Denticulate Ligaments: How They Anchor the Spinal Cord

Denticulate ligaments are thin, ribbon-like structures that anchor the spinal cord inside the spinal canal, keeping it centered within its protective sleeve of fluid and membranes. Running along each side of the spinal cord, they send roughly 21 tooth-shaped projections outward to attach to the tough outer membrane called the dura mater. Despite being easy to overlook, these small ligaments play a surprisingly large role in spinal cord stability, cerebrospinal fluid dynamics, and neurosurgical planning.

Where They Are and What They Look Like

Each denticulate ligament begins as a narrow strip of fibrous tissue running along the lateral surface of the spinal cord, positioned between where the front (motor) and back (sensory) nerve roots exit. From this continuous strip, triangular or tooth-like projections fan outward at regular intervals and anchor into the inner lining of the dura mater. Anatomical studies describe them as roughly 20 to 21 pairs of projections, though some counts place the number at 21 to 22 pairs, depending on how the most superior and inferior attachments are defined.¹²

Their shape changes as you move down the spine. In the cervical region (the neck), the projections are distinctly triangular. In the thoracic region (mid-back), the shape shifts into more of a Y configuration. The most caudal, or lowest, denticulate ligament sits at about the L1-2 vertebral level, roughly where the spinal cord itself tapers off into the conus medullaris.³ The spacing between each projection and the nearest nerve root also increases from the cervical to the lower thoracic segments, meaning the ligaments are packed more closely together near the top of the spine and spread farther apart as they descend.

The ligaments are covered by a thin layer of the same membranes that envelop the spinal cord itself, the pia mater and arachnoid. This leptomeningeal covering means they sit within the subarachnoid space, bathed in cerebrospinal fluid on all sides.⁴

What They Are Made Of

Denticulate ligaments are primarily collagen, but the collagen fibers are not arranged randomly. In the narrow continuous strip that runs along the spinal cord, the fibers are oriented longitudinally, running up and down the length of the spine. In the triangular projections that extend outward to the dura, the fibers switch to a transverse and oblique orientation, fanning sideways to absorb forces coming from different directions.⁵

The ligaments are not uniform from top to bottom. The collagen fibers are thicker and more densely packed at cervical levels than at thoracic levels.⁶ This makes intuitive sense: the cervical spinal cord carries signals for both the arms and the legs and sits in a region with a wide range of neck motion, so the anchoring structures there need to be more robust. The thoracic spine, by contrast, is braced by the ribcage and moves less, so less structural reinforcement from the ligaments is needed.

How They Stabilize the Spinal Cord

The core job of the denticulate ligaments is mechanical. They suspend the spinal cord within the dural sac the way guy-wires hold up a tent pole, preventing it from sagging or shifting to one side. Without them, the spinal cord would be held in place mainly by the nerve roots exiting at each segment and by the fluid pressure of the surrounding cerebrospinal fluid, which would be a less stable arrangement.

Tensile testing of cadaveric denticulate ligaments reveals that despite their small size, they behave like typical biological soft tissues under load. The average rupturing force measured in one study was about 1 newton, roughly the weight of a small apple resting in your hand. No meaningful differences in ultimate strength or stiffness were found between ligaments harvested from different vertebral levels, suggesting that even though the cervical ones are thicker, the mechanical properties per unit area are fairly consistent along the spine.⁷ Finite element computer models built to mimic these ligaments accurately predicted how they stretch and deform, which gives researchers a tool for studying what happens to the spinal cord under different loading conditions, such as during trauma or certain disease states.

