A conventional brain MRI resolves detail down to roughly a millimeter. An axon is about one micron across. That gap of three orders of magnitude is where most concussions live.
Most concussions come from rotation rather than a straight-line blow: the head snaps or turns, the brain lags behind the skull for a few milliseconds, and the tissue inside twists against itself. That force rarely breaks anything a structural scan can see. It strains the long wiring bundles running between brain regions.
Diffusion tensor imaging — the second study in this series on the eight imaging types used to evaluate mild traumatic brain injury — is built for that scale. It is also the most oversold.
How Does DTI See Something a Standard MRI Can't?
Water in the brain is always moving. In open fluid it moves equally in all directions. Inside an axon it cannot: the membrane and the myelin around it obstruct sideways movement, so water travels along the fiber. Picture a bundle of drinking straws — water poured in runs down them, not across. Crush the bundle, or loosen it so the straws splay, and the movement becomes less directional.
DTI is a specialized MRI sequence — same magnet, no contrast, no radiation — that measures diffusion direction in every small volume of tissue. Combined across the brain, those directions reconstruct the tracts and estimate how organized each one is.
The number reported is **fractional anisotropy**, scaled from 0 to 1. Zero means water moves equally in all directions; one means it moves along a single axis. Higher values are read as more organized tissue.
What Kind of Injury Is DTI Looking For?
Gray and white matter differ slightly in density, so under rapid rotation they decelerate at different rates and the tissue where they meet gets dragged. The long tracts crossing between hemispheres and running toward the brainstem sit where that strain concentrates. It is the same mechanism that makes [whiplash and concussion produce overlapping symptoms](post-why-whiplash-and-concussion-produce-the-same-symptoms.html) — the neck delivers the rotation, and the brain is on the end of it.
The clinical term is traumatic axonal injury. Bruggeman and colleagues collected its definitions, mechanisms, and imaging in a 2021 review (Bruggeman et al., 2021) — a narrative review, which organizes the concept without establishing how accurately any test detects it.
One framing correction. Mild traumatic brain injury is diagnosed by its characteristic symptoms and neurological effects, not by the absence of visible damage. Visible damage is simply rare here, and that rarity creates the puzzle: real impairment alongside a report saying nothing is wrong.
What Have the DTI Studies Actually Found?
Aoki and Inokuchi pooled DTI studies in a voxel-based meta-analysis and found that "fractional anisotropy values were significantly lower in mTBI patients than in control in three clusters" (Aoki & Inokuchi, 2016). The largest peaked in "the left thalamus" and extended "to the splenium of the corpus callosum and to the anterior thalamic radiation." The others involved the "left forceps minor" and the "right superior longitudinal fasciculus III."
Two limits matter, both stated by the authors. This is a group-level result: patients as a population differed from controls as a population, which does not establish that an individual can be diagnosed from their own scan. And the meta-analysis exists because prior studies "yielded inconsistent results."
Does Injury Always Lower Fractional Anisotropy?
No, and this is the finding that should make anyone cautious about a single DTI report. Eierud and colleagues found the direction of change depends on how much time has passed: "acute mTBI is associated with **elevated** anisotropy values and chronic mTBI complaints are correlated with **depressed** anisotropy" (Eierud et al., 2014).
The measurement moves in opposite directions depending on chronicity, so a scan taken days after an injury and one taken a year later are not interpretable against the same reference. That review also documents "a strong anterior-to-posterior gradient," and its authors conclude that "much more work in this area is required… to achieve clinically-relevant capabilities for diagnosis."
Can DTI Diagnose an Individual Person?
Not on current evidence. Wortzel's review of advanced neuroimaging in mTBI litigation reports that the American College of Radiology and a consensus statement hold "there remains insufficient evidence… to conclude that these advanced techniques can be used for routine clinical use at the individual patient level" (Wortzel, 2022). On DTI it names three barriers.
