TBI or from dental braces) or incomplete imaging sequences due to claustrophobia, equipment failure, or participant refusal. In addition, two children with severe TBI had such extensive brain abnormalities as a result of their injuries that their scans were not amenable to automated image analysis.
Procedure
Institutional review boards approved all study procedures prior to recruitment, and informed parental consent and child assent were obtained prior to participation. As part of a larger assessment, children completed MRI and ToM tasks and parents provided ratings of psychosocial adjustment. Classroom data collection occurred after the assessment, and was preceded by written and phone contact with principals, to obtain their permission to contact participants’ teachers. The study was explained to teachers, who distributed and collected parental consent forms from students. To protect confidentiality, the project was described to students as a study of friendships without mentioning traumatic brain injuries or identifying the participating child. Questionnaires were administered during a single group session in the primary classroom for elementary school students or a required academic subject for students in middle school. Classroom data were not collected during the first 2 months of the school year to ensure that children were familiar with one another prior to completing ratings. The groups did not differ in the average number of months that had passed during the school year prior to classroom collection (M = 6.30, SD = 2.09).
MRI and voxel-based morphometry
Magnetic field strength was 1.5 Tesla for all MRI studies. The Toronto and Columbus sites used GE Signa Excite scanners, and the Cleveland site used a Siemens Symphony scanner. All sites acquired the following sequences on each participant: thin slice, volume acquisition T1-weighted ultrafast 3D gradient echo, commonly referred to as MPRAGE or FSPGR (depending on scanner manufacturer); a dual-echo proton density (PD)/T2-weighted sequence; FLAIR; and GRE. Standard phantoms were used for quality control and to calibrate scans across sites.
FreeSurfer (surfer.nmr.mgh.harvard.edu) was used to derive whole-brain volumetric measurements, using methods described by Bigler et al. (2010). Volumes obtained included total white matter (WM), total gray matter (GM), and total ventricular volume, along with total brain volume (TBV), summing for right and left hemisphere structures. Also, a ventricle-to-brain ratio (VBR) was obtained by using the FreeSurfer computation of total ventricular volume divided by TBV and multiplied by 100 to provide whole number values (Tate, Khedraki, Neeley, Ryser, & Bigler, 2011).
For VBM, the major preprocessing and analysis steps were done in SPM8 (Wellcome Trust Centre for Neuroimaging, London, UK; available at http://www.fil.ion.ucl.ac.uk/spm) using the VBM8 toolbox (available at http://dbm.neuro.uni-jena.de/vbm). The images were (a) corrected for bias-field inhomogeneities, (b) registered using a linear (i.e., 12-parameter affine) and a nonlinear transformation, and (c) stripped of nonbrain tissue in the T1-weighted images. The SPM8 unified segmentation package (Ashburner & Friston, 2005) was used to initialize a VBM8 algorithm-based classification of GM, WM, and CSF, following the methods outlined by Ziegler et al. (2012).
GM and WM population templates were generated separately using the entire image dataset, using the Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL) technique (Ashburner, 2007). We performed affine registration of the DARTEL templates to the tissue probability maps in Montreal Neurological Institute (MNI) space (see http://www.mni.mcgill.ca/), along with nonlinear warping of images to the DARTEL templates in MNI space with a 1.5 mm3 resolution (as recommended for the DARTEL procedure). GM and WM volumes at each voxel were obtained through modulation, which was performed by multiplying the respective GM or WM concentration map by the Jacobian determinants derived from the nonlinear spatial normalization step, followed by smoothing of volume images with a full width at half maximum kernel of 8 mm. After all these spatial preprocessing methods were completed, the smoothed, modulated, and normalized volume maps for GM and WM were used for statistical analysis.
Theory of mind tasks
Three tasks were administered to assess three types of ToM: cognitive, affective, and conative (Dennis, Simic, Bigler, et al., 2013). Cognitive ToM is the original mind-reading sense of ToM, as reflected in understanding of false beliefs. Affective ToM is reflected in the understanding of the distinction between felt versus displayed emotion (i.e., emotive communication, in which the expression on the face is deceptive; Hein & Singer, 2008). Conative ToM refers to forms of social communication in which we try to influence the mental and emotional state of others. Ironic criticism and empathic praise are prototypical forms of conative ToM.
