Abstract: Incidence of traumatic brain injuries (TBIs) among children in the United States is high, and may result in both changes in brain structure and cognition.  While cortical thinning among adult patients experiencing TBIs is well documented, the effect of pediatric TBIs on cortical thickness has not been as extensively studied.  Advanced Normalization Tools (ANTS) software was used to compare cortical thickness of ten pediatric TBI patients to ten orthopedically injuried (OI) matched controls.  There were no significant differences in cortical thickness between the two groups (p=0.05).  The null effect may be the result of extraneous variables, and the subject warrants further research.  
Introduction
    Traumatic brain injuries (TBIs) are prevalent in the United States, and are a leading cause of both mortality and emergency room visits among children.  On average, 500,000 children under the age of 14 suffer a TBI each year (Wilde et al, 2012).  TBIs may prove to have serious consequences, as they may be associated with neuroanatomical changes for the patient.  For example, studies have demonstrated volumetric reductions in the hippocampus and corpus callosum for pediatric TBI subjects compared to healthy controls (Dennis et al, 2017).  Evidence also exists for expansion of the lateral ventricles and volumetric reductions throughout the brain in regions such as the cingulate gyrus, thalamus, and fusiform gyrus (Dennis et al, 2016). 
    Various studies have also connected TBIs to reductions in cortical thickness.  Among adult populations, reductions in cortical thickness have been demonstrated following injury among military veterans (Michael et al, 2015), patients at risk for neurodegenerative diseases (Hayes et al, 2017), and motorists experiencing an accident (Wang, 2015).  These studies, however, all feature populations aged 18 years or older.  There are comparatively fewer studies examining cortical thickness among pediatric populations, and the relationship between pediatric TBI and cortical thickness remains relatively unknown (Urban et al, 2017).  Difficulties arise when examining pediatric cortical thickness, and results are obscured in some cases.  Some cortical thinning is normal as a product of childhood development and maturation (Urban et al, 2017).  Thus it may be difficult to separate thinning that occurs naturally from thinning that may be attributed to brain insult.  While there are studies demonstrating significant cortical thinning of pediatric TBI patients compared to normally developing matched controls (Merkley et al, 2008), the overall body of research remains inconclusive.  Additional studies are needed to corroborate the reported results of cortical thinning among pediatric TBI patients. 
    Although potentially problematic to analyze, the cortical thinning among pediatric TBI patients merits additional research.  Pediatric TBI is often connected with distressing symptoms such as problems in emotional and behavioral control (Wilde et al, 2012).  A better understanding of structural changes in the brain following pediatric TBI may aide in the diagnosis and treatment of patients.  The purpose of this study, therefore, is to further examine the relationship between pediatric TBI and cortical thickness.  It is hypothesized that significant differences in cortical thickness will be observed in accordance with available research, with pediatric TBI participants having thinner cortices than their matched control counterparts.    
Methods
Participants
    The present study analyzed data from the Social Outcome of Brain Injury in Kids (SOBIK) data set.  The SOBIK study examined social outcomes in children following TBI, and Magnetic Resonance images (MRIs) of these children were obtained at multiple sites.  Participants included ten brain injured, school aged children and ten matched orthopedic injured (OI) controls, none of whom had experienced loss of consciousness or had any known head injury.  Both TBI and OI children were injured and hospitalized.    
MRI Acquisition
    T1-weighted images were acquired with an 8-channel head coil, using a 1.5 T GE Signa Excite scanner.  Slices were `acquired after applying a 3D magnetization-prepared rapid gradient echo pulse sequence.  The 166 contiguous sagittal slices were obtained, with a slice thickness set to 1.2 mm.  Pulse echoes had a flip angle of 8ยบ, repetition time of 8.9 ms, echo time of 3.8 ms, field of view of 192 mm, acquisition matrix 512 x 512 mm, and in-pane resolution 0.47 x 0.47 mm.  A total of 166 contiguous slices were acquired with slice thickness of 1.2 mm.  Images were processed and analyzed at the Brigham Young University Fulton Supercomputing Lab.  The lab provides 21, 552 CPU cores across 972 compute nodes, and all nodes run Red Hat Enterprise Linux 6.6.  
MRI Preprocessing
    All images were preprocessed before analysis.  To standardize the images, they were converted from a DICOM format to a NIFTI format using the dcm2niix program.  Version  v1.0.20170821 of the dcm2niix program was downloaded from https://github.com/rordenlab/dcm2niix.  Options were included in the dcm2niix code to crop (-x) and compress (-z) the images.  This program automatically removed participant identifying information.