The vestibular system is the balance organ of the body. It sits deep inside the ear, in a small bony chamber called the labyrinth, and it tells the brain how the head is moving so that the eyes can stay steady on the world, the body can stay upright, and the brain can know where it is in space.

It has two parts. The peripheral part is the labyrinth itself — five separate organs that detect different kinds of head movement. The central part is the network of brain regions that receive those signals and act on them: the brainstem, the cerebellum, and parts of the cortex.

When any part of this network is damaged, the brain receives a false signal of motion. That false signal is vertigo.

The peripheral vestibular apparatus comprises three semicircular canals (transducing angular acceleration in three orthogonal planes), two otolith organs — utricle and saccule (transducing linear acceleration and head tilt), and the eighth cranial nerve, which carries the resulting afferent traffic to the brainstem.1,11 The peripheral organs share a single fluid space, filled with potassium-rich endolymph, sitting within an outer perilymphatic space that is continuous with cerebrospinal fluid via the cochlear aqueduct.

Centrally, the four vestibular nuclei (superior, lateral, medial, inferior) in the dorsolateral pons and medulla form the first relay. From there, signals diverge to four destinations: the vestibulocerebellum for calibration, the oculomotor nuclei via the medial longitudinal fasciculus for the vestibulo-ocular reflex, the thalamus and parieto-insular cortex for conscious perception of motion, and the spinal cord via the vestibulospinal tracts for postural control.7,9

Vertigo is the perceptual signature of asymmetry anywhere in this network — most commonly a peripheral lesion that leaves one labyrinth firing more than the other, less commonly a central lesion that disrupts the integration step.

The vestibular system is best conceived as a push–pull neurophysiological circuit. Each pair of canals on opposite sides of the head is coplanar and reciprocally innervated: a rotation that excites one canal's afferents inhibits the contralateral coplanar canal's afferents, and the brain derives motion from the difference in firing rate, not the absolute output of either side.1,11 This is why complete bilateral vestibular failure is, paradoxically, often less disabling at rest than a unilateral lesion, and why central compensation works at all: the substrate is asymmetry, and asymmetry can be reweighted.

The clinically relevant divisions of the eighth nerve track the hair-cell territories: the superior vestibular nerve carries afferents from the anterior and lateral canals, the utricle, and a small slip of the saccule, while the inferior vestibular nerve carries afferents from the posterior canal and the body of the saccule.4 Two facts follow at the bedside. Vestibular neuritis typically affects the superior division (sparing the posterior canal and most of the saccule — which is why cVEMP is often preserved while oVEMP is reduced), and vestibular schwannomas at the internal acoustic meatus most often arise from the inferior division (which is why cVEMP may be the earlier electrophysiological signal of an evolving schwannoma).

Centrally, the vestibular network is unique among sensory systems in lacking a single primary cortex. Vestibular afferents reach the parieto-insular vestibular cortex (PIVC) via thalamic relay, but they also distribute to anterior cingulate, hippocampus, and somatosensory cortex — a topography that explains the cognitive and affective sequelae of vestibular disease as much as it explains the perception of motion.9,8

The labyrinth contains three loop-shaped tubes — the semicircular canals — that detect spinning movements, and two pouches — the utricle and saccule — that detect movement in straight lines and the pull of gravity.

The canals sit at right angles to one another, like the corner of a room, so that any rotation of the head is sensed by at least one canal on each side. The utricle senses horizontal motion (driving in a car), and the saccule senses vertical motion (an elevator).

At the front of the labyrinth is the cochlea, the organ of hearing. Because the two share a fluid space, ear diseases often produce both hearing changes and vertigo.

The three semicircular canals — anterior (superior), lateral (horizontal), and posterior — are aligned approximately orthogonally so that any angular head movement projects onto at least one canal in each labyrinth. Each canal terminates in a dilation, the ampulla, containing a sensory ridge (crista ampullaris) capped by a gelatinous cupula. Hair cells in the crista project their stereocilia into the cupula; endolymph flow through the canal, lagging behind the canal walls during head rotation, deflects the cupula and bends the stereocilia.1,4

The utricle and saccule are otolith organs. Their sensory epithelia (the maculae) are oriented in approximately perpendicular planes — utricular macula roughly horizontal, saccular macula roughly vertical — so they jointly transduce linear acceleration in all directions. Hair cells in each macula are overlaid by an otolithic membrane studded with calcium-carbonate crystals (otoconia), which add mass; linear acceleration shears the otoconial mass relative to the hair cells, deflecting their stereocilia.5,6

In benign paroxysmal positional vertigo, otoconia dislodged from the utricular macula migrate into the (usually posterior) semicircular canal. Once there, they make a canal that should only respond to angular acceleration become responsive to gravity — the defining mechanical fault of BPPV.

The orientation of the canals matters at the bedside. The lateral canal lies in a plane that is about 30° above the horizontal in the upright head; tilting the head 30° backwards (or, conversely, lying supine with the head flexed 30° on the pillow) aligns the lateral canal vertically — the geometry of the Dix-Hallpike and supine roll manoeuvres exploits this directly. The posterior canal sits roughly in the sagittal plane, angled posteriorly, which is why the Dix-Hallpike position (head turned 45°, then dropped into extension) preferentially stimulates the dependent posterior canal.

Beyond the macroscopic geometry, the maculae have an internal topography that explains otolithic dysfunction patterns. The striola is a curved line dividing each macula into regions where hair cells point in opposite directions, producing bidirectional sensitivity. Type I hair cells (calyx-bearing, irregular afferents) cluster centrally near the striola; type II hair cells (bouton synapses, regular afferents) sit peripherally.4,3 Irregular afferents from type I cells are the substrate for short-latency reflexes such as VEMPs; regular afferents subserve sustained postural tone.

The labyrinth communicates with the middle ear via the oval window (footplate of the stapes) and the round window (secondary tympanic membrane), and with the intracranial space via the cochlear aqueduct, vestibular aqueduct, and any anatomical or pathological dehiscence. When the bony enclosure fails — at the superior semicircular canal in SCDS, or at a window in perilymph fistula — the resulting third window allows sound and pressure to enter the labyrinth pathologically, generating the Tullio and Hennebert phenomena.

The actual sensor — the part of the system that turns movement into a nerve signal — is a microscopic hair cell. Each cell has a bundle of tiny hairs sticking out of its top, and one slightly taller hair (the kinocilium) at one edge of the bundle.

When the head moves, the fluid in the labyrinth pushes the hair bundle to one side. If the hairs bend toward the kinocilium, the cell fires the nerve faster. If they bend away, it fires more slowly. The brain reads the change in firing rate as motion.

