How to Test Cranial Nerve 3 (Oculomotor Nerve)

Testing the oculomotor nerve (cranial nerve III, or CN III) comes down to checking three things at the bedside: how the eyes move, how the pupils react, and whether the eyelids open properly. Each of these functions maps onto a different set of fibers within the nerve, so a thorough exam can reveal not just whether CN III is working but where along its path something might be going wrong. The distinction matters because some causes of CN III dysfunction are medical emergencies, while others resolve on their own.

What the Oculomotor Nerve Actually Does

CN III is the workhorse of eye movement. It controls four of the six muscles that move each eyeball: the medial rectus (turns the eye inward), the superior rectus (turns it upward), the inferior rectus (turns it downward), and the inferior oblique (tilts and elevates). It also innervates the levator palpebrae superioris, the muscle that raises the upper eyelid. On top of all that motor work, CN III carries parasympathetic fibers that travel to the ciliary ganglion and then project to the sphincter muscle of the iris and the ciliary body, controlling pupil constriction and the focusing mechanism known as accommodation.¹ When you test CN III, you are really probing all three of these domains: eye movements, eyelid function, and pupil responses.

Testing Eye Movements with the H-Pattern

The standard bedside test for CN III’s motor function is the H-pattern test, sometimes called the “six cardinal positions of gaze” test. You hold a target (a finger, a penlight, or a small toy for children) about 30 to 40 centimeters from the patient’s face, then slowly trace an H shape in the air while asking the patient to follow the target with their eyes and keep their head still. Each arm of the H isolates a different extraocular muscle and, by extension, the cranial nerve supplying it. The assessment looks for any underaction or overaction of the extraocular muscles and can reveal structural or neurological problems affecting the central nervous system.²

For CN III specifically, you are watching what happens when the patient looks up, down, and inward. If the medial rectus is weak, the eye will not adduct (turn toward the nose) fully. If the inferior rectus is affected, the eye lags when looking down. If the superior rectus or inferior oblique are involved, upward gaze on that side will be incomplete. A complete CN III palsy is hard to miss: the affected eye drifts outward and slightly downward (a “down and out” position) because the lateral rectus and superior oblique, which are supplied by other cranial nerves, pull unopposed.

One subtlety worth noting: when you watch a patient track your finger through the H pattern, you are testing all three motor cranial nerves at once (CN III, IV, and VI). The positions of gaze are designed so that each muscle is the primary mover in a specific direction. For CN III, the key positions are medial gaze, upward gaze (especially when the eye is slightly abducted), and downward gaze (when the eye is slightly abducted). If only one of those directions shows a deficit, you may be dealing with a partial or divisional palsy rather than a complete one.

Checking the Pupil

The pupil exam is arguably the most important part of CN III testing because it helps separate dangerous causes from benign ones. The parasympathetic fibers that constrict the pupil run along the outer surface of the oculomotor nerve. That location makes them vulnerable to compression from an expanding aneurysm or a herniating brain, but relatively protected from ischemic damage that affects the nerve’s interior.

Start by checking pupil size in dim lighting. Compare both pupils. A difference of more than about a millimeter that was not present before is significant. Then shine a bright light into one eye and observe whether that pupil constricts (the direct response) and whether the opposite pupil constricts as well (the consensual response). Repeat on the other side. A pupil that is fixed and dilated on the side of an eye-movement deficit is the classic alarm signal for compressive CN III palsy.

You should also test the near response. Ask the patient to focus on a distant object, then quickly shift their gaze to your finger held about 10 centimeters away. Normally, both pupils constrict when the eyes converge for near vision. A pupil that constricts to a near target but not to light is a dissociation that points to specific diagnoses, including aberrant regeneration of CN III fibers, a phenomenon sometimes seen after chronic compression or congenital palsy.

Assessing the Eyelid

CN III innervates the levator palpebrae superioris, so a third nerve palsy often produces ptosis, a drooping of the upper eyelid on the affected side. In a complete palsy, the lid may droop so far that it covers the pupil entirely. In partial palsies, the ptosis can be subtle, with only a millimeter or two of difference between the two sides.

A useful clinical measurement is the margin-reflex distance (MRD1), defined as the distance in millimeters between the center of the pupillary light reflex and the upper eyelid margin. In a normal eye this distance is roughly 3 to 5 millimeters. In studies of congenital ptosis, an MRD1 below 0.5 millimeters indicated severe levator weakness that responded better to a frontalis suspension procedure, while a value at or above 0.5 millimeters corresponded to milder weakness amenable to levator resection.³ Even if you are not a surgeon, measuring MRD1 with a ruler gives you a reproducible way to track whether ptosis is getting better or worse over time.

Keep in mind that ptosis is not unique to CN III problems. Horner syndrome produces a milder ptosis with a small pupil (the opposite of CN III ptosis, which pairs with a large pupil). Myasthenia gravis can cause fatigable ptosis that worsens over the course of the day. These patterns help you narrow down what you are dealing with.

