What Does Traumatic Brain Injury Look Like on MRI?

Traumatic brain injury on MRI can look like anything from large, bright patches of bleeding and swelling to subtly disrupted nerve fibers invisible on routine scans. The appearance depends heavily on injury severity, how much time has passed since the trauma, and which MRI sequences the radiologist uses. A severe TBI might show obvious hemorrhages, contusions, and swelling on standard imaging, while a mild concussion can produce a scan that looks perfectly normal to the naked eye, even when the patient is clearly symptomatic. Understanding the range of what MRI reveals, and what it misses, matters for patients and families navigating a diagnosis.

Why CT Comes First and Where MRI Takes Over

If you arrive at an emergency department after a head injury, a CT scan is almost always the first imaging study ordered. CT is fast, widely available, and excellent at spotting the things that need immediate surgical attention: skull fractures, large bleeds, and dangerous shifts in brain tissue. MRI, by contrast, takes longer and requires the patient to lie still inside the scanner, which can be impractical or even unsafe in someone who is unstable. That speed difference is why CT remains the go-to tool in the first hours after injury.

Where MRI earns its place is in sensitivity. A systematic review comparing the two modalities confirmed that while CT is essential for rapidly identifying life-threatening hemorrhage and fractures, MRI is far better at detecting subtle lesions, tiny bleeds, and a pattern of injury called diffuse axonal injury that CT routinely misses.¹ A separate meta-analysis reached a similar conclusion, noting that MRI’s strength lies primarily in post-acute evaluation and prognosis rather than in guiding emergency decisions.² In practice, many patients get both: CT in the ER and MRI days or weeks later to characterize the full extent of injury.

What Standard MRI Sequences Show

On a conventional MRI, the most recognizable TBI findings include contusions, hemorrhages, and edema. Contusions are bruised patches of brain tissue, usually at the surface, that appear as areas of abnormal signal on standard sequences. They tend to cluster along the undersides of the frontal and temporal lobes because those areas sit against bony ridges inside the skull. On T2-weighted and FLAIR sequences, contusions show up as bright areas surrounded by swelling. If a contusion involves bleeding, it will appear darker or brighter depending on the age of the blood, because hemoglobin changes its magnetic properties as it breaks down over hours and days.

Extra-axial collections, meaning blood that pools between the brain and skull, are another common finding. Epidural hematomas (between the skull and the dura), subdural hematomas (between the dura and the brain surface), and subarachnoid hemorrhage each have characteristic shapes and locations on MRI. The signal intensity of these collections changes predictably over time: fresh blood looks different from blood that has been sitting for a week, which helps radiologists date when the bleeding occurred.

Traumatic meningeal enhancement is a subtler finding. After contrast dye is injected, inflamed meninges (the membranes covering the brain) can light up on certain sequences. Research has shown that a post-contrast FLAIR sequence catches this enhancement more reliably than a standard post-contrast T1 sequence; in one study, more than a third of patients with meningeal enhancement visible on FLAIR were scored as negative on T1 imaging.³ This is the kind of finding that might go unnoticed if the radiologist is not specifically looking for it using the right sequence.

Diffuse Axonal Injury and the Microbleed Signature

Diffuse axonal injury, often abbreviated DAI, is one of the most consequential TBI findings and one that conventional CT almost always misses. DAI happens when the brain’s long nerve fibers are stretched and torn by rotational forces during the injury. On MRI, it typically appears as small, scattered lesions in the white matter, the corpus callosum (the bridge between the two brain hemispheres), and the brainstem. These lesions are often tiny, just a few millimeters across, and many of them involve microscopic hemorrhages.

Susceptibility-weighted imaging, or SWI, is particularly good at revealing these tiny bleeds. SWI exploits the magnetic properties of blood breakdown products to make even minuscule hemorrhages pop out as dark spots against the surrounding tissue. Standard MRI sequences frequently miss them entirely.⁴ The number and location of these microbleeds carry prognostic weight. Research on DAI grading found that hemorrhagic lesions in specific brainstem structures, particularly the substantia nigra and the central part of the midbrain, were independently associated with worse long-term outcomes. The number of lesions in those areas and the patient’s age were the only factors that independently predicted poor recovery in a combined statistical model.⁵

Not every dark spot on SWI means trauma, though. Microbleeds can also result from high blood pressure, cerebral amyloid angiopathy (a condition in which a protein builds up in blood vessel walls), and several other causes. Telling traumatic microbleeds apart from other types depends on where they appear, what other MRI findings surround them, and what the patient’s medical history looks like.⁶ Traumatic microbleeds tend to cluster at the gray-white matter junction and in the corpus callosum, while hypertension-related ones favor deeper brain structures.