Vestibular hair cells are mechanotransducers: they convert the physical deflection of their stereociliary bundles into changes in membrane potential.2,3 Each cell carries a bundle of stereocilia of graded height plus a single, taller kinocilium at one end. Adjacent stereocilia are connected at their tips by a fine extracellular tether — the tip link — which is physically coupled to a mechanoelectrical transduction (MET) ion channel near the top of the shorter stereocilium.

When stereocilia deflect toward the kinocilium, tip-link tension rises and the MET channel opens. The endolymph that bathes the apical surface is unusual: it is potassium-rich, so the open channel admits K⁺ (and Ca²⁺) into a cell that is otherwise sitting at a hyperpolarised resting potential. The cell depolarises, voltage-gated Ca²⁺ channels open at the basolateral synapse, and glutamate release onto the afferent terminal rises — increasing the afferent firing rate. Deflection in the opposite direction has the inverse effect, lowering firing rate below baseline.

The tonic discharge of vestibular afferents averages around 90 spikes per second at rest, with a range across the population of roughly 10–200 spikes/s.1 This tonic firing is the baseline against which deflection-driven modulation is read; it is also what allows bidirectional sensitivity, because inhibition has somewhere to go from.

The biophysical detail behind vestibular transduction has clinical consequences. The MET channel is non-selective for small cations and is gated by mechanical force on the time scale of microseconds, making the vestibular afferent the fastest sensory transducer in the body. Two hair-cell subtypes — type I (flask-shaped, ensheathed by calyx terminals) and type II (cylindrical, contacted by bouton terminals) — feed two functionally distinct afferent populations.4,3

Irregular afferents, originating predominantly from the striolar zone via type I hair cells, carry phasic high-gain signals well suited to driving short-latency reflexes; these are the fibres that subserve the cervical and ocular VEMPs, and they are the fibres preferentially compromised by aminoglycoside ototoxicity. Regular afferents, with their lower thresholds and more linear response, carry the tonic position-and-velocity signal that supports the vestibulo-ocular reflex. Galvanic vestibular stimulation differentially activates these two populations — a fact exploited experimentally to dissect their separate contributions to gaze and postural control.

Otoconial integrity is dependent on calcium homeostasis, which is why osteoporosis and vitamin D deficiency have been associated with recurrent BPPV.6 The otoconia themselves turn over slowly across the lifespan; age-related loss of otoconia is one driver of presbyvestibulopathy, the gradual bilateral decline of otolith function that contributes to falls in older adults.

The vestibular nerve carries signals from the labyrinth into a cluster of four small brain regions called the vestibular nuclei, located in the brainstem. From there, the signals fan out to four places: the cerebellum (which fine-tunes balance), the muscles that move the eyes (which keep vision steady), the higher brain (which gives us a conscious sense of motion), and the spinal cord (which keeps us upright).

A problem at any one of these places can cause vertigo. The trick of diagnosis is working out which one.

The vestibular nuclear complex sits in the dorsolateral brainstem, spanning the pontomedullary junction. It comprises four nuclei — superior, lateral (Deiters'), medial, and inferior — each with characteristic afferent inputs and efferent projections.7

Four principal output projections leave the complex. To the cerebellum, via mossy-fibre inputs to the flocculus, nodulus, and uvula, which together form the vestibulocerebellum and are responsible for calibrating the VOR and suppressing inappropriate vestibular reflexes. To the oculomotor (III), trochlear (IV), and abducens (VI) nuclei via the medial longitudinal fasciculus — the anatomical substrate of the VOR. To the ventral posterior thalamus and from there to the parieto-insular vestibular cortex (PIVC), generating conscious perception of motion and orientation.9 And to the spinal cord via the lateral and medial vestibulospinal tracts, maintaining antigravity tone and head-on-trunk stability.10

Because vestibular signals do not converge on a single primary cortex, vestibular lesions produce a richer mix of cognitive, affective, and spatial deficits than other sensory lesions.8

The functional segregation of the vestibular nuclei matters at the bedside. The medial vestibular nucleus is the main relay for canal-driven VOR signals to the oculomotor nuclei; the lateral (Deiters') nucleus is the origin of the lateral vestibulospinal tract; the inferior nucleus carries cerebello-vestibular and commissural traffic; the superior nucleus carries canal afferents to ocular motor nuclei via the MLF. Lesions of the medial nucleus or the MLF produce the gaze and pursuit abnormalities that distinguish central from peripheral vertigo, while lesions of the lateral nucleus produce ipsilateral postural lateropulsion (a feature of lateral medullary infarction).7,10

The vestibulocerebellum's role in VOR calibration is the substrate of central vestibular compensation. After a unilateral peripheral lesion, the asymmetric afferent input drives a recalibration process in the flocculus, nodulus, and uvula that reweights canal and otolith signals and adjusts the gain of brainstem reflex pathways. This adaptation takes days to weeks, is the rate-limiting step in clinical recovery from vestibular neuritis, and is accelerated by early vestibular rehabilitation — patients told to stay still recover more slowly than patients prescribed gaze-stabilisation exercises.10

The cortical vestibular network — PIVC plus its connections to the anterior cingulate, hippocampus, parietal operculum, and somatosensory cortex — has lateralisation: the right hemisphere in right-handed individuals appears dominant for vestibular processing, which may explain why right hemispheric strokes more often produce vertigo and visuospatial neglect than left hemispheric strokes of comparable size.8,9

BPPV is what happens when tiny calcium-carbonate crystals (otoconia) escape from the part of the inner ear that shouldcontain them — the utricle — and drift into a place where they don't belong — one of the semicircular canals, usually the one at the back (the posterior canal).

These misplaced crystals make the canal sensitive to gravity. When the head moves into a position where gravity pulls the crystals through the canal, the patient experiences a brief, intense spinning sensation that fades within a minute. Typical triggers are lying down, rolling over in bed, looking up, or bending forward.

The good news: a single bedside manoeuvre, the Epley repositioning procedure, restores the crystals to where they belong in around 80% of patients on the first attempt. No surgery, no medications.

BPPV is by far the most common peripheral vestibulopathy, with a lifetime prevalence of 3–10% in the general population and rising steeply with age.1 First described by Dix and Hallpike in 1952 as a positional vertigo with a stereotyped torsional nystagmus on provocative manoeuvre,2 its mechanism was explained two decades later by Hall, Ruby, and McClure: free otoconia, displaced from the utricular macula, migrate into a semicircular canal where they sit at the most dependent point.3

The posterior canal is involved in roughly 85–95% of BPPV. The lateral (horizontal) canal accounts for most of the remainder; anterior canal involvement is rare. The dominance of the posterior canal reflects geometry — it is the most dependent canal in the head when the patient is supine, so dislodged otoconia tend to fall into it.