For each type of ToM, tasks were developed to permit comparisons of ToM trials to control trials that involve the same task demands but do not require ToM. The Jack and Jill task (Dennis et al., 2012) measures cognitive ToM with a series of trials that measure false belief, as compared to a series of control trials that measure true belief. For this study, percentage accuracy on false belief trials was used as a measure of cognitive ToM. The Emotional and Emotive Faces task (Dennis, Agostino, et al., 2013) measures affective ToM with trials requiring identification of emotions expressed for social purposes (i.e., emotive communication) as compared to control trials requiring identification of emotions actually felt. For this study, percentage accuracy on trials involving emotive communication was treated as a measure of emotional ToM. The Ironic Criticism and Empathic Praise task (Dennis, Simic, Agostino, et al.,, 2013) measures conative ToM with trials requiring the child to identify the beliefs and intentions underlying communications involving irony and empathy, as compared to control trials probing for beliefs and intentions underlying literally true statements. For this study, percentage accuracy on trials involving irony and empathy was treated as a measure of conative ToM. Sample stimuli for the three types of ToM, along with more task details, are available in Dennis, Simic, Bigler, et al. (2013).
Classroom measures
Extended class play (ECP)
Participants completed an extended version of the Revised Class Play (Masten, Morison, & Pellegrini, 1985) to assess social behavior with peers. Administration of the ECP requires students to imagine that they are the director of a play and to “cast” one boy and one girl from the class into 31 hypothetical roles. Factor analysis has identified five subscales: Popular-Sociable, Prosocial, Aggressive, Rejected-Victimized, and Shy-Withdrawn (Rubin, Wojslawowicz, Burgess, Rose-Krasnor, & Booth-LaForce, 2006; Wojslawowicz Bowker, Rubin, Burgess, Rose-Krasnor, & Booth-LaForce, 2006). Tallies of nominations received from classmates for each role were standardized within sex in each class to adjust for unequal class size, composition, and participation rates, and then summed. The summed scores were standardized (M = 0, SD = 1) within sex in each class to create subscales. The scores reflect nominations relative to same-sex peers.
Peer acceptance ratings
Students rated how much they liked each classmate on a 5-point scale (Asher, Singleton, Tinsley, & Hymel, 1979). Mean acceptance ratings were standardized within sex for each class.
Best friend nominations
Students recorded the names of their three “best friends” from the class. Scores were computed for total friendship nominations and for mutual friendship nominations, which reflects the number of friendship nominations each child received from classmates who also identified the child as a friend (Bukowski & Hoza, 1989). Both scores were standardized within sex for each class.
Measures of psychosocial adjustment
Parents rated children’s social adjustment using the second edition of the Adaptive Behavior Assessment System (ABAS-II; Reynolds & Kamphaus, 2004). The ABAS-II is a parent self-report of behavioral skills that are important in coping with the demands of daily life in multiple settings (e.g., home and school). It assesses multiple adaptive areas; in the current study, the Social subscale was used as a measure of social adjustment. The ABAS-II is well standardized, demonstrates good reliability and internal consistency within scales, and has been shown to be sensitive to developmental differences and clinical disorders.
Parents rated their children’s emotional and behavioral adjustment using the second edition of the Behavior Assessment System for Children (BASC-2; Harrison & Oakland, 2003). It provides multiple subscales; for the purposes of the current study, we chose the Aggression, Withdrawn, Functional Communication, and Social Skills scales as measures of psychosocial adjustment, because they focus on behaviors that are predominantly social in nature. We also examined the Behavioral Symptoms Index as a summary measure of psychopathology. The BASC-2 is well standardized and has shown good internal consistency and test–retest reliability. The construct validity of the scales is well documented, as is its predictive validity and sensitivity to clinical disorders.
Statistical analyses
Brain volumes and theory of mind
Overall measures of brain volume (total GM, WM, total ventricular volume, and VBR) were first correlated with the three ToM measures, to determine if social information processing was related to general brain development or atrophy. Partial correlations were computed within the OI and TBI groups separately and across the entire sample, controlling for age at testing (group was also controlled when the entire sample was combined).
Voxel-based partial correlation analyses were then carried out in SPM8 to test associations between GM and WM volumes and ToM. Age, sex, and MRI acquisition site were treated as covariates, to reduce error variance and to control for variations between scanners across sites. We first identified significant individual voxels using an uncorrected threshold of T of 2.5 (p < .008; Friston, Holmes, Poline, Price, & Frith, 1996). We next identified clusters of voxels that were significant based on a threshold of p < .01 after familywise error correction. We interpreted only those clusters that surpassed a predetermined size threshold, which was the expected number of voxels per cluster, as computed by SPM8 based on random field theory. Analyses were completed separately for the TBI and OI groups, as well as with both groups combined; in the latter analyses, group was also treated as a covariate.