Convergence and Accommodation

Because CN III supplies the medial recti (for convergence) and carries the parasympathetic fibers controlling the ciliary body (for accommodation), testing these functions adds another layer to your exam. To check convergence, bring a target slowly toward the bridge of the patient’s nose and watch both eyes turn inward symmetrically. Note the distance at which one eye breaks outward, losing alignment; this is the near point of convergence (NPC). Clinical measurement of NPC typically uses an accommodative target on a ruler, and the examiner records the distance from the break point to the lateral canthus of the eye.⁴ A normal NPC is usually within about 6 to 10 centimeters; values further out can suggest CN III weakness, convergence insufficiency, or other disorders.

Accommodation is harder to measure at the bedside without instruments, but you can get a rough sense by asking the patient to read small print at near distance. If focusing ability is markedly reduced on one side (tested with the other eye covered), the parasympathetic component of CN III may be compromised.

Complete Versus Partial Palsy

A complete CN III palsy produces the full constellation: the eye is down and out, the upper lid droops, and the pupil is fixed and dilated. But many CN III palsies are incomplete, and the pattern of what is affected tells you about the anatomy of the injury. The nerve divides into a superior division (supplying the superior rectus and levator) and an inferior division (supplying the medial rectus, inferior rectus, inferior oblique, and the parasympathetic pupil fibers). A superior division palsy may show only ptosis and a limitation looking up, with normal pupil responses and normal adduction. One reported case of superior division CN III palsy due to diabetic microvascular ischemia presented with ptosis as essentially the only sign.⁵ An inferior division palsy, by contrast, might spare the lid but show poor adduction, downgaze limitation, and a dilated pupil.

Recognizing these partial patterns matters because a clinician who expects the full “down and out with a blown pupil” picture may overlook a subtle partial palsy, and the workup for a partial palsy is often the same urgency as a complete one, depending on whether the pupil is involved.

Why the Pupil Decides the Urgency

The single most important distinction in a new CN III palsy is whether the pupil is involved. A dilated, unreactive pupil alongside eye-movement deficits raises the immediate concern for a posterior communicating artery aneurysm (PCoAA), which can present with oculomotor nerve palsy in roughly one out of five affected patients.⁶ A ruptured aneurysm at this location is life-threatening, so a pupil-involving CN III palsy typically triggers emergency neuroimaging, often CT angiography, right away.

Pupil involvement is also the hallmark of uncal herniation, where rising intracranial pressure pushes the temporal lobe downward and compresses CN III against the edge of the tentorium. Unilateral pupillary dilation in this context, historically called a Hutchinson pupil, was later linked specifically to compression or angulation of the oculomotor nerve rather than to direct midbrain damage.⁷ In a trauma or neurosurgical setting, a newly dilated and fixed pupil on one side is treated as a neurosurgical emergency until proven otherwise.

On the other hand, a CN III palsy that spares the pupil is more commonly caused by microvascular ischemia, often in patients with diabetes or hypertension. The explanation is that ischemia damages the interior of the nerve, where the motor fibers live, while the superficial parasympathetic fibers retain their blood supply. That said, pupil-sparing is not perfectly specific. A case report documented pupil-sparing in a complete third nerve palsy caused by a midbrain stroke, not a peripheral ischemic insult at all.⁸ So the pupil helps stratify risk, but it does not guarantee the diagnosis by itself.

Diabetic Oculomotor Palsy

Diabetes is one of the most common causes of isolated CN III palsy in adults, especially older adults. The mechanism involves hyperglycemia-induced damage to the nerve’s own small blood vessels, leading to ischemic injury within the nerve trunk.⁵ The infarction tends to affect the interior of the nerve first, which explains why the pupil is typically spared or only mildly affected in diabetic cases.⁹

Diabetic CN III palsy usually improves on its own over weeks to a few months as the nerve recovers from the ischemic insult. The typical management is to control blood sugar, treat any coexisting vascular risk factors, and recheck the patient periodically. If the palsy does not begin to improve within two to three months, or if it worsens, clinicians generally revisit the diagnosis and consider imaging to rule out compression.

Signs of Aberrant Regeneration

When CN III is damaged and then partially regrows, the new fibers sometimes reach the wrong targets. This aberrant regeneration produces distinctive and somewhat bizarre examination findings. The most recognized of these is the pseudo-Graefe sign (also called the Fuchs sign), where the upper eyelid retracts when the patient looks down, the opposite of what should happen. Other signs include the pupil constricting when the patient looks to one side (the so-called pseudo-Argyll Robertson pupil), and retraction of the globe when attempting vertical eye movements.¹⁰

Aberrant regeneration is clinically important for two reasons. First, its presence suggests a chronic or slowly progressive cause of CN III damage, such as a meningioma or cavernous sinus lesion, rather than an acute ischemic event. Ischemic CN III palsies that resolve cleanly do not typically produce aberrant regeneration. Second, these misdirected connections do not resolve, so surgery on the opposite, healthy eye is sometimes considered to reduce the cosmetic and functional mismatch between the two sides.