BPPV is sometimes idiopathic, sometimes secondary. Common associations include head trauma (often minor), prolonged bed-rest, viral labyrinthitis or vestibular neuritis, ear surgery, and ageing of the otoconial membrane.11 Low bone mineral density and vitamin D deficiency are increasingly recognised as risk factors for both incident and recurrent BPPV.12

Canalithiasis — free otoconia in the canal — is the mechanism in the great majority of BPPV cases. The minority variant, cupulolithiasis (otoconia adherent to the cupula itself, first proposed by Schuknecht in 1969),4 produces a persistent positional nystagmus rather than the brief paroxysm characteristic of canalithiasis. Distinguishing the two matters because cupulolithiasis is more resistant to conventional repositioning and may require liberatory manoeuvres (Semont, Yacovino) rather than the standard Epley.6

The Bárány Society's 2015 consensus diagnostic criteria define three classical BPPV variants and three emerging ones.7 The classical: posterior canal canalithiasis, horizontal canal canalithiasis (geotropic nystagmus on supine roll), and horizontal canal cupulolithiasis (apogeotropic, direction-changing nystagmus). The emerging: anterior canal BPPV, multiple-canal BPPV, and probable/possible BPPV — the last categories carved out to capture patients with a compatible history and either spontaneous resolution or non-diagnostic test results.

Two diagnostic pitfalls deserve named attention. Persistent downbeat positional nystagmus during a Dix-Hallpike, especially if it does not fatigue and is not associated with vertigo, is a central sign and warrants imaging — it is not BPPV. And bilateral posterior-canal BPPV, more common after head trauma, produces an upbeat-torsional nystagmus that changes torsional direction depending on which ear is dependent — a pattern easily missed if only one Dix-Hallpike is performed.

Here is what is happening inside the ear in the figure above. When the patient sits upright (position 1), the otoconia rest at the bottom of the canal — they don't move, so the patient feels nothing. As soon as the head is tipped backward in the Dix-Hallpike position (position 2), gravity drags the otoconia down the canal. This pulls the canal's fluid with them, which in turn deflects the cupula — the gel-like flap that the nerve uses to detect head movement. The cupula wrongly tells the brain the head is spinning, so the patient feels intense vertigo and the eyes show a characteristic jerky movement called nystagmus.

The Epley manoeuvre exploits the same physics in reverse: by rotating the head through a sequence of positions, gravity is used to carry the otoconia all the way around the canal and back out into the utricle, where the body can dispose of them naturally.

Posterior-canal canalithiasis produces a stereotyped clinical signature.7,2 A latency of approximately 2–20 seconds between attaining the provocative position and the onset of nystagmus reflects the time taken for the otoconial mass to begin sliding. The nystagmus is upbeating (vertical fast phase upward, from the activation of the ipsilateral inferior oblique and contralateral superior rectus) and torsional, with the upper pole of the eye beating toward the dependent ear — the dependent ear being the affected one. Intensity rises in a crescendo and then fades over 30–60 seconds as the otoconia reach their new resting position. Reversal of the nystagmus on returning to sitting is characteristic, as is fatigability on repeat testing.

The Epley manoeuvre is the canonical treatment.5 Each of its four positions is held until the otoconia have settled (usually about 30 seconds, or until the nystagmus subsides): begin in the Dix-Hallpike position with the affected ear dependent (1), rotate the head 90° to the opposite side (2), roll the patient onto the unaffected side with the head facing the floor (3), and sit up with the head still angled forward (4). At each step the otoconia move further along the canal, with the goal of expelling them through the common crus into the utricle. Single-treatment success rates are around 80%; a second manoeuvre rescues most of the remainder.9

The biomechanics are worth understanding in detail because they explain almost every variant of BPPV and almost every treatment failure. The posterior canal lies approximately in the sagittal plane, angled posteriorly and inferiorly. In the upright head, the canal's most dependent point is near the ampulla; dislodged otoconia therefore rest there at baseline, where their weight produces no flow because they are pressed against the cupula at its base rather than dragging fluid past it. The Dix-Hallpike manoeuvre — head turned 45° toward the affected side, then dropped into 20° extension below couch level — aligns the posterior canal with gravity such that the otoconia begin a long slide along the canal, away from the cupula. The dragged endolymph deflects the cupula ampullofugally, which is excitatory for the posterior canal afferent, producing the upbeat-torsional nystagmus.2,7

Recognising why this matters at the bedside: an absent latency (nystagmus the instant the patient is positioned), a failure to fatigue, or persistent rather than paroxysmal nystagmus argues against canalithiasis. Apogeotropic horizontal-canal BPPV is the most common diagnostic pitfall in the lateral canal because the geotropic and apogeotropic forms have opposite implications for the affected ear: in the geotropic form, the affected ear is the one toward which the patient is rolled when nystagmus is more intense; in the apogeotropic form, the affected ear is the one toward which nystagmus is less intense.10

When the Epley fails, three causes deserve consideration in order of frequency: canal conversion (otoconia transitioned to the horizontal canal during repositioning — confirmable with a supine roll test before the patient stands up), wrong canal (the supine roll test was never performed), or cupulolithiasis rather than canalithiasis (consider Semont or Yacovino). Persistent failure after correctly performed repositioning manoeuvres on at least two visits should prompt reconsideration of the diagnosis — central positional vertigo from a posterior fossa lesion can mimic BPPV closely enough to fool a careful examiner on a single visit.

One of the most useful things about BPPV is what it does not do. It does not cause hearing loss, tinnitus, or a feeling of pressure in the ear. If a patient describes positional vertigo along with any of those symptoms, the diagnosis is something else.

The audiogram in BPPV is normal — and that normality is itself diagnostic. Cochlear involvement in a patient with positional vertigo redirects the differential toward conditions that affect both labyrinths together (Ménière's disease, perilymph fistula, vestibular schwannoma, advanced otosclerosis with secondary BPPV) or toward central positional vertigo. The audiogram is therefore a routine part of the BPPV workup not because BPPV affects it but because finding an abnormality rules BPPV out as the sole diagnosis.

BPPV may co-exist with hearing-affecting pathology, and the combination matters. BPPV is described after stapedectomy and cochlear implantation (mechanical disruption of the utricle liberating otoconia), in established Ménière's disease (theorised mechanism: the same endolymphatic distortion that produces hydrops also disturbs the otolithic membrane), and following sudden sensorineural hearing loss with vestibular involvement.8 In these settings, the audiogram and the positional findings are independent pieces of information, both of which need acknowledging in the management plan. A clean audiogram in an isolated positional vertigo, by contrast, is the green light to proceed with repositioning without further workup in most patients.

