Mild traumatic brain injury (mTBI) is a growing public health problem, which represents more than 80% of all traumatic brain injury. However, our understanding of the underlying pathophysiology associated with mTBI is limited. A few studies using voxel-based morphometry (VBM)¹ show morphometric changes to the GM after mTBI². Spatially normalized and modulated GM density from VBM analysis is typically interpreted as GM volume with a decrease in density equated with atrophy and an increase in density equated with GM thickening^2,3. VBM is, however, sensitive not only to changes in volume but to variations in T1 relaxation⁴. In this study, we investigated alterations in GM density using simulated images and VBM followed by an in vivo study of cortical GM morphometry using GM density and cortical thickness in mTBI patients and matched controls.Simulation: Effect of altering T1 relaxation on VBM GM density was studied by using simulated images from a representative T1 weighted image (T1WI) of a normal individual (25 year old man). 40% of voxels within a frontal opercular region of interest (ROI) (total 833 voxels through 7 slices) were selected for simulation (Fig. 2 (top)). Hyperintense voxels were simulated via increasing voxel signal intensity by 5%-20% whereas hypointense lesions were simulated via decreasing intensity by 5-20% (Fig.2 (top)). Human Subjects: We studied 17 patients with mTBI within two weeks of injury (age, 34.4±9.5; 21-52 years old) and 14 age- and sex-matched normal controls (NC) (age, 33.8±9.6; 19-50 years old). MRI Acquisition: MR imaging was performed on 3T MR scanner (Skyra; Siemens). 3D MP-RAGE acquisition was performed to obtain structural T1WI (FOV = 256 × 256 mm², matrix = 256 × 256, resolution = 1 × 1 × 1 mm³, slices = 192, TR/TE/TI = 2100/3.19/900 ms, FA = 8°, and bandwidth = 260 Hz/pixel). Image Analyses: FSL-VBM⁵ carried out with FSL tools was used to assess GM density changes. All data underwent noise reduction, brain-extraction, segmentation, registration, modulation and smoothness. Regional cortical thickness was estimated using FreeSurfer^6,7 and the results were visually inspected and manually corrected by an expert neuroradiologist.In Fig.1, GM/modulated GM densities were higher in the hyperintense regions, while they were lower in the hypointense regions. We found that the cortical boundaries of the simulated lesions were almost identical as shown in Fig. 2 (bottom). Fig.2 presents the VBM results showing comparisons between NC and mTBI. We found three clusters having significantly decreased modulated GM densities in mTBI (p < 0.001; shown in red-yellow; > 350 voxels): right precuneus (cluster 1), left superior temporal (cluster 2), and left precentral regions (cluster 3). The mean values of modulated GM densities for the NC and mTBI groups were 0.50 ± 0.054 and 0.40 ± 0.051 in cluster 1, 0.56 ± 0.083 and 0.45 ± 0.03 in cluster 2, and 0.54 ± 0.054 and 0.45 ± 0.048 in cluster 3, respectively. However, we found no significant differences in cortical thickness and GM volume from FreeSurfer between NC and mTBI groups.In this study, which used both VBM and cortical thickness analyses, we demonstrate that changes in VBM results can be sensitive to any alterations in the GM composition after mTBI, as opposed to the volume of GM. We observed: 1) decreased GM/modulated GM densities in mTBI within three regions (right precuneus, left posterior superior temporal gyrus, left precentral gyrus); and 2) no significant differences in cortical thickness in these regions. Absence of cortical thickness changes in the setting of positive VBM findings, raises the possibility that differences in GM volume may not be the cause of this finding. As our simulation results demonstrate, a measured decrease in GM density can be due to changes in signal intensity within the region. Decreases in signal intensity without concomitant changes in GM volume, may arise in a variety of situations such as gliosis, resolving edema such as in the case of subacute infarction, as well as areas of hyperacute and acute hemorrhage (oxy- and deoxyhemoglobin).This study indicates that used together, VBM and cortical thickness analyses can provide greater detail regarding GM changes. Detecting and understanding GM injury after mTBI is critical for further investigating the mechanisms of tissue damage and recovery. Focal changes in GM density as detected by VBM analysis may represent areas of signal changes rather than true volume change, particularly in the setting of maintained cortical thickness.