Theory of mind and peer relationships
The three ToM measures were entered as predictors of rejection-victimization in a hierarchical multiple regression analysis. We examined the change in R2 accounted for by ToM, over and above group differences (TBI vs. OI), as well as the individual β for each ToM measure, to assess their unique contributions. Test of Group × ToM interactions were conducted to determine whether the association between ToM to rejection-victimization varied across the two groups, but none of them was significant, so they are not reported here.
Peer relationships and psychosocial adjustment
In a series of hierarchical multiple regression analyses, rejection-victimization was entered as a predictor of mean acceptance ratings, total friendship nominations, and mutual friendship nominations provided by classmates, as well as parent ratings on the ABAS-II Social scale and BASC-2 Aggression, Withdrawn, Functional Communication, Social Skills, and Behavioral Symptom Index scales. We examined the change in R2 accounted for by rejection-victimization, over and above group (TBI vs. OI). Tests of the Group × Rejection-Victimization interaction were conducted to determine whether the association between rejection-victimization and psychosocial adjustment varied across the two groups, but none was significant, so they are not reported here.
Results
Brain volumes and theory of mind
Several significant correlations (all p < .05) were obtained between overall brain volumes and the three ToM measures. Across both groups, conative ToM was positively correlated with GM and WM volumes (r = .29 and .25, respectively) and negatively correlated with the VBR (r = −.17), after controlling for age and group membership. Within the OI group, cognitive ToM was positively correlated with GM volume (r = .32), affective ToM was negatively correlated with the VBR (r = −.30), and conative ToM was positively correlated with GM volume (r = .29). Within the TBI group, conative ToM was positively correlated with GM and WM volumes (r = .29 and .25, respectively) and negatively correlated with total ventricle volume and VBR (r = −.27 and −.31, respectively).
VBM identified two significant clusters of voxels associated with ToM, but only within the OI group (see Tables 3 and 4). After controlling for age, sex, and MRI acquisition site, cognitive ToM within the OI group was positively correlated with a WM cluster located largely within the cerebellum and brainstem (see Fig. 1a). In addition, conative ToM was positively correlated with GM volume within a cluster that involved regions within the thalamus, basal ganglia, and limbic system. Specific regions within the cluster included the medial dorsal nucleus, putamen, globus pallidus, subcallosal gyrus, anterior cingulate, and amygdala (see Fig. 1b).
Results of VBM Analysis Showing Positive Correlation of WM Volume With Cognitive ToM in the OI Group
Note: MNI = Montreal Neurological Institute; OI = orthopedic injury; ToM = theory of mind; VBM = voxel-based morphometry; WM = white matter.
Results of VBM Analysis Showing Positive Correlation of GM Volume With Conative ToM in the OI Group
Note: GM = gray matter; MNI = Montreal Neurological Institute; OI = orthopedic injury; ToM = theory of mind; VBM = voxel-based morphometry.
Significant results of voxel-based morphometry (VBM) analyses. (a) Voxel cluster in which white matter (WM) volume was positively correlated with cognitive theory of mind in the orthopedic injury (OI) group (right hemisphere on right side). Regions within the cluster included the posterior and anterior lobes of the cerebellum, cerebellar tonsils, pons, and medulla. (b) Voxel cluster in which gray matter (GM) volume was positively correlated with conative theory of mind in the OI group. Regions within the cluster included the medial dorsal nucleus, putamen, globus pallidus, subcallosal gyrus, anterior cingulate, and amygdala. For both analyses, individual voxels were identified with uncorrected threshold of T = 2.5 (p < .008) and significant clusters were identified with familywise error corrected threshold of p < .01, with a cluster size threshold that reflected the expected number of voxels per cluster based on random field theory. Analyses controlled for age, sex, and MRI acquisition site. Brain template is an averaged image based on T1 images from children with orthopedic injuries. Scale represents T values for individual voxels.
Theory of mind and peer relationships
Taken together, the three ToM measures accounted for an additional 10% of the variance in rejection-victimization, F(3, 72) = 2.72, p = .051, adding to the 5% accounted for by group membership, F(1, 75) = 3.96, p = .05. Only affective ToM was a unique predictor, β = −.27, t = −2.18, p = .032, with better ToM predicting less rejection-victimization. However, cognitive ToM was marginally associated with rejection-victimization when considered in isolation after controlling for group (partial r = −.20, p = .08), again with better performance predicting less rejection-victimization.