Diagnosis is purely clinical — by history and the Dix-Hallpike manoeuvre at the bedside. No imaging, no blood tests, no hearing tests are required to make the diagnosis (although a hearing test is often done to be sure nothing else is going on in the ear).

Treatment is the Epley manoeuvre. It works the first time in about 80% of patients; if it doesn't, it can be repeated at the same visit or a few days later. Medications like betahistine, prochlorperazine, or benzodiazepines do not treat BPPV — they only mask the symptom, and they should not be used routinely.

The Bárány Society's diagnostic criteria for definite posterior-canal canalithiasis require:7

1. Recurrent attacks of positional vertigo provoked by lying down or turning over in the supine position.
2. Duration of attacks < 1 minute.
3. Positional nystagmus elicited after a latency of one or a few seconds by the Dix-Hallpike manoeuvre or side-lying manoeuvre — upbeating and torsional with the upper pole of the eyes beating toward the lower ear; duration < 1 minute.
4. Not attributable to another disorder.

The AAO-HNS 2017 clinical practice guideline is the authoritative reference for management.8 Its strong recommendations include: treat posterior-canal BPPV with a canalith repositioning procedure; do not recommend post-procedural postural restrictions (an older practice disproven by trials); and do not routinely treat with vestibular suppressants such as antihistamines or benzodiazepines. Vestibular rehabilitation may be offered as an adjunct. Imaging is not indicated unless atypical features raise concern for central pathology.

Patients should be told that recurrence is common (rates of 20–50% over five years), that recurrence does not imply treatment failure, and that the same Epley manoeuvre will generally work again.

A complete BPPV examination has three parts: a Dix-Hallpike to each side, a supine roll test (head turned 90° to each side from the supine starting position) for the lateral canals, and a straight head-hanging manoeuvre when the Dix-Hallpike provokes an atypical pattern. Video-oculography, where available, increases sensitivity for low-intensity nystagmus and is particularly helpful for identifying the torsional component, which can be missed under direct observation.

Treatment selection follows the canal and the mechanism. Posterior-canal canalithiasis: Epley manoeuvre, repeated if necessary. Posterior-canal cupulolithiasis: Semont liberatory manoeuvre. Horizontal-canal canalithiasis (geotropic): Lempert/Gufoni manoeuvre toward the unaffected ear, or the "barbeque roll" (360° rotation in 90° increments). Horizontal-canal cupulolithiasis (apogeotropic): Gufoni for cupulolithiasis (toward the affected ear), then re-test; conversion to a geotropic pattern is favourable. Anterior canal BPPV (rare): straight head-hanging or Yacovino manoeuvre.10,6

Failed repositioning after two or three attempts at the same visit should prompt an exit strategy rather than persistence. Re-examine 3–7 days later; if still failing, reconsider the canal, reconsider canalithiasis vs cupulolithiasis, and — if examination supports it — consider Brandt-Daroff habituation exercises as a self-administered alternative. Surgical management (posterior-canal occlusion) is reserved for the vanishingly rare patient with intractable, debilitating BPPV refractory to repeated correctly-performed manoeuvres.8

Vestibular neuritis is an inflammation of the balance nerve running from one ear to the brain. The body, often after a cold or viral infection, mistakenly attacks part of this nerve. The result is a sudden onset of intense vertigo that lasts several days, with nausea and unsteadiness, and a strong urge to lie completely still.

Hearing is not affected — this is the key feature that distinguishes it from inner-ear infections that involve the cochlea too.

The symptoms typically peak within a day and gradually improve over one to two weeks, although a slight feeling of imbalance can persist for months. Most patients recover completely, helped along by vestibular rehabilitation exercises.

Vestibular neuritis is one of the most common causes of acute vestibular syndrome (AVS): sudden-onset, persistent, continuous vertigo lasting at least 24 hours, with associated nausea, head-motion intolerance, and unsteady gait. Estimated annual incidence is around 3.5 per 100,000.1

The Bárány Society's 2022 consensus criteria prefer the term acute unilateral vestibulopathy(AUVP), reflecting uncertainty about whether every case is genuinely inflammatory — but the older "vestibular neuritis" remains widespread in clinical practice and the two are used interchangeably.4

The diagnostic challenge is not making the diagnosis — the history and bedside examination are stereotyped — but making it safely. Posterior circulation stroke (PICA, AICA) can present identically, and missing such a stroke is the single most consequential vestibular error in emergency medicine. Most of this module is about the bedside discriminators.

Vestibular neuritis preferentially affects the superior division of the vestibular nerve — the bundle that carries afferents from the anterior and lateral canals and the utricle. The inferior division (posterior canal, saccule) is usually spared. This anatomical selectivity is a clinical fingerprint: the head impulse is abnormal in the horizontal plane (lateral canal failure), oVEMP amplitudes are reduced (utricular failure), but cVEMPamplitudes are preserved (saccular sparing), and posterior-canal BPPV may develop in the aftermath as the dependent canal's otoconia are liberated by inflammation.1,2

The aetiology is most commonly proposed as reactivation of latent herpes simplex virus type 1 in Scarpa's ganglion. HSV-1 DNA has been demonstrated in postmortem vestibular ganglia, and the geographical and seasonal patterns of neuritis cluster with viral epidemics.3 Alternative aetiologies — autoimmune, microvascular, post-vaccination — have been proposed but lack consistent evidence. Bilateral simultaneous neuritis is exceedingly rare and should prompt consideration of ototoxicity, autoimmune inner ear disease, or central pathology.

The three parts of HINTS each test something slightly different about how the brain and inner ear interact. The head impulse tests whether the inner ear can keep the eyes locked on a target while the head is suddenly turned. The nystagmus test looks for an abnormal pattern of eye jerking. The skew test looks for one eye drifting up while the other drifts down — a sign of brainstem trouble.

Used together, these three tests reliably separate the common inner-ear problem (vestibular neuritis) from the rarer but dangerous brain stroke. The key teaching point: an abnormal head impulse test is reassuring — it means the inner ear is at fault, not the brain.

