Can You See Neuropathy on an MRI Scan?

MRI can reveal many types of neuropathy, though what the scan actually shows is not the nerve pain or numbness itself but the structural damage behind it. Swollen nerves, abnormal signal patterns, and wasting in the muscles those nerves supply all show up on modern MRI sequences. The catch is that not every neuropathy leaves a visible footprint, and the type of MRI protocol matters enormously. A standard brain or spine MRI ordered for back pain, for example, is not designed to evaluate the tiny peripheral nerves in your leg or wrist. Specialized protocols known as MR neurography have pushed the boundaries of what clinicians can see, yet even these have blind spots worth understanding.

What MRI Actually Shows in a Damaged Nerve

Healthy peripheral nerves have a predictable look on MRI: they are small, uniform in thickness, and relatively dark on most sequences. When neuropathy sets in, two changes tend to appear first. The nerve brightens on fluid-sensitive sequences (called T2-weighted images), and it gets physically larger. In diabetic neuropathy, for instance, high-resolution 3-T MRI can pick up abnormal T2 brightness and swelling of individual nerve bundles (fascicles) during the acute and subacute stages, while chronically damaged nerves look shrunken and infiltrated with fat.¹ Those two markers, brightness and size, form the backbone of MRI-based neuropathy diagnosis across many conditions.

Nerve enlargement is measurable. In a study comparing patients with the inherited neuropathy CMT1A and the inflammatory condition CIDP against healthy controls, the median cross-sectional area of the sciatic nerve was roughly three times normal in CMT1A and about one-and-a-half times normal in CIDP.² That kind of enlargement is obvious on a properly acquired scan, and it can even help distinguish one neuropathy from another when the clinical picture is unclear.

Muscle Changes as an Indirect Clue

Sometimes the nerve itself is hard to image directly, but the muscles it feeds tell the story. When a nerve stops sending signals to a muscle, that muscle goes through a predictable sequence of changes visible on MRI. In the early weeks after denervation, the muscle lights up on fluid-sensitive sequences while still looking normal on standard T1-weighted images. Over months, the muscle shrinks and fat replaces the muscle fibers, producing bright T1 signal and obvious volume loss.³ Chronically denervated muscles show conspicuous fatty replacement and marked atrophy.⁴

This pattern is clinically useful. If your doctor suspects nerve damage but cannot pinpoint where, seeing which muscles have started to waste on MRI can localize the problem to a specific nerve or nerve root. In inherited neuropathies like Charcot-Marie-Tooth disease, MRI of the lower legs reveals characteristic patterns of fatty muscle replacement that can even help narrow down the genetic subtype. Patients with certain CMT mutations, for example, show the most severe fatty infiltration in the calf muscles, particularly the soleus, and the damage is typically worse further down the leg.⁵

Carpal Tunnel and Ulnar Neuropathy

The two most common entrapment neuropathies, carpal tunnel syndrome and ulnar neuropathy at the elbow, are among the best-studied conditions on MRI. In carpal tunnel syndrome, the median nerve swells at and just before the tunnel entrance. One study found that using a nerve cross-sectional area cutoff could diagnose carpal tunnel syndrome with perfect sensitivity and near-perfect specificity.⁶ Advanced diffusion-based sequences pushed accuracy even higher, with one study reporting that a specific diffusion metric could identify carpal tunnel syndrome with sensitivity above 96% and specificity above 86%.⁷

Ulnar neuropathy at the elbow follows a similar logic. The nerve brightens and thickens at the cubital tunnel. MR neurography has shown excellent ability to distinguish ulnar neuropathy from a normal finding, and nerve caliber measurements can separate severe cases from mild ones.⁸ For many patients with entrapment neuropathies, MRI does not just confirm that something is wrong; it shows exactly where the compression is occurring, which can guide surgical planning.

Diabetic Neuropathy

Diabetic peripheral neuropathy is one of the most common neuropathies worldwide, and MR neurography has proven capable of detecting it at surprisingly early stages. A study using specialized sequences found that nerve lesion burden was significantly elevated even in people with mild-to-moderate diabetic neuropathy, and the damage was concentrated at thigh level rather than the feet, where symptoms usually appear first.⁹ That finding is noteworthy because it challenges the traditional understanding that diabetic nerve damage starts in the longest nerves at the toes and works its way up. MRI is revealing a more complex picture.