Peer relationships and psychosocial adjustment
After controlling for group differences, rejection-victimization was significantly associated with several measures of psychosocial adjustment: mean peer acceptance ratings, F(1, 83) = 77.06, p < .001; total friendship nominations, F(1, 83) = 24.00, p < .001; mutual friendship nominations, F(1, 80) = 10.08, p = .002; the BASC-2 Withdrawn scale, F(1, 68) = 6.40, p = .014; and the BASC-2 Behavioral Symptom Index scale, F(1, 67) = 5.49, p = .022. Rejection-victimization accounted for significant variance in all measures (47% of the acceptance ratings, 22% of total friendship nominations, 11% of mutual friendship nominations, 8% of the BASC-2 Withdrawn scale, and 7% of the BASC-2 Behavioral Symptom Index), with greater rejection-victimization associated with poorer psychosocial adjustment. Rejection-victimization did not predict significant variance in the ABAS-II Social scale (ΔR2 = .02) or the BASC-2 Aggression (ΔR2 = .015), Functional Communication (ΔR2 = .02), or Social Skills scales (ΔR2 = .00).
Discussion
Children with TBI are at risk for poor social outcomes across a variety of levels, including social cognition, social interaction, and social adjustment (Rosema et al., 2012; Yeates et al., 2007). The current findings also confirm that the outcomes within these domains are likely to be linked, such that deficits in social cognition are associated with problematic social interactions, which in turn are associated with poorer social adjustment. Moreover, volumetric reductions in the brain, both global and local, are associated with deficits in social cognition. Thus, the findings are consistent with the multilevel model of social competence in childhood brain disorder that guided our project (Yeates et al., 2007).
One interesting aspect of the findings is that they generally held true for both children with TBI and children with OI. That is, associations among brain volumes, ToM, peer relationships, and psychosocial adjustment were evident in both groups. In setting forth the model of social competence that guided our study of childhood TBI (Yeates et al., 2007), we argued that the model might also help to characterize social competence in healthy children and thereby contribute to our understanding of both normal and aberrant social development. Our hope was that the model would provide a heuristic framework for research regarding the neural and cognitive-affective substrates of children’s social behavior. The current findings suggest that the links among brain, social cognition, social interaction, and social adjustment extend to children without overt brain disorder.
The associations between brain volumes and social information processing were both global and local in nature. Global measures of brain volume were positively associated with ToM across and within groups, suggesting that social information processing may be related to general brain development in children of these ages, and specifically to diffuse atrophy in children with TBI. This is perhaps not surprising, given that the social brain involves a broad array of spatially remote regions thought to interact as a network (Adolphs, 2001; Johnson et al., 2005). Individual differences in overall brain maturation should involve many of these regions, as should the diffuse atrophy that occurs after more severe TBI (Bigler et al., 2010; Merkley et al., 2008).
At the same time, VBM identified specific, local brain regions that correlated especially strongly with ToM among children with OI. The finding of an association between cerebellar WM and cognitive ToM is at first glance somewhat surprising, as the cerebellum is not often mentioned as a region implicated in social information processing. However, recent research has demonstrated reciprocal connections between the cerebellum and superior temporal sulcus, a region known to play a key role in social information processing, and specifically in ToM (Sokolov et al., 2012). Moreover, several major neurodevelopmental disorders, including autism spectrum disorder and Fragile X syndrome, have been shown to involve cerebellar abnormalities that also may be associated with deficits in ToM (Cornish et al., 2005; Riva et al., in press). The association between conative ToM and subcortical GM in regions including the thalamus, basal ganglia, and limbic system is more compatible with existing knowledge about the social brain network. This network is known to include the ventral striatum, anterior cingulate, and amygdala; in addition, the medial dorsal nucleus of the thalamus has important projections to orbitofrontal and prefrontal regions that also are integrally involved in the network (Adolphs, 2001).
Notably, the specific associations between local brain volumes and ToM were found only within the OI group, and did not extend to the children with TBI. The reasons for this inconsistency are unclear, but the diffuse and multiple brain insults that are often associated with TBI, and which are very heterogeneous and nonoverlapping across children with TBI, may play a dominant role in affecting social information processing, and hamper the detection of more focal brain-behavior relationships (Bigler et al., 2013). In contrast, relationships between specific brain regions and social information processing may be more apparent in children without brain injuries.
ToM was a significant predictor of rejection-victimization, consistent with the general assumption that social information processing affects social behavior. Notably, affective ToM was the only unique predictor of rejection-victimization. Children’s ability to understand others’ emotions has long been acknowledged as an important predictor of their social behavior and peer relationships, being associated with increased submissive behavior and reduced peer acceptance (Hubbard & Coie, 1994). However, relatively little research has focused specifically on children’s understanding of socially deceptive emotions. The current findings show that a poorer understanding of socially deceptive emotion, or affective ToM, is associated with increased rejection-victimization by peers.