Each HINTS component has a specific peripheral and central signature. The head impulse, first described as a bedside sign by Halmagyi and Curthoys in 1988,5tests the high-acceleration horizontal VOR by rapidly rotating the head while the patient fixates on the examiner's nose. A working VOR keeps the eyes locked on the target; a failed VOR produces a corrective saccade as the eyes catch up. In vestibular neuritis, the head impulse is abnormal when the head is rotated toward the affected ear. In posterior fossa stroke, the head impulse is typically normal — because the labyrinth itself is intact.7

The video head impulse test (vHIT) is the instrumented refinement, using lightweight goggles with high-frame-rate cameras to quantify VOR gain and detect covert saccades. Sensitivities exceed 90% for canal-specific paresis in experienced hands.6

Nystagmus in vestibular neuritis is unidirectional, mostly horizontal with a torsional component, beating toward the healthy ear, enhanced by removing fixation (Frenzel glasses). It obeys Alexander's law — intensity rises when looking toward the fast phase. Central nystagmus, by contrast, often changes direction with gaze, may be purely vertical or torsional, and is not suppressed by fixation.

Skew deviation is the rarest of the three components but the most specific for central pathology. Detected with alternate cover testing: the uncovered eye refixates vertically as it takes up fixation. It reflects an ocular tilt reaction from disruption of the graviceptive pathways in the brainstem.

The discriminating power of HINTS depends entirely on operator skill and on careful patient selection. The original Kattah cohort comprised patients with continuous vertigo lasting at least 24 hours and at least one stroke risk factor — exactly the AVS population where HINTS performs as advertised.8 Applied to patients with episodic or positional vertigo, or to patients without persistent spontaneous nystagmus, HINTS loses sensitivity and specificity dramatically. Do not test HINTS in a patient without nystagmus — the test is built on the asymmetry of spontaneous vestibular tone.

Three failure modes deserve specific mention. First, AICA infarction can disrupt the labyrinth's own blood supply and produce an abnormal head impulse alongside the central signs — so an "abnormal HI" in a patient with new hearing loss is a red flag, not a green light. Second, small cerebellar infarcts of the posterior inferior cerebellar artery territory may produce HINTS findings indistinguishable from neuritis on one or two components but typically violate the rule on the third. Third, skew deviation can be subtle — an experienced examiner will pick up minor vertical refixation that is missed on inspection alone.

The HINTS'Plus' modification adds bedside finger-rub hearing as a fourth component; new unilateral hearing loss in AVS classifies the case as central. This increases sensitivity for AICA strokes (which involve the labyrinthine artery) at the cost of a few additional "central" calls in labyrinthitis patients.9,10

A normal hearing test is part of confirming the diagnosis of vestibular neuritis. If hearing has dropped on one side, the problem is more likely to be in the inner ear itself (a condition called labyrinthitis) or, more worryingly, a small stroke affecting both the balance nerve and the hearing supply at the same time.

Audiometric assessment is recommended in every AVS workup, not because vestibular neuritis affects hearing but because hearing changes redirect the differential. Acute SNHL with persistent vertigo is concerning for AICA infarction — the labyrinthine artery is a terminal branch — and warrants vascular imaging regardless of an otherwise "peripheral" HINTS.10 A conductive hearing loss in the same context should prompt examination for tympanic membrane pathology and consideration of acute otitis media with secondary labyrinthitis.

Beyond the audiogram, supportive evidence for the diagnosis of vestibular neuritis includes a unilateral caloric paresis on bithermal caloric testing, reduced VOR gain on vHIT for the affected horizontal canal, and reduced or absent oVEMPs with preserved cVEMPs (reflecting selective superior-division involvement). None of these are required for diagnosis under the Bárány criteria, but they aid localisation and have medicolegal value when stroke must be definitively excluded.4

Treatment focuses on symptom control, prevention of long-term imbalance, and recovery. Anti-nausea medications are used for the first few days; longer-term use slows recovery and is discouraged. Once the worst of the vertigo has passed, the patient should be encouraged to get up and move — staying still actually makes recovery slower. Vestibular rehabilitation, a programme of head and eye exercises supervised by a physiotherapist, accelerates recovery.

Corticosteroids may be used in the first week to limit inflammation; the evidence is suggestive but not definitive. Most patients are substantially better at one month and back to normal by three.

The Bárány Society's 2022 diagnostic criteria for acute unilateral vestibulopathy require:4

1. Acute or subacute onset of sustained spinning or non-spinning vertigo (an acute vestibular syndrome) of moderate to severe intensity, lasting at least 24 hours.
2. Spontaneous peripheral vestibular nystagmus — direction-fixed, enhanced by removing fixation, with a trajectory appropriate to the involved canal afferents (generally horizontal-torsional).
3. Unambiguous evidence of a reduced VOR function on the affected side (clinical head impulse test, vHIT, calorics, or rotational testing).
4. No accompanying acute audiological or central neurological signs.
5. Not better accounted for by another disease or disorder.

Management proceeds along three axes. Symptomatic relief with anti-emetics and vestibular suppressants (e.g. prochlorperazine, antihistamines) is appropriate in the first 72 hours but should be discontinued promptly thereafter — prolonged use suppresses the central compensation process.1Corticosteroids (typically methylprednisolone) within the first 72 hours improved caloric recovery at 12 months in the Strupp 2004 RCT, although a follow-up Cochrane review considered the overall evidence inconclusive.11Antiviral therapy alone or in combination with steroids has not been shown to add benefit. Vestibular rehabilitation started early significantly shortens functional disability and is the best-evidenced intervention.12

Three management decisions in vestibular neuritis warrant deliberate consideration in every case. First, the safety workup. Even after a peripheral HINTS, patients with stroke risk factors and an unconvincing examination should have MRI/MRA — the cost of missing a small posterior fossa stroke in a patient who later deteriorates is high enough to justify imaging liberally. Second, the steroid decision. The Strupp 2004 trial demonstrated improved vestibular recovery on caloric testing at 12 months but did not show improved subjective recovery — many neurologists offer steroids within 72 hours by default, others reserve them, and the evidence supports either practice. Third, the vestibular rehabilitation referral. Early VR halves the time to symptom resolution and is the only intervention with consistently strong evidence; arrange it before the patient leaves the acute setting if practicable.12,11

Long-term outcomes: most patients recover to a functional state but maintain a measurable canal paresis on caloric testing indefinitely. Persistent imbalance — particularly visual-vestibular mismatch in busy visual environments — is common at three to six months and usually resolves with continued vestibular rehabilitation. Failure to compensate after six months should prompt consideration of persistent postural-perceptual dizziness (PPPD) as a secondary diagnosis, with cognitive behavioural therapy and SSRIs as additional management options.

Recurrence is uncommon (around 2% per year) and should prompt reconsideration of the diagnosis — recurrent acute vestibular syndrome is more often vestibular migraine, Ménière's, or vestibular paroxysmia than recurrent neuritis.