The Highest Denticulate Ligament and Its Intracranial Neighbors

Most of the denticulate ligaments live within the spinal canal, but the very first one, the highest denticulate ligament, crosses into intracranial territory. It attaches to the dura near the marginal sinus, just above where the vertebral artery pierces the dura to enter the skull. Anatomical dissection confirms that this uppermost ligament consistently passes between the vertebral artery and the spinal accessory nerve, two major structures in the craniocervical junction.⁸

That tight real estate matters because blood vessels in the area can loop around the ligament. A cadaveric case report documented the posterior inferior cerebellar artery, one of the major arteries feeding the cerebellum, taking an extreme lateral course around the edge of the highest denticulate ligament. The artery made a sharp 180-degree turn at the ligament and was mildly compressed against the floor of the posterior cranial fossa.⁹ Variants like this are worth knowing about because a surgeon operating in the craniocervical junction who encounters an unexpectedly tethered artery needs to understand where the tethering comes from and how to release it safely.

Why They Matter in Chiari Malformation

Chiari I malformation is a condition where the lower part of the brain, the cerebellar tonsils, descend through the opening at the base of the skull and crowd the space around the upper spinal cord. One consequence is disrupted cerebrospinal fluid flow, which can cause headaches, nerve symptoms, and sometimes the formation of a fluid-filled cavity within the spinal cord (a syrinx). Physicians have long observed unusual flow patterns on imaging in Chiari patients, including flow that splits into distinct front-and-back compartments and bidirectional currents during moments of flow reversal, but the physical reason for these patterns was not well understood.

Computational fluid dynamics modeling now points to the denticulate ligaments as a key piece of the puzzle. Patient-specific simulations showed that simply making the anatomical boundaries more realistic did improve the match with actual velocity measurements from MRI, but the characteristic anterior-posterior compartmentalization and bidirectional reversal patterns only appeared when denticulate ligaments were included in the model.¹⁰ In other words, the ligaments act as physical dividers within the subarachnoid space, channeling fluid into distinct compartments and creating flow features that would not exist without them. For Chiari patients, this finding helps explain why symptoms can be so variable: slight differences in the ligaments’ position, thickness, or number could reshape how fluid flows around the compressed cord and brainstem.

The clinical takeaway is still emerging, but the research opens the door to using computational models, personalized with a patient’s own anatomy, to predict which Chiari patients might benefit most from surgical decompression and where the fluid bottleneck actually sits.

Seeing Denticulate Ligaments on MRI

For most of the history of spinal imaging, denticulate ligaments were too small and too similar in signal to surrounding structures to be seen clearly. That has changed with advances in MRI. Using volumetric sequences on standard clinical scanners, radiologists can now identify most of the cervical denticulate ligaments. One study comparing 1.5-Tesla and 3-Tesla MRI found detection rates of roughly 70 to 91 percent at 1.5T and 68 to 98 percent at 3T, depending on the vertebral level. The two field strengths performed similarly overall, though 3T scanners were better at picking up the ligaments in the upper cervical levels.¹¹

Ultra-high-field MRI pushes the boundary further. Imaging healthy volunteers at 7 Tesla revealed not only fine white-and-gray-matter detail within the spinal cord but also structures that are invisible at lower field strengths, including denticulate ligaments, individual nerve roots, and tiny blood vessels running along the cord.¹² Seven-Tesla scanners are still largely research tools, not widely available in clinical practice, but the proof of concept is important. Being able to see the ligaments noninvasively could eventually help surgeons plan approaches to spinal tumors or decompressions, and it could help researchers study whether ligament thickening or calcification contributes to symptoms in conditions like Chiari malformation, spinal cord tethering, or chronic myelopathy.

Surgical Significance

Neurosurgeons encounter denticulate ligaments regularly during open spinal cord procedures. The ligaments serve as reliable anatomical landmarks: because they sit between the ventral (motor) and dorsal (sensory) nerve roots, dividing the cord into front and back halves, they help surgeons orient themselves inside an operative field where the anatomy can be distorted by tumors or other pathology. Cutting a denticulate ligament and retracting it is a standard step during removal of intradural extramedullary tumors such as meningiomas and schwannomas. Severing the ligament at its dural attachment allows the surgeon to gently rotate the spinal cord a few degrees to access a lesion on the far side without excessive retraction, which reduces the risk of cord injury.

