HJAR Jul/Aug 2024

HEALTHCARE JOURNAL OF ARKANSAS I  JUL / AUG 2024 15 JAMA NETWORK OPEN ORIGINAL INVESTIGATION / PEDIATRICS all vertices in surfaces. The individual ReHo maps were smoothed by a 3D Gaussian ker- nel of 6 mm FWHM for further statistical analysis. Analyses for Seed-Based RS-FC Analysis As with ALFF and ReHo, the seed-based FC analyses were conducted on the corti- cal surface. First, temporal filtering was applied to the fMRI data, specifically in the frequency range of 0.01 to 0.1 Hz. The mean CSF signal was removed by linear regression. Then, the 3-dimensional vol- umes comprising 210 images for each participant were transformed to the sur- face. Normalization procedures, including a smoothing operation within the surface space to a 9-mm radius, were performed. For the surface mapping of the fMRI data, a weighted-mean method that uses a Gauss- ian kernel for mapping along the normal was applied. Lastly, regions of interest extraction was performed within the surface space. The bilateral DLPFC was selected as our region-of-interest (ROI) because of its involvement in brain injury. 24 APearson cor- relation coefficient was generated between the DLPFC and the whole brain. The cor- relation coefficients were transformed into Fisher z scores for statistical analysis. Statistical Analysis Demographic differences between the football and control groups were assessed by 2-sample t tests for continuous vari- ables and x 2 tests for categorical variables. Mann-Whitney U test was used for General Anxiety Disorder-7 (GAD-7) and Patient Health Questionnaire-9 (PHQ-9) due to significant results in Shapiro-Wilk normal- ity tests. Group comparison on morpho- logical metrics was performed using CAT12 and analyzed via a nonparametric permu- tation technique (5000 permutations). Age, body mass index (BMI), number of previous concussions, depression score via PHQ-9, anxiety score via GAD-7, and intracra- nial volume were included as covariates. The threshold-free cluster enhancement (TFCE) was used in the permutation test, which gives cluster-based thresholding for familywise error correction, and the level of significance was set to P  <.02 to reflect 3 morphological outcomes. For the RS-fMRI analyses, the 2-tailed, independent-samples t tests were per- formed in the DPABISurf toolkit to com- pare group differences in mean and peak ALFF and ReHo. For ALFF and ReHo, the statistical significance was set to P  <.01. We used TFCE to obtain a significant group difference. Similarly, the 2-tailed, indepen- dent-samples t tests were used to compare RS-FC strengths between groups. For all RS- fMRI analyses, age, BMI, previous concus- sions, and mean head motion were included as covariates. Multiple comparisons were corrected by TFCE, and the level of statisti- cal significance was set at 2-tailed P  <.03 to reflect bilateral DLPFC as seeds. The number of permutations was set at 5000. Data were analyzed from February to November 2023.