Ménière's disease is a chronic condition of the inner ear in which fluid builds up where it shouldn't, distorting the hearing and balance organs. Episodes are dramatic but self-limiting: the patient experiences intense spinning vertigo lasting two to twelve hours, with a drop in hearing on one side, loud ringing, and a feeling of pressure in the ear. Between episodes the patient is usually well, although hearing can fluctuate and a baseline imbalance is common.

Over years, hearing on the affected side typically declines progressively. The condition can affect one ear (most common) or both. There is no cure, but a stepwise approach — starting with diet and lifestyle, escalating to medications and, in refractory cases, procedures — controls symptoms for most patients.

Ménière's disease is an episodic peripheral vestibular disorder defined by the simultaneous involvement of the vestibular and cochlear partitions of one (sometimes both) inner ear. Estimated prevalence is 17–46 per 100,000, with a peak onset between 40 and 60 years of age.1 The condition is named for Prosper Ménière, who in 1861 first attributed the syndrome to the inner ear rather than to a cerebral cause as had been previously assumed.

The defining pathology, demonstrated by Hallpike and Cairns in their seminal 1938 temporal-bone study, is endolymphatic hydrops — distension of the membranous labyrinth by an accumulation of endolymph, particularly affecting the scala media of the cochlea and the saccule.2Whether hydrops itself causes Ménière's symptoms or is a downstream marker has been debated; temporal-bone studies show hydrops in essentially every clinical case, but also in some asymptomatic individuals, which complicates the causal argument.3

Most cases are idiopathic. Around 10% of cases are familial, with a polygenic or autosomal-dominant inheritance pattern emerging in genome-wide studies.7Secondary causes — "Ménière's syndrome" rather than disease — include autoimmune inner-ear disease, otosyphilis, post-traumatic hydrops, and large vestibular aqueduct syndrome.

The pathophysiological cascade is best modelled as impaired endolymph homeostasis: production by the stria vascularis exceeds absorption by the endolymphatic sac, the membranous labyrinth distends, and an acute rise in endolymphatic pressure either ruptures Reissner's membrane locally or transiently opens K⁺ channels at the hair-cell apex — in either case the result is paradoxical hair-cell depolarisation followed by transient functional loss. The hydrops itself is a chronic structural change; the acute attack is a biochemical event superimposed upon it.3

In vivo demonstration of hydrops became possible after Nakashima's 2007 description of intratympanic-gadolinium MRI, which highlights perilymph and reveals the endolymphatic spaces as filling defects.4The technique is not required for diagnosis under current criteria, but it has been useful for confirming hydrops in atypical cases, for distinguishing Ménière's from vestibular migraine in refractory presentations, and for research into the natural history of hydrops. The 3T delayed-gadolinium protocol is the current standard.

Three observations from the temporal-bone literature continue to shape clinical practice. First, saccular hydrops can be isolated — the inferior-division vestibular afferents are preferentially affected, which is why cVEMPs are commonly abnormal in Ménière's while oVEMPs may be preserved. Second, contralateral hydrops emerges on histology before it manifests clinically; bilateral involvement increases over time and reaches 15–40% by 20 years from onset. Third, the endolymphatic sac itself is reduced in size in many patients — the structural basis for the sac-decompression and duct-blockage surgical procedures.1

The figure above is a cross-section through one turn of the cochlea — the snail-shaped hearing organ. Inside, there are three fluid-filled chambers separated by thin membranes. The middle chamber, scala media, contains a special fluid (endolymph) with a high concentration of potassium. The chambers above and below contain a different fluid (perilymph) with low potassium.

In Ménière's disease, endolymph accumulates and the middle chamber swells, stretching the membrane above it (Reissner's membrane). When the pressure becomes too great, this membrane ruptures, the two fluids mix, and the hair cells — which are exquisitely sensitive to potassium concentration — misfire wildly. That is the Ménière's attack.

The membrane eventually heals and normal fluid balance is restored, which is why the attacks are episodic. But the damage is cumulative: with each cycle, a little more hearing is permanently lost.

The cochlea is divided into three parallel scalae running its full length: scala vestibuli (perilymph, continuous with the vestibule), scala media (endolymph, an enclosed compartment with a high K⁺/low Na⁺ profile maintained by the stria vascularis), and scala tympani (perilymph, continuous with the round window niche). Reissner's membrane separates scala vestibuli from scala media; the basilar membrane (with the organ of Corti on top) separates scala media from scala tympani.1

Endolymph is produced primarily by the stria vascularis at a steady rate and absorbed by the endolymphatic sac. When that balance is upset — by sac fibrosis, hypoplastic sac anatomy, autoimmune injury, viral infection, vascular insufficiency, or other proposed causes — endolymph accumulates, scala media distends, and Reissner's membrane bulges upward. Step through the figure to see the progression: normal anatomy → early hydrops (mild distension) → severe hydrops (large distension, hair cells stressed) → acute rupture (membrane fails, K⁺ contaminates perilymph, hair cells depolarise, attack ensues).3

The vestibular sequelae mirror this. The saccule, anatomically contiguous with scala media via the ductus reuniens, distends in parallel and presses against the macula utriculi and the lateral canal wall. With each episode, saccular function is progressively compromised, which is why cVEMPs deteriorate across the disease course.

The hydrops-as-cause hypothesis has limits worth understanding. Merchant et al.'s reanalysis of the temporal-bone literature noted that hydrops is found postmortem in approximately 9% of subjects who never reported Ménière's symptoms — and Ménière's symptoms can occur in the absence of demonstrable hydrops.3 The most parsimonious current view is that hydrops is a marker for a broader endolymphatic dysregulation; the symptomatic attack requires a second event (membrane rupture, transient ion-channel opening, or vascular precipitant) that the hydrops makes possible but does not guarantee.

For the clinician, the practical consequences are three. First, in-vivo MRI demonstration of hydrops (post-gadolinium) supports but does not confirm a diagnosis of Ménière's.4 Second, the goal of medical therapy is to reduce endolymph accumulation (diuretics, low-salt diet, betahistine) rather than to treat the attack itself. Third, in refractory disease the surgical target is the endolymphatic sac — decompression or duct-blockage procedures aim to restore endolymph absorption capacity.11

A hearing test in Ménière's disease almost always shows worse hearing on one side, and the pattern is unusual: the deepest sounds (low frequencies) are affected most. Most forms of hearing loss show the opposite pattern, with high frequencies affected first. This "upside-down" loss is one of the most useful clues.

The Bárány/AAO-HNS 2015 criteria require audiometric documentation of low-to-medium-frequency sensorineural hearing loss in the affected ear for the diagnosis of definite Ménière's disease.5 The required pattern, in any of the inter-ictal or peri-ictal recordings, is: pure-tone average across 500, 1000, and 2000 Hz at least 30 dB worse on the affected side than on the contralateral side (or 35 dB if the contralateral ear is also affected, with additional criteria).

