Diffusion tensor imaging (DTI) is now widely used in both clinical and experimental research for studying pathology related to traumatic brain injury. However, studies report a wide range of DTI indices that are not easily ascribed to post-injury time-point, injury severity or developmental stage, likely indicating a complex underlying pathology. In order to help interpret these complex changes, we obtained DTI data before and after TBI using the well-known, clinically relevant rodent controlled cortical impact (CCI) model of TBI. We used the known post-injury changes in axonal and dendritic pathology^1,2 that are associated with this model, as well as additional dye-tract tracing and immunostaining (IHC) experiments to interpret the changes in the scalar indices.DTI data were acquired on a 7Tesla Bruker MRI from adult rats (n=17) under isoflurane sedation before and at 7 and 28 days after unilateral CCI injury. A 4-shot, spin echo, echo planar imaging sequence (6250/32ms TR/TE) was used to acquire diffusion-weighted images with 30 different directionally-encoded gradients, a b value of 1000 s/mm2, using Δ=20ms and δ=3ms, and five images with a b value of 0 s/mm2. All images were acquired with a 128-read and 128-phase-encoding matrix within a 35mm2 field-of-view and 25x0.75mm contiguous, coronal slices. Anatomical RARE data were acquired with 50x0.5mm slices, the same matrix, a TR/TE 5000/60ms, RARE factor of 8 and 2 averages. Data were fit to the tensor model to generate scalar maps of fractional anisotropy (FA), axial and radial diffusivity (AD, RD), mean diffusivity (MD) and tensor shape as the mode of the tensor (MO). Scalar data were analyzed using tract based spatial statistics³ and tested for group differences using an unpaired t-test against pre-injury data and corrected for multiple comparisons using cluster-based thresholding (c=1.7,family-wise error rate of P<0.05). Contusion volume at 28days post-injury was estimated from the anatomical data and used as a covariate of no interest within the statistical design. Bayesian estimation of diffusion parameters was used to model 2 fibers per voxel⁴ to test for gross differences in crossing fibers. Fiber tract density (FTD) data were generated from pre-injury and post-injury 28d data by deterministic fiber-tracking⁵. Interhemispheric retrograde tract-tracing was performed in 3 rats by injecting fluorogold and cholera toxin B either side of the main injury site and measuring the number of labelled cells in the contralateral cortex. Additional brains (n=6) were also immunostained by gliosis, neurofilaments and myelin basic protein using standard IHC.There were marked reductions in FA at 1wk in corpus callosum that were underpinned by increases in RD (P<0.05) that was associated with myelin breakdown on IHC. FA was negatively associated with injury severity at this post-injury time (r=-0.78;P<0.001). Similar regional reductions in FA were found at 4wks post-injury but underpinned by increases in AD, RD and MD, corresponding to neurofilament and myelin loss on IHC and an almost 50% reduction in interhemispherically, dye-labelled cells (P<0.05). FA and AD/RD/MD were negatively and positively associated with injury severity, respectively (r=-0.67;0.43-0.48; P<0.004) but unlike at 1wk, this was bilateral across the corpus callosum, possibly reflecting on-going degeneration. We also found increases in FA at both 1 and 4wks post-injury, bilaterally in caudate thalamus, internal capsule, peduncle and at 4wks only in the posterior callosum. Increased AD was the clear driver of this at 1wk (P<0.05), although it was variable at 4wks (P>0.05). Contralateral internal capsule FA was positively associated with injury severity at 4weeks (r=0.78;P<0.001) which contrasts to 1wk when RD and MD were negatively associated with severity only in this brain region (r=-50, -0.52;P<0.004). We found that increased gliosis was not associated with increased FA in the corticospinal tract. Similarly, a reduction in crossing fibers by degeneration (resulting in increased FA through partial volume averaging, and indicated by increases in MO) did not occur in this region, although MO was increased around the primary injury site. However, in addition to tractography-generated FTD maps confirming reduced interhemispheric fibers shown by dye-tract-tracing (P<0.05), increases in FTD were found limited to the same subcortical regions where FA was increased (P<0.05). Fiber tract length analysis showed an increase in mean length in this region compared to preinjury (6.24±0.24 versus 5.21±0.18mm, respectively, P<0.01) while ipsilaterally it was reduced compared to preinjury (3.58±0.13 versus 5.07±0.22mm, respectively, P<0.001; 2-tailed t test).In addition to the cellular changes underpinning the the DTI scalars after TBI, the most novel aspect of this study is the increased FA within subcortical regions that we have shown to be unrelated to gliosis and fiber tract degeneration, but positively associated with increased fiber density and length.