The relationship between the highest denticulate ligament and the vertebral artery described earlier is directly relevant to surgeries at the craniocervical junction. Procedures like far-lateral and posterior fossa approaches can bring surgeons into close quarters with this ligament, and knowing that a major artery may loop around or be tethered by it helps avoid catastrophic vascular injury.¹³

There has also been interest in whether deliberately cutting denticulate ligaments could benefit certain patients. The theory, sometimes called “denticulotomy,” is that releasing the tethering effect of the ligaments might relieve tension on the spinal cord in patients with conditions like tethered cord syndrome or certain spinal cord injuries. The evidence for this remains limited and the practice is not standard, but it reflects a broader appreciation that these ligaments are not just passive bystanders. They transmit forces to the cord, and altering that force transmission could, in principle, change clinical outcomes.

Denticulate Ligaments in Other Animals

Denticulate ligaments are not unique to humans. They have been described in a range of vertebrates, including amphibians. In urodeles (salamanders and newts), the ligaments share the same basic composition: collagen, elastin, and fibroblasts concentrated in the segmental lateral projections that extend from the cord.¹⁴ The structure looked similar across three different amphibian species examined in one study, suggesting the design is evolutionarily conserved. Of particular interest, the study also identified neurons at the margins of the amphibian spinal cord closely associated with the ligaments, a finding that raises questions about whether the ligaments have sensory or signaling roles in addition to their mechanical ones.

The presence of elastin in amphibian denticulate ligaments is worth noting because it is not prominently described in human specimens, where the emphasis falls on collagen fiber orientation. Whether human ligaments contain functionally meaningful amounts of elastin, or whether this is a feature more developed in animals that move in fundamentally different ways (swimming versus upright bipedal locomotion), remains an open question.

A Brief History of Denticulate Ligament Anatomy

The denticulate ligaments were first described in 1739 by Johann Jacob Huber, a Swiss anatomist.¹⁵ Since then, there has been persistent disagreement about some of their basic descriptive features: how many pairs exist, what shapes they take at different spinal levels, exactly where they attach, and what function they serve beyond simple mechanical support. Part of the confusion stems from natural variability between individuals. The ligaments can differ in number, attachment point, and thickness from person to person, and some projections may be absent or duplicated at a given vertebral level. This variability is clinically relevant because a surgeon relying on a textbook description of ligament positions might not find them exactly where expected in a particular patient.

Modern microsurgical and imaging studies have done much to clarify the picture, but the field still does not have a definitive atlas that accounts for the full range of normal variation. That gap matters because as procedures become more minimally invasive and as computational models become more patient-specific, the precision with which we understand these small structures will increasingly affect outcomes.

Common Misconceptions

People who encounter denticulate ligaments for the first time, typically in anatomy courses, sometimes come away thinking they are just passive connective tissue straps with no real functional importance. The research paints a different picture. Their role in shaping cerebrospinal fluid dynamics, for instance, was only recently appreciated and changes how clinicians think about conditions like Chiari malformation.¹⁶

Another misconception is that the ligaments are identical from top to bottom. As described above, their shape, fiber thickness, and spacing all change along the spine. A cervical denticulate ligament is a structurally different entity from a lower thoracic one, even if both share the same name and basic plan.¹⁷ This matters practically: a surgical technique that works well for dividing a cervical ligament may not translate directly to the thoracic spine, where the ligaments are thinner and the geometry of surrounding structures is different.

Finally, the idea that denticulate ligaments can always be seen on routine MRI is overstated. While high detection rates have been reported with dedicated volumetric MRI sequences, standard clinical protocols may not visualize them reliably, especially in the lower cervical and thoracic regions.¹⁸ If a radiologist’s report does not mention them, it does not mean they are absent or abnormal; it more likely means the scan was not optimized to show them.