Three caveats deserve attention. First, the loss fluctuates in early disease — a single normal audiogram does not exclude Ménière's; serial audiometry across attacks is more revealing than any single recording. Second, the pattern flattens with disease progression; late-stage Ménière's often shows a moderately severe pancochlear loss that no longer demonstrates the upsloping configuration. Third, similar low-frequency SNHL patterns can be seen in autoimmune inner-ear disease, large vestibular aqueduct syndrome, and intracranial hypotension — the audiogram supports the diagnosis but does not make it alone.1

In stage 1 Ménière's, hearing recovers between attacks — sometimes to normal — and the audiogram is a snapshot, not a baseline. Schedule serial audiometry particularly close to attack times if achievable; the fluctuation is itself diagnostic. Stage 2 brings persistent low-frequency loss that no longer recovers; stage 3 brings expansion to mid- and high frequencies; stage 4 (end-stage) is a flat moderate-to-severe sensorineural pattern with speech discrimination scores out of proportion to pure-tone thresholds and a "dead ear" phenomenon in some patients. The Bárány/AAO-HNS criteria use the four-tone pure-tone average (500, 1000, 2000, 3000 Hz) for staging.5

Electrocochleography (ECochG) measures the summating potential (SP) and the action potential (AP) of the cochlear nerve. An elevated SP:AP ratio (commonly >0.4–0.5) supports a diagnosis of endolymphatic hydrops but is neither sensitive nor specific enough to be a primary diagnostic test — useful as supportive evidence in equivocal cases.

The diagnosis is made by the doctor recognising the pattern: episodic vertigo lasting hours, hearing changes on one side, ringing, and pressure in the ear. There is no single test — a hearing test, MRI, and balance tests are usually done to support the diagnosis and rule out other things.

The 2015 international consensus criteria (Bárány Society, Japan Society for Equilibrium Research, EAONO, AAO-HNS, Korean Balance Society) define two categories:5

Definite Ménière's disease:

1. Two or more spontaneous episodes of vertigo, each lasting 20 minutes to 12 hours.
2. Audiometrically documented low- to medium-frequency sensorineural hearing loss in one ear, defining the affected ear, on at least one occasion before, during, or after one of the episodes of vertigo.
3. Fluctuating aural symptoms (hearing loss, tinnitus, or fullness) in the affected ear.
4. Not better accounted for by another vestibular diagnosis.

Probable Ménière's disease: Two or more episodes of vertigo or dizziness, each lasting 20 minutes to 24 hours; fluctuating aural symptoms; not better accounted for by another vestibular diagnosis. The probable category accommodates patients with the clinical picture but without audiometric confirmation of hearing loss.

Notably, vertigo episodes outside the 20-minute to 12-hour window should redirect the differential: shorter than 20 minutes suggests vestibular migraine or TIA; longer than 24 hours suggests AUVP or stroke.

The 2015 criteria represent a deliberate departure from the 1995 AAO-HNS criteria, broadening the category of probable disease, requiring no specific test other than audiometry, and explicitly framing the diagnosis as one of pattern recognition rather than exclusion. In practice, three differentials cost the most diagnostic mistakes. Vestibular migraine — particularly in patients who lack a migraine history, or in whom auditory symptoms are mild — accounts for many cases initially labelled Ménière's; the temporal pattern of episodes (longer in Ménière's) and the presence or absence of documented hearing loss is the cleanest discriminator. Autoimmune inner ear disease can produce bilateral fluctuating SNHL with vestibular symptoms and progresses faster than Ménière's. Vestibular schwannoma, while not typically episodic, can present with a single attack of acute vertigo plus unilateral hearing loss — image with gadolinium-enhanced MRI in any unilateral SNHL of unclear cause.

Familial Ménière's, while uncommon (≈10%), warrants pedigree analysis when present.7 An autosomal-dominant inheritance pattern with incomplete penetrance is the most common picture; the underlying genes are heterogeneous, and large-scale GWAS are emerging.

Treatment starts conservatively and escalates only if the patient continues to have attacks. The first steps — cutting salt, caffeine, and alcohol; managing stress; taking a water tablet (diuretic) — help around half of patients. If attacks persist, medications such as betahistine are added, then injections into the ear, then in rare cases surgery.

During an attack, the patient should lie still, take an anti-sickness tablet, and wait it out. Most attacks resolve within a few hours. Between attacks, vestibular rehabilitation can help with any baseline imbalance.

The AAO-HNS 2020 clinical practice guideline organises management as a stepwise ladder.6 Begin at the lowest effective rung and escalate only with documented failure.

1. Lifestyle:low-salt diet (typically <1500 mg sodium/day), reduction of caffeine, alcohol, and smoking. Stress and sleep optimisation. Counselling on the natural history. Symptomatic treatment of acute attacks with antihistamines or benzodiazepines as needed.
2. Diuretics: typically thiazides (e.g. hydrochlorothiazide–triamterene) — weak evidence but low cost and good safety profile.
3. Betahistine: a histamine H₃-antagonist / weak H₁-agonist, widely used in Europe and parts of Asia, not FDA-approved in the US. The BEMED trial found no significant benefit over placebo for vertigo attack frequency at high or standard doses, although some patients report subjective benefit.8
4. Intratympanic corticosteroids: (typically dexamethasone) inject into the middle ear through the tympanic membrane. Useful for vertigo control with hearing preservation. Cochrane evidence supports a modest benefit.9
5. Intratympanic gentamicin: aminoglycoside preferentially toxic to vestibular hair cells. Effective for vertigo control in ≈80–90% of refractory cases, with a 25% risk of further sensorineural hearing loss.10 Avoid in bilateral disease (bilateral vestibular failure is a catastrophic outcome).
6. Surgery: endolymphatic sac decompression, endolymphatic duct blockage, or — as last resort and only in unilateral disease with non-serviceable hearing — vestibular nerve section or labyrinthectomy. Surgical evidence is mostly observational; the Cochrane review noted insufficient high-quality RCT data.11

Vestibular rehabilitation has a place between attacks for patients with persistent inter-ictal imbalance; it does not prevent attacks but accelerates compensation for any cumulative vestibular loss.

