Normative Gray Matter Stiffness Gradients in the Human Brain Predict Patterns of Cortical Injury after Concussion

Finite-element modeling and in vivo MR strain mapping show that mechanical strain concentrates in the cortex, and late-life neurodegenerative sequelae of traumatic brain injury (TBI) are predominantly gray matter disorders. Nonetheless, in vivo evidence of acute gray matter damage after mild TBI (mTBI) has remained elusive. The empirical gap derives from a limitation of conventional diffusion tensor metrics, which are blind to the cortex's isotropic yet mechanically relevant solid-phase matrix of soma (glial and neural), dendrites, and extracellular scaffold. Here, we leveraged constrained spherical deconvolution (CSD)-derived "total" apparent fiber density (AFD) to index this solid-phase microarchitecture to test two predictions: (1) regional AFD covaries with magnetic resonance elastography (MRE)-derived cortical stiffness, and (2) AFD can detect gray matter injury that tensor metrics miss. We tested the first hypothesis by relating AFD from 349 healthy adults who underwent diffusion MRI to measures of shear modulus from an independent cohort of 59 healthy adults scanned with MRE. The regional distribution of AFD explained 74% of the variance in MRE-measured shear stiffness, indicating AFD is strongly coupled to the microarchitectural features that influence tissue rigidity. We then tested the clinical utility of AFD in three cross-sectional mTBI cohorts-acute (∼72 h), subacute (2 weeks to 90 days), and chronic (>90 days)-each compared with age- and sex-matched controls. Effect sizes were thresholded using Cohen's d; parcels or tracts with |d| ≥ 2.0 were chosen to isolate effects that are both statistically extreme and robust to distributional effects and technical noise. Using those criteria, 11 cortical parcels in the acute cohort showed decreased AFD. This expanded to 116 parcels in the subacute group and 106 parcels in the chronic cohort; fractional anisotropy detected no parcels, and mean diffusivity flagged only 7-9 parcels. MRE-based stiffness estimates in healthy controls further stratified the observed abnormalities: compliant cortex (∼1.6 kPa) showed AFD gains during recovery; by contrast, the stiffest cortex (∼3.0 kPa) showed persistent decreases, with baseline modulus accounting for >50% of variance in ΔAFD. Across parcels, baseline stiffness from healthy controls predicted the magnitude of AFD change in both the subacute and chronic cohorts: stiffer cortex showed larger AFD decreases; less stiff cortex showed AFD increases (Spearman ρ ≈ -0.72 to -0.74, p < 0.001). AFD also revealed robust abnormalities in 12 major white matter tracts across all mTBI cohorts, outperforming diffusion tensor metrics. Because MRE and diffusion MRI were acquired in independent cohorts, these findings should be interpreted as showing that normative region-wise stiffness gradients predict the direction and magnitude of postinjury AFD alterations. This is the first in vivo evidence that joint diffusion MRI and MRE analysis sharpens mechanistic interpretations of gray matter microarchitecture and detects gray matter disruption across the mTBI timeline.