Diffuse axonal injury (DAI) is observed commonly in traumatically brain injured humans. However, traditional histologic methods have proven of limited use in identifying reactive axonal change early (< 12 h) in the posttraumatic course. Recently, we have reported, in both humans and animals, that antibodies targeting neurofilament subunits are useful in the light microscopic recognition of early reactive change. In the present study, we extend our previous efforts in humans by analyzing the progression of traumatic brain injury (TBI)-induced axonal change at the ultrastructural level. This effort was initiated to follow the subcellular progression of reactive axonal change in humans and to determine whether this progression parallels that described in animals. Two commercially prepared antibodies were used to recognize reactive axonal change in patients surviving from 6 to 88 h. The NR4 antibody was used to target the light neurofilament subunit (NF-L), and the SMI32 antibody was used to target the heavy neurofilament subunit (NF-H). Plastic-embedded tissue sections were screened for evidence of reactive axonal change, and once identified, this reactive change was analyzed at the ultrastructural level. At 6 h survival, focally enlarged, immunoreactive axons with axolemmal infolding or disordered neurofilaments were seen within fields of axons exhibiting no apparent abnormality. By 12 h, some axons exhibited continued neurofilamentous misalignment, pronounced immunoreactivity, vacuolization, and, occasionally, disconnection. At later stages, specifically 30 and 60 h survival, further accumulation of neurofilaments and organelles had led to the further expansion of the axis cylinder, and clearly disconnected reactive swellings were recognized. These contained a dense core of disordered immunoreactive neurofilaments partially encompassed by a cap of less densely aggregated organelles. At 88 h, the reactive axons were larger and elongated, consistent with the continued delivery of organelles by axoplasmic transport. At the later time points, considerable heterogeneity was observed, with focally enlarged disconnected axons being observed in relation to axons showing less advanced reactive change. Our findings suggest that neurofilamentous disruption is a pivotal event in axonal injury.