Fronto-Parietal Brain Metabolites Following Traumatic Brain Injury-A Proton MR Spectroscopy Study

Background: Conventional structural CT and MRI may be normal after mild traumatic brain injury (mTBI), while magnetic resonance spectroscopic imaging (¹H-MRSI) can provide regional measures of brain metabolites. We examined supracallosal fronto-parietal metabolite ratios in active-duty service members with military-related TBI and non-TBI military controls, including comparisons by injury mechanism and exploratory clinical associations. Methods: The study used a retrospective cross-sectional cohort. Participants underwent 3T 2D multivoxel ¹H-MRSI covering a 6 × 8 × 1 cm³ supracallosal fronto-parietal PRESS/CSI volume of interest. Regional N-acetylaspartate (NAA), choline-containing compounds (Cho), and creatine/phosphocreatine (Cr) measures were processed with SIVIC and summarized as NAA/Cr and Cho/Cr ratios. Linear mixed-effects models were adjusted for age and sex, with Bonferroni correction for regional group comparisons. Male-only analyses were performed as a sensitivity analysis. Associations with clinical and neurocognitive measures were exploratory. Results: The final analytic cohort reported in the manuscript comprised 741 service members with TBI (34.3 ± 8.1 years old; M/F = 711/30; mean 327.4 ± 226.8 days post-injury) and 75 non-TBI controls (M/F = 53/22; mean age: 33.1 ± 10.0 years). TBI participants showed lower Cho/Cr area ratios in multiple fronto-parietal regions, including right precuneus (estimate = -0.0781, SE = 0.0176, t = -4.43, pBonf < 0.0001, Cohen's d = -0.36, 95% CI [-0.1127, -0.0435]), left superior frontal gyrus (estimate = -0.0715, SE = 0.0163, t = -4.38, pBonf < 0.0001, Cohen's d = -0.36, 95% CI [-0.1035, -0.0394], and bilateral fronto-parietal WM (right: estimate = -0.0741, SE = 0.0169, t = -4.39, pBonf < 0.0001, Cohen's d = -0.35, 95% CI [-0.1072, -0.0410]; left: estimate = -0.0656, SE = 0.0161, t = -4.08, Cohen's d = -0.33, pBonf = 0.0011, 95% CI [-0.0971, -0.0340]). Injury mechanism stratification revealed a widespread reduction in Cho/Cr area in individuals with blast-related TBI relative to unexposed non-TBI controls over the left superior frontal gyrus (estimate = -0.0941, SE = 0.0176, t = -5.35, pBonf < 0.0001, Cohen's d = -0.44, 95% CI [-0.1287, -0.0596]), right precuneus (estimate = -0.0900, SE = 0.0193, t = -4.66, Cohen's d = -0.38, pBonf = 0.0010, 95% CI [-0.1279, -0.0521]), and right fronto-parietal WM (estimate = -0.0895, SE = 0.0182, t = -4.91, Cohen's d = -0.39, pBonf < 0.0001, 95% CI [-0.1253, -0.0537]). Twelve regional differences remained significant in the primary analysis, and the principal pattern was substantially preserved in the male-only sensitivity analysis. Blast-related and multiple-blast comparisons also showed regional Cho/Cr differences; individuals with multiple blast exposures mainly drove Cho/Cr reduction. Exploratory clinical analyses found inverse associations between Cho/Cr and Post-Traumatic Stress Checklist-Civilian version (PCL-C) and Neurobehavioral Symptom Inventory (NSI) scores, and positive associations between selected NAA/Cr measures and cognitive performance. Specifically, higher NAA/Cr area over the left caudal anterior cingulate correlates positively with verbal fluency performance (r = 0.358, p < 0.0001). Higher NAA/Cr peak heights in the left caudal middle frontal cortex correlated with better response inhibition accuracy (r = 0.278, p = 0.0003). Concurrently, lower Cho/Cr values over default mode network hubs-specifically the left caudal anterior cingulate gyrus (estimate = -0.00144, t = -4.02, uncorrected p < 0.0001, 95% CI [-0.00215, -0.00074]), left posterior cingulate gyrus (estimate = -0.00115, t = -3.76, uncorrected p = 0.00019, 95% CI [-0.00175, -0.00055], and the left precuneus gyrus (estimate = -0.00120, t = -3.62, p = 0.00033, 95% CI [-0.00186, -0.00055])-associated significantly with greater self-reported post-traumatic stress severity on the PCL-C. In addition, Cho/Cr in the left fronto-parietal WM was inversely associated with PCL-C (estimate = -0.00115, t = -3.87, p = 0.00012, 95% CI [-0.00174, -0.00057]) and NSI (estimate = -0.00117, SE = 0.00030, t = -3.92, p < 0.0001, 95% CI [-0.00176, -0.00058]). Injury chronicity did not correlate with metabolic alterations. Conclusions: ¹H-MRSI identified regional metabolite-ratio differences in this military TBI cohort. The findings support further investigation of regional Cho/Cr and NAA/Cr measures as correlates of clinical heterogeneity after TBI. The widespread depression of Cho/Cr suggests persistent membrane depletion and/or chronic axonal degeneration, particularly in TBI individuals with multiple blast exposures. Although the cross-sectional design does not establish causality, diagnostic sensitivity, biomechanical strain patterns, or longitudinal recovery, the results imply the utility of MRSI as a robust translational tool for phenotyping chronic military brain injury and objectively monitoring long-term neural recovery.