Researchers have also shown that the amount of T2-bright nerve lesion on MRI correlates with clinical severity scores and with nerve conduction slowing. Patients with painful diabetic neuropathy tend to have a larger percentage of damaged nerve volume than those with painless neuropathy.¹⁰ In practical terms, this means MRI can do more than simply confirm the presence of diabetic neuropathy; it can give a rough gauge of how extensive the damage is and even help explain why some patients have more pain than others. Quantitative T2 measurements of the tibial nerve have been proposed as a potential noninvasive biomarker for diabetic neuropathy diagnosis.¹¹

Inflammatory and Inherited Neuropathies

CIDP, a chronic autoimmune condition that attacks nerve insulation, produces some of the most dramatic findings on MRI. The nerve roots and plexuses (the large nerve bundles near the spine) become diffusely swollen and bright. In one study, about two-thirds of CIDP patients showed nerve enlargement and abnormal signal on fluid-sensitive sequences in the brachial and lumbar plexuses.¹² Quantitative measurements confirm that nerve roots in CIDP are larger, brighter, and structurally more disorganized than those in healthy people.¹³

In inherited neuropathies like CMT, whole-body MR neurography has shown nerve thickening and bright signal involving both the arm and leg nerve plexuses, with the most pronounced thickening in the demyelinating subtypes (CMT type 1).¹⁴ Diffusion tensor imaging of the sciatic nerve in CMT1A patients shows structural disorganization that correlates closely with nerve conduction slowing.¹⁵ For clinicians trying to distinguish CIDP (treatable with immunotherapy) from CMT (genetic and currently without a cure), the pattern and degree of nerve enlargement on MRI can be a meaningful additional data point alongside nerve conduction studies and genetic testing.

The Small-Fiber Blind Spot

Here is where expectations often collide with reality. Small-fiber neuropathy, which causes burning pain, tingling, and autonomic symptoms like abnormal sweating, affects the thinnest nerve fibers that MRI simply cannot resolve. Standard nerve conduction studies also miss these fibers because they are too small and slow to register on electrical testing. The established diagnostic tool for small-fiber neuropathy is a skin biopsy, which counts the tiny nerve endings in a small punch of skin. MRI, no matter how advanced, is not a substitute for this.

This distinction matters because small-fiber neuropathy is increasingly recognized as a common condition, and patients often arrive at an MRI expecting it to explain their symptoms. If your neuropathy is purely small-fiber, a normal MRI result does not mean nothing is wrong. It means the type of nerve damage you have lives below the resolution of the scanner.

Diffusion Tensor Imaging and Quantitative MRI

Beyond simply looking at whether a nerve appears bright or swollen, newer MRI techniques can probe nerve microstructure. Diffusion tensor imaging (DTI) measures how freely water molecules move within tissue. In a healthy nerve, water moves preferentially along the length of the nerve fibers, producing a high “fractional anisotropy” (FA) value. When nerve fibers are damaged, water movement becomes more random and FA drops. Damaged sciatic nerves consistently show lower FA and higher water diffusion values compared to healthy controls.¹⁶

These diffusion metrics correlate with electrical nerve function. In ulnar neuropathy, FA values tracked with nerve conduction velocity, and visual assessment of FA maps identified patients with conduction slowing with about 80% sensitivity.¹⁷ In diabetic neuropathy, DTI can distinguish between axonal injury and demyelination, two fundamentally different types of nerve damage that call for different clinical approaches. Patients with axonal injury showed lower FA values than those with demyelination.¹⁸

Ultra-high-field MRI at 7 Tesla (7T), roughly twice the strength of the scanners used in most hospitals, pushes resolution even further. At 7T, individual nerve fascicles and their surrounding sheaths become visible, with sharper contrast than at the standard 3T.¹⁹ Diffusion-weighted imaging at 7T has demonstrated the ability to trace individual nerve fibers in the forearm using fiber tractography.²⁰ These tools remain largely in the research domain for now, but they represent the direction the field is moving.