**Absence of normative data.** To call your value low, you need to know what it was before, or what is expected for someone like you on that scanner. Neither is usually available.
**Reliance on standard population analyses.** Those statistics compare groups, and applying a group difference to one person is a category error — an easy one to make when the picture is colorful and the patient is suffering.
**Lack of specificity in distinguishing different forms of pathology.** Reduced anisotropy is not a signature of trauma, and the measurement cannot separate trauma from other processes affecting white matter.
So Why Order It At All?
What follows is my clinical reasoning, not a research finding. I think about imaging after head injury in tiers. Tier 1 is gross structural damage — bleeds, lesions, fractures — and [conventional MRI answers that well](post-what-an-mri-of-the-craniocervical-junction-shows.html). Tier 2 is diffuse axonal injury from rotational shear, a different scale of problem needing a different sequence. When the structural scan is clean and the patient is clearly impaired, DTI is often where something appears — reduced values along the corpus callosum, internal capsule, and cervicomedullary tracts.
That is a clinical position, and I want to be precise about why. None of the studies above enrolled only patients with confirmed-normal MRI, so the confident sentence "DTI detects injury when MRI is normal" is not something this literature demonstrates. It is a rationale for ordering the study, not a proven property of it.
What DTI legitimately contributes is a data point that either converges with the rest of the evidence or does not — which is why the series runs to eight modalities. A finding that lines up with the mechanism, the exam, and the [symptoms someone is living with](condition-post-concussion-syndrome.html) is worth something. Alone, it is worth little.
Does This Help Separate a Brain Problem From a Neck Problem?
This is the part of the question I care most about, and it cuts both ways.
Most of my clinical work concerns the craniocervical junction, and a fair reading of that could be that I expect the neck to be the answer every time. I do not, and an imaging protocol that could only ever confirm one conclusion would not be worth ordering.
Tractography is a brain study. It images axonal wiring, and it has nothing to say about the ligaments at the top of your neck. That is precisely what makes it useful to me, because it is one of the modalities capable of pointing away from my own territory.
A patient with a clean craniocervical examination and a consistent anisotropy pattern through the corpus callosum is describing a different problem from a patient with a demonstrably unstable junction and unremarkable tracts. Those two people need different care, and in the second case a good deal of that care is mine while in the first case much of it is not.
Sometimes the finding is cervical. Sometimes it is genuinely in the brain, and saying so plainly is part of the job. Sometimes it is both — which, in the people I see, is the most common of the three. Running brain-only modalities alongside structural imaging of the neck is how you tell those apart rather than assuming.
Been Told There Is Nothing Wrong?
If your structural imaging came back clean and your symptoms did not, the answer is not that the scan was wrong. It answered a question about millimeters, and the injury may be at microns.
Sorting out which tier of injury explains your presentation, and being straight about what each test can and cannot prove, is the work we do at Cerebral. If you'd like a real evaluation, we're here.
References
- Aoki Y, Inokuchi R. A voxel-based meta-analysis of diffusion tensor imaging in mild traumatic brain injury. *Neuroscience & Biobehavioral Reviews*. 2016;66:119–126. https://pubmed.ncbi.nlm.nih.gov/27133211/
- Eierud C, Craddock RC, Fletcher S, et al. Neuroimaging after mild traumatic brain injury: review and meta-analysis. *NeuroImage: Clinical*. 2014;4:283–294. https://pubmed.ncbi.nlm.nih.gov/25061565/
- Bruggeman GF, Haitsma IK, Dirven CMF, et al. Traumatic axonal injury (TAI): definitions, pathophysiology and imaging — a narrative review. *Acta Neurochirurgica*. 2021;163(1):31–44. https://pubmed.ncbi.nlm.nih.gov/33006648/
- Wortzel HS. Advanced neuroimaging and mild traumatic brain injury litigation, revisited. *Journal of the American Academy of Psychiatry and the Law*. 2022;50(3):336–341. https://jaapl.org/content/50/3/336