When the MRI Looks Normal but the Patient Does Not Feel Normal

This is one of the most frustrating aspects of mild TBI and concussion. A patient can have persistent headaches, difficulty concentrating, mood changes, and fatigue, yet a standard MRI comes back clean. In a study of 127 patients with post-concussion syndrome, only about 3% had clearly abnormal structural findings like microhemorrhages or areas of tissue softening. White matter hyperintensities (small bright spots commonly seen in healthy people too) appeared in about 22% of patients, but at roughly the same rate as in age-matched controls without any head injury.⁷

The disconnect makes sense once you understand that conventional MRI is looking at brain structure at a fairly coarse level. It can see a bruise or a bleed, but it cannot see a nerve fiber that has been stretched and is conducting signals more slowly, or a metabolic shift in brain chemistry. That is where advanced MRI techniques come in, and they reveal a very different picture.

Diffusion Tensor Imaging and White Matter Tract Damage

Diffusion tensor imaging, or DTI, measures how water molecules move along the brain’s white matter tracts. In healthy nerve fibers, water tends to flow in one direction along the fiber, like water in a garden hose. When fibers are damaged, water starts leaking sideways, and DTI can measure that change. The key metric is fractional anisotropy, a number that reflects how directional the water flow is. Damaged tracts generally show lower fractional anisotropy.

In a large multicenter study, patients with mild TBI had measurably disrupted white matter compared to uninjured controls.⁸ Another longitudinal study found that people with mild TBI showed abnormal diffusion values in several white matter regions, including the splenium of the corpus callosum and the optic radiation, and that these changes were linked to neuropsychological performance.⁹ One complicating detail is that the direction of the change is not always consistent. At least one prospective study found that fractional anisotropy was actually higher (not lower) in the corpus callosum and several left-hemisphere tracts in the early phase after mild TBI, a pattern thought to reflect acute tissue swelling that compresses nerve fibers and temporarily constrains water flow.¹⁰

Newer diffusion techniques are pushing beyond DTI’s limitations. A method called neurite orientation dispersion and density imaging, or NODDI, provides a more detailed picture by separating signals from nerve fibers, surrounding fluid, and everything else. A longitudinal study found that early changes after mild TBI on standard DTI corresponded to elevated free water (edema) on NODDI, while longer-term changes showed declining neurite density, suggesting ongoing nerve fiber degeneration. The researchers concluded that NODDI was more sensitive and specific than DTI alone for tracking white matter changes after concussion.¹¹ A separate study of concussed athletes found that white matter abnormalities persisted even after athletes were medically cleared to return to play, raising the question of whether clinical recovery outpaces structural recovery.¹²

What Brain Chemistry and Blood Flow Reveal

Magnetic resonance spectroscopy, or MRS, is essentially MRI applied to brain chemistry rather than brain structure. Instead of producing a picture, it produces a graph showing the concentrations of various metabolites in a chosen brain region. After TBI, the most consistent finding is a drop in a chemical called N-acetylaspartate, or NAA, which is a marker of healthy, functioning neurons. A meta-analysis spanning studies of both overt brain injuries and subconcussive hits found a reliably lower NAA in injured subjects compared to controls. The same analysis found elevated choline, which reflects cell membrane breakdown or inflammation.¹³ NAA reductions have been shown to predict cognitive outcomes and to appear even in patients whose conventional MRI scans look normal.¹⁴ That sensitivity makes MRS valuable precisely in the cases where standard imaging falls short.¹⁵

Perfusion imaging looks at blood flow through the brain. One technique, arterial spin labeling, or ASL, uses magnetically labeled blood as its own tracer, so no dye injection is needed. After mild TBI, researchers have found reduced blood flow in multiple regions, particularly the frontal and occipital lobes, even when conventional MRI showed no structural abnormalities.¹⁶ A review of the ASL literature found that patterns of reduced blood flow correlate with worse cognitive outcomes and that region-specific perfusion changes map onto particular symptom clusters, including mood and emotional difficulties.¹⁷

Functional Connectivity Changes on Resting-State fMRI

Resting-state functional MRI measures how different brain regions synchronize their activity when a person is lying quietly in the scanner, not performing any task. Several brain networks become disrupted after TBI. A study of mild TBI patients found reduced connectivity in the default mode network, the central executive network, and motor-auditory areas. The degree of disruption tracked with symptom severity: patients with worse post-concussive symptoms had more reduced connectivity. At a six-month follow-up, connectivity had increased and symptoms had improved in parallel.¹⁸

In adolescents, the picture is slightly different. A study comparing concussed teenagers to teens with non-brain injuries (like broken bones) found increased connectivity within certain networks after concussion, not decreased. Specifically, the salience network and the central executive network showed heightened within-network connectivity in the concussion group, while between-network connectivity between the default mode and executive networks was reduced.¹⁹ These patterns may reflect the injured brain’s attempt to compensate, though that interpretation is still being worked out.