When Contrast Dye Helps and When It Does Not

Gadolinium-based contrast agents are routinely used in MRI to highlight areas where the blood-tissue barrier has broken down. In the brain, contrast enhancement reliably flags areas of inflammation or tumor. In peripheral nerves, the situation is less straightforward. The blood-nerve barrier does break down after compression or injury, and experimental studies have shown that contrast enhancement appears in nerves subjected to significant mechanical force.²¹ Newer experimental contrast agents have shown promise in animal models for detecting blood-nerve barrier disruption.²²

However, standard gadolinium contrast does not consistently enhance injured peripheral nerves, even when the blood-nerve barrier is demonstrably open.²³ Contrast-enhanced MRI can be useful in specific scenarios, such as identifying nerve tumors or active inflammatory lesions, but it is not a reliable standalone indicator of neuropathy. Most dedicated MR neurography protocols rely primarily on non-contrast sequences.

MRI Around Metal Implants

A common practical concern is whether neuropathy can be assessed on MRI when the patient has orthopedic hardware near the nerve in question. Metal creates distortion artifacts that can obscure nearby structures. This is not a trivial problem: some of the neuropathies that most urgently need imaging, such as sciatic nerve injury after hip replacement, occur right next to large metal implants. Dedicated artifact-reduction techniques using specific pulse sequences can shrink the distortion zone substantially. Researchers have successfully visualized sciatic nerve injuries adjacent to hip prostheses by carefully selecting sequences and positioning the patient to minimize artifact.²⁴ Specialized metal artifact reduction software has reduced artifact volume by as much as 69% around dental implants when imaging the inferior alveolar nerve.²⁵ The technology keeps improving, but imaging near metal still requires an experienced team and a tailored protocol. A generic MRI order will not cut it.²⁶

How MRI Compares to Nerve Conduction Studies

Nerve conduction studies and electromyography (EMG) have been the gold standard for diagnosing neuropathy for decades. They measure how fast electrical signals travel along a nerve and whether the muscles are receiving those signals properly. MRI and nerve conduction studies answer fundamentally different questions. Electrical testing tells you how the nerve is functioning right now. MRI tells you what the nerve looks like structurally, where the damage is located, and what the surrounding tissues are doing.

The two methods are increasingly seen as complementary rather than competing. MR neurography has demonstrated high sensitivity in predicting nerve conduction abnormalities and can be particularly valuable when electrical testing produces inconclusive results.²⁷ MRI also has clear advantages in preoperative planning, showing the surgeon exactly where the nerve is compressed, scarred, or disrupted. Nerve conduction studies cannot tell you that a neuroma has formed at a specific point along a nerve, or that the nerve is being compressed by an adjacent cyst. MRI can.

The tradeoff is cost and accessibility. An MRI-first approach to diagnosing ulnar neuropathy at the elbow, for example, is the most expensive diagnostic pathway. A cost-effectiveness analysis found that electrodiagnostic testing alone costs substantially less than MRI alone, and combining modalities drives costs higher still. The decision usually comes down to clinical context: straightforward cases can be diagnosed electrically, while complex, surgical, or diagnostically uncertain cases benefit from adding MRI.

MRI for Children With Nerve Injuries

MR neurography is not limited to adults. High-resolution nerve imaging can detect and characterize peripheral nerve injury in children, including traumatic injuries beyond the relatively well-known birth-related brachial plexus palsy.²⁸ Because children heal and adapt differently from adults, and because nerve surgery outcomes depend heavily on timing, having a clear structural picture of the nerve injury can influence treatment decisions. The main practical challenge in pediatric imaging is keeping a young child still long enough for a high-resolution scan, which often means sedation or general anesthesia for younger patients.

Automated Nerve Analysis With AI

One of the bottlenecks in MR neurography is the time it takes a radiologist to manually trace nerves through dozens of image slices. Deep-learning models are being developed to automate this task. A fully automated 3D segmentation model achieved high overlap with expert-drawn nerve outlines in both healthy volunteers and patients with neuropathy, and the resulting segmentations could potentially be used to calculate nerve volume and signal abnormalities automatically.²⁹ Earlier work using fully convolutional neural networks demonstrated that automated segmentation could detect both focal and extended nerve lesions based on cross-sectional area or signal changes, potentially serving as a biomarker for lesion burden.³⁰ None of these tools have replaced the radiologist yet, but they point toward a future where MR neurography could become faster, more reproducible, and more widely available outside specialized academic centers.