Testing CN III in Children

CN III palsy in children has a different profile from adults. In a study of pediatric ocular motor cranial nerve palsies, the third nerve was the most commonly affected, involved in about 39% of cases. More than half of pediatric CN III palsies were congenital in origin, and amblyopia (reduced vision due to disuse) was associated with third nerve palsy in roughly 77% of those children.¹¹

Congenital CN III palsy often looks quite different from the acquired version. In a series of 14 patients with congenital oculomotor palsy, 10 had associated neurologic disorders, challenging the assumption that the congenital form is an isolated finding. The pupil on the affected side was involved in 12 out of 14 patients, but in eight of those, the pupil was actually small (miotic) rather than dilated, likely because of aberrant reinnervation that develops over time. Two children whose pupils were dilated at birth later developed miotic pupils within months as aberrant connections formed.¹² High-resolution MRI in children with congenital CN III palsy has revealed a range of nerve abnormalities, from diffuse thinning to complete absence of the nerve, with corresponding atrophy of the muscles it should supply.¹³

Testing a young child’s eye movements obviously requires some creativity. Instead of asking a toddler to follow your finger, you use a bright toy, a video on a phone, or a moving light to attract their gaze in different directions. Ptosis is usually visible on inspection, and pupil reactions can be checked with a penlight in a dim room. The H-pattern still works, but you may need to move the target more slowly and wait for the child’s attention to catch up.

Automated Pupillometry

The traditional penlight exam is somewhat subjective: different examiners may disagree on whether a pupil is sluggish versus briskly reactive. Automated pupillometers, handheld devices that use infrared light to measure pupil size and the speed and extent of constriction, have improved reproducibility. Studies have shown that automated pupillometry provides better accuracy and interrater agreement compared with bedside eye examinations, and a derived metric called the Neurological Pupil Index (NPI) correlates reliably with third nerve function.¹⁴

In the specific context of third nerve palsy, digital pupillometry has been tested as a way to distinguish compressive from microvascular causes. An abnormal pupillary constriction ratio turned out to be highly specific for compressive lesions, meaning that if the ratio was abnormal, a compressive cause was very likely present. However, the sensitivity was not high, so a normal pupillometry reading could not reliably exclude compression.¹⁵ Automated pupillometry is becoming more common in neurosurgical and neurological intensive care units, where detecting subtle changes in pupil reactivity over hours can signal worsening brain swelling or a new compressive lesion before other signs appear.

Pharmacological Pupil Testing

Sometimes the bedside exam does not fully clarify what is happening with the pupil, and pharmacological testing can help. One relevant scenario involves Adie’s tonic pupil, a condition in which one pupil is dilated and responds poorly to light but slowly constricts with prolonged near effort. Adie’s pupil can be confused with a CN III palsy, especially if the examiner does not test near response carefully.

The dilute pilocarpine test exploits a principle called denervation supersensitivity. When the nerve supply to the iris sphincter has been chronically disrupted (as in Adie’s), the muscle becomes hypersensitive to cholinergic stimulation. A very dilute concentration of pilocarpine, too weak to constrict a normal pupil, will cause an Adie’s pupil to constrict noticeably. Research comparing two concentrations found that the 0.0625% solution performed significantly better than the 0.125% solution for detecting Adie’s tonic pupil, achieving 100% sensitivity and about 83% specificity when a constriction of at least 0.5 millimeters was used as the cutoff.¹⁶ This test is useful precisely because it separates Adie’s pupil from an acute CN III problem. In a true acute CN III compressive palsy, the iris sphincter has not had time to develop denervation supersensitivity, so the dilute pilocarpine will not constrict it.

Other pharmacological agents are used in different clinical scenarios. Cocaine drops test for Horner syndrome (a sympathetic problem, not CN III), and hydroxyamphetamine can localize where the sympathetic chain is interrupted. These are not CN III tests per se, but they come into play when a clinician is sorting through the differential diagnosis of a pupil that does not look right.

Conditions That Can Mimic a Third Nerve Palsy

Not every combination of ptosis and double vision is a CN III palsy. Myasthenia gravis is the most common impersonator. It can produce ptosis, limitation of eye movement in any direction, and even a dilated pupil in rare presentations, closely mimicking a third nerve palsy. The key distinguishing features are fatigability (symptoms worsen with sustained effort and improve with rest) and variability (the pattern of weakness shifts over days or even hours). An ice test, where placing ice on the closed eyelid for two minutes temporarily improves myasthenic ptosis, can be a simple bedside clue.

Thyroid eye disease can restrict eye movements by enlarging the extraocular muscles themselves, which sometimes looks like a nerve palsy on casual inspection. However, the restriction pattern is different: the eye has trouble moving away from the direction of the tight muscle, rather than toward the direction of the weak one. Orbital imaging usually makes this distinction clear. Cavernous sinus lesions, such as tumors or inflammatory conditions like Tolosa-Hunt syndrome, may also involve CN III alongside other cranial nerves (IV, V1, V2, VI), producing a more complex pattern of deficits than a pure CN III palsy. When multiple cranial nerves supplying the eye are affected simultaneously, the localization shifts from the nerve itself to the cavernous sinus or orbital apex.