Patients who develop persistent post-concussion syndrome show specific patterns of network disruption that differ from those who recover quickly. One study found that in the post-concussion syndrome group, temporal and thalamic regions were most affected in the early weeks, while frontal regions became the primary sites of disruption months later, suggesting the injury’s footprint evolves over time.²⁰

Progressive Brain Atrophy After TBI

TBI does not just cause damage at the moment of impact. Months and years afterward, the brain can continue to lose volume in a way that is visible on serial MRI scans. In patients with moderate-to-severe TBI, volumetric MRI showed significant increases in cerebrospinal fluid space and decreases in hippocampal volume between roughly four months and two-and-a-half years after injury, a rate of shrinkage well beyond what normal aging would explain.²¹ A larger study measuring annualized brain volume change found that TBI patients lost about 1.5% of total brain volume per year, compared to about 0.2% in healthy controls.²²

The pattern of this atrophy differs from other neurodegenerative diseases. Compared to Alzheimer’s disease, TBI-related atrophy tends to hit white matter and structures like the corticospinal tracts and corpus callosum harder, while Alzheimer’s predominantly affects the cortex. Deep white matter loss appears to be a TBI-specific signature.²³ Recognizing these distinct patterns can help radiologists differentiate between post-traumatic neurodegeneration and early dementia when a patient with a history of head injury presents with cognitive decline years later.

The Glymphatic System and Waste Clearance

A more recent area of research uses MRI to study the brain’s waste-clearance system, called the glymphatic system. This network of fluid-filled channels flushes metabolic waste out of the brain, and disruption of this system after TBI may contribute to ongoing damage. Researchers have developed a way to estimate glymphatic function from diffusion MRI data. Patients with TBI consistently show lower glymphatic activity than healthy controls. One study found the index measuring this activity was significantly reduced in TBI patients, with worse suppression in patients who had visible structural damage.²⁴ Even in mild TBI, reduced glymphatic function has been linked to greater white matter damage and cognitive impairment.²⁵ This line of research is still young, but it offers a potential explanation for why some patients continue to deteriorate months after injury: if the brain cannot efficiently clear its own waste, damaging byproducts may accumulate.

Imaging Chronic Traumatic Encephalopathy

Chronic traumatic encephalopathy, or CTE, is the neurodegenerative condition linked to repeated head impacts, most famously in contact-sport athletes. As of now, CTE can only be definitively diagnosed through examination of brain tissue after death. But there is active work trying to find MRI biomarkers that could flag it during life. Structural and functional imaging studies in people with histories of repetitive head trauma have found patterns of frontotemporal and medial temporal lobe degeneration, along with increased rates of a structural variant called cavum septum pellucidum, an abnormal gap between membranes deep in the brain.²⁶ PET scans targeting tau protein (the hallmark of CTE pathology) have been explored, but an optimal tracer that binds well to the specific type of tau tangles in CTE has not yet been identified.²⁷ The honest assessment is that neuroimaging for CTE is promising but unproven. Small sample sizes and the inability to confirm whether living participants actually have CTE limit what can be concluded from existing studies.

Pediatric Considerations

Children’s brains are still developing, which makes interpreting TBI imaging more complicated. Myelination, the process by which nerve fibers gain their insulating sheath, is ongoing throughout childhood and adolescence. A longitudinal study of children with mild TBI and uninjured controls found that cortical myelination increased significantly over a four-month window in both groups, with no group-level difference that survived statistical correction.²⁸ That result is cautiously reassuring for mild injuries, but it also highlights how normal developmental change can mask or mimic injury-related change, making pediatric TBI imaging inherently harder to interpret.

There is also a practical challenge: younger children move more in the scanner. A study of children aged 8 to 16 found that younger age and male sex were both associated with significantly more motion artifacts on both structural and diffusion-weighted images. Children with any recent injury, whether to the brain or to an arm or leg, had more motion-corrupted data than typically developing children, meaning that the population you most want to scan is the one most likely to produce poor-quality images. Sedation can solve the motion problem but adds its own risks, which is why clinicians weigh the decision carefully.

Artificial Intelligence and Automated Detection

Machine-learning tools are being developed to assist radiologists in reading TBI scans. Deep learning models can be trained on thousands of images to spot subtle features that a human eye might miss, including tiny microbleeds, subtle swelling, or early signs of structural damage. A scoping review of AI applications in TBI found that these tools are being developed to classify injury severity from imaging data and to improve both speed and accuracy of diagnosis.²⁹ The technology is not yet part of standard clinical workflows for TBI, but it represents one of the more plausible near-term advances. Given that many of the most meaningful TBI findings involve detecting dozens of tiny lesions scattered across the brain, automated counting and localization tools could reduce both missed findings and the time a radiologist spends on each case.