The hardest decisions in Ménière's management cluster at the boundary between intratympanic steroid and intratympanic gentamicin. Both deliver drug directly to the inner ear; gentamicin is more effective for vertigo control but ototoxic. The decision pivots on hearing status: serviceable hearing on the affected side argues for trying intratympanic steroid first; non-serviceable hearing makes gentamicin a reasonable early step. A network meta-analysis suggested IT gentamicin gives the highest probability of complete vertigo control, with IT steroid plus betahistine second, both well ahead of placebo.10

Three practical considerations in long-term care. First, contralateral disease is the rate-limiting threat — by 20 years from onset, 15–40% of patients have developed bilateral involvement, and patients on the gentamicin track must be re-evaluated before any contralateral intervention. Second, the natural history is favourable for many patients: attacks may "burn out" over years, with progressive hearing loss but reducing vertigo frequency — a course that argues for patience before destructive procedures. Third, comorbid vestibular migraine is common and often underdiagnosed; refractory disease that responds to migraine prophylaxis deserves reconsideration of the primary diagnosis.1

Drop attacks (Tumarkin's otolithic crises) — sudden falls without warning or loss of consciousness, attributed to abrupt utricular dysfunction — are a feature of advanced disease and an indication to escalate definitively. Intratympanic gentamicin reliably abolishes Tumarkin attacks; their persistence is one of the few absolute indications for surgical destruction of the labyrinth.

The inner ear is normally a sealed bony chamber with two small windows for sound and pressure to enter and leave — the oval window (where the smallest bone in the body, the stapes, sits) and the round window (a flexible membrane lower down). In superior canal dehiscence syndrome, a thin spot or hole in the bone above one of the semicircular canals — the canal that arches up under the floor of the brain — creates an unintended third window.

The consequences are striking. Loud sounds can cause vertigo (the Tullio phenomenon). Holding the breath, lifting heavy objects, or even sneezing can do the same (the Hennebert phenomenon). Patients may hear their own heartbeat, eye movements, or footsteps inside the ear (autophony). Some lose hearing in the affected ear at low pitches — but paradoxically hear better than normal through bone conduction.

The condition is uncommon but well-defined. A CT scan showing the missing bone, plus the symptoms, plus an abnormal VEMP test, confirms the diagnosis. Mild cases are managed with reassurance and avoidance of triggers; severe cases benefit from a small surgical repair.

Superior canal dehiscence syndrome was first described in 1998 by Lloyd Minor and colleagues at Johns Hopkins, in a series of patients with sound- and pressure-induced vertigo whose CT scans showed an absent or thinned bony roof over the superior semicircular canal.1The mechanism is now conceptualised as a "third mobile window" in the bony labyrinth: in addition to the normal oval and round windows, the dehiscence creates a third compliant interface through which pressure and sound energy can be diverted away from the cochlea and into the dura or directly onto the membranous canal.

Anatomic dehiscence of the superior semicircular canal is present in approximately 0.5–1% of temporal bones at postmortem, with thinning (<0.1 mm) considerably more common.5 Symptomatic SCDS is rarer — most anatomic dehiscences are clinically silent, probably protected by the overlying dura. The syndrome typically presents in adulthood (4th to 6th decades), with up to half of patients showing bilateral anatomic dehiscence on imaging even when symptoms are unilateral.2

The diagnostic question is rarely "is there a dehiscence?" — CT will answer that — but rather "is the dehiscence the cause of these symptoms?". The Bárány Society 2021 criteria address this by requiring three elements together: characteristic symptoms, a physiologic test demonstrating third-window behaviour, and radiological confirmation of the anatomic defect.2

The biophysics of SCDS hinge on impedance mismatch. Normally, acoustic energy entering at the oval window travels through the cochlear partition (where it is transduced) and exits at the round window — a closed two-window system in which all input pressure must be balanced by an equal-volume displacement at the round window. With a third window present, a fraction of the input energy short-circuits through the dehiscence, producing two effects simultaneously: a relative reduction in air-conducted sound reaching the cochlea (manifest as a low-frequency air-bone gap) and an enhanced response to bone-conducted sound and to inertial stimuli (because the dehiscence becomes a low-impedance pathway for these stimuli to drive cochlear-fluid motion).6,4

The clinical fingerprint of SCDS therefore has two parts — vestibular and audiological — that often present together. Vestibular: sound- or pressure-induced vertigo (Tullio, Hennebert) with vertical-torsional nystagmus in the plane of the affected superior canal. Audiological: pseudo-conductive hearing loss with supranormal bone conduction (negative dB HL thresholds at low frequencies), autophony, pulsatile tinnitus, and audible eye movements or footfalls. The first dehiscent canal patient described by Minor reportedly heard her own pulse so loudly that she could not sleep on the affected side.3

The simulator above shows two ears side by side — one normal, one with the bone gap. Choose "Loud sound" or "Valsalva" to see the difference. In the normal ear, the pressure wave moves harmlessly through the cochlea. In the dehiscent ear, some of the energy escapes upward through the hole in the bone and pushes on the balance organ — which is why a loud noise can make a patient with SCDS feel like the room is spinning.

In the dehiscent labyrinth the bony roof of the superior canal is partly or wholly absent, leaving the membranous canal in contact with the overlying dura (and through it, CSF). This creates a fluid pathway between the labyrinth and the intracranial compartment that is normally impossible. Pressure that should remain bounded within the labyrinth — sound-driven stapes motion, ICP fluctuations, Valsalva-induced venous pressure — can now displace endolymph through the canal opening, deflecting the cupula of the superior canal and generating an aberrant excitatory signal to the central vestibular system.

The direction of the resulting eye movement is stereotyped: vertical-torsional, in the plane of the superior canal — slow phase up and torsional with the upper pole rotating away from the affected ear, fast phase the reverse. This pattern is pathognomonic when seen synchronously with sound or pressure stimulus.3

On the cochlear side, the third window changes the impedance of the inner ear in a frequency-dependent way. At low frequencies, where the stapes drives slow, sustained pressure changes, the dehiscence acts as a pressure-relief valve — energy escapes upward instead of driving the basilar membrane. The result is a low-frequency air-bone gap that superficially mimics otosclerosis. Crucially, bone conduction at low frequencies is enhanced rather than normal: vibrations applied to the skull are preferentially channelled through the low-impedance third window to the cochlear fluids, producing thresholds better than 0 dB HL.6

The supranormal bone-conduction threshold is the audiometric feature that makes SCDS impossible to confuse with otosclerosis on a careful audiogram. In otosclerosis, fixation of the stapes raises air-conduction thresholds but bone conduction is normal (or shows the Carhart notch at 2 kHz — a 5–15 dB depression that disappears after successful stapedectomy). In SCDS, bone conduction is supranormal — the low-frequency points sit above 0 dB HL on the audiogram. Miss the supranormal threshold and the patient gets a stapedectomy that fails to improve symptoms (and, worse, may produce a perilymph gusher).

Vestibular-evoked myogenic