As expected, rejection-victimization by peers was associated with multiple aspects of social adjustment, including peer acceptance, mutual friendships, social withdrawal, and general psychopathology. This finding is in keeping with an established literature showing that submissive, atypical, and withdrawn behavior is associated with rejection and victimization by peers, which in turn predicts an increased likelihood of poorer psychosocial outcomes (Rubin, Shulz Begle, & McDonald, 2012). In the current study, rejection-victimization was more strongly associated with peer acceptance and friendship than with general aspects of social adjustment. However, that result could reflect shared rater variance, because rejection-victimization, peer acceptance, and friendships were all rated by children’s classmates, whereas broader aspects of adjustment were rated by parents. The latter consideration may also help account for the lack of a significant association between rejection-victimization and parent ratings of social adjustment on the ABAS-II and BASC-2.
Comparability across different sites and scanners in the VBM analyses is a potential challenge for the current study. The sample size was too small to permit separate analyses by site. However, prior studies using VBM with larger samples have shown that multisite acquisitions result in reliable findings (Bendfeldt et al., 2012; Schnack et al., 2010). The authors worked closely with the radiology staff at each site to duplicate the imaging sequences across different platforms, and both human and machine phantoms were used to ensure reliability during the study. In the VBM analyses, moreover, site was treated as a covariate.
Sample size was a more general limitation. Unfortunately, classroom and neuroimaging data could not be obtained for all children in the larger parent study. Although children with and without classroom and neuroimaging data did not differ on demographic or injury characteristics, the restricted sample reduced the power of statistical analyses and precluded the use of more sophisticated statistical methods to test the overall model.
Another limitation is that the effects of SES on the assessed outcomes cannot be determined, because of the intrinsic relationship of SES to TBI. However, the lack of evidence for recruitment bias and the absence of group differences in SES when injury mechanism is taken into account indicate that the group difference in SES is not a function of biased sampling. A final limitation was the lack of information regarding the children’s preinjury peer relationships. We could have obtained retrospective ratings of social adjustment from parents, but the study was conducted at least 1 year postinjury, and retrospective ratings would be subject to substantial recall bias.
In the future, significant advances in our understanding of the social outcomes of pediatric TBI are likely to require a prospective, longitudinal study that uses innovative methods to examine the associations among the different levels of social outcomes across time and to assess the causal relationships among those outcomes. The ideal study would involve a large sample of children recruited from multiple sites shortly after their injuries and followed longitudinally into adulthood. This design would permit the identification of individual trajectories in social outcomes and their relationship to various risk and resilience factors that moderate those outcomes (Booth-LaForce et al., 2012; Oh et al., 2008). Friendship quality, for example, may be a critical moderator of the effects of peer rejection on social adjustment, such that good friendships buffer children against the negative effects of rejection whereas poor friendships exacerbate those effects (Laursen, Bukowski, Aunola, & Nurmi, 2007). The study would also incorporate advanced neuroimaging to better characterize the changes in the social brain network that occur following TBI and their relationship to various social outcomes. For instance, both structural and functional connectivity of the network may be altered by TBI, with likely implications for social information processing (Levin et al., 2011). The application of naturalistic social cognitive paradigms using functional MRI would provide greater insight into how changes in brain function mediate changes in social information processing following TBI (Zaki & Ochsner, 2009). This emphasis could be extended to the observation of children’s social behavior in naturalistic contexts, to provide a better appreciation of the relationship of social information processing to everyday social interaction (Hawkins, Pepler, & Craig, 2001; Pepler & Craig, 1995). The study of critical longer-term correlates of social competence, such as employment status and intimate relationships, would likely help to clarify the deficits in social adjustment and overall quality of life reported among the adult survivors of childhood TBI (Anderson, Brown, Newitt, & Hole, 2009).
Obviously, the study envisioned here would necessitate substantial fiscal resources to support an interdisciplinary approach combining expertise in social neuroscience, developmental psychology, neuropsychology, and social psychology. However, in addition to furthering scientific understanding, the study would likely have significant clinical implications by (a) identifying which children are at highest risk for social problems at what point postinjury, so that interventions can be undertaken with the most vulnerable children when they are most likely to be effective; (b) clarifying the nature of social problems after TBI, how they develop across ages and time since injury, and their neural bases, as a means of recognizing distinct trajectories of social problems and guiding interventions that are specific to core deficits; and (c) elucidating the medical and social-environmental factors that contribute to the evolution, resolution, or exacerbation of social problems after TBI, thereby facilitating the design of interventions that are most likely to facilitate children’s recovery from TBI.
Footnotes
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.
Funding
The preparation of this article was supported in part by Grant 5 R01 HD048946 from the National Institute of Child Health and Human Development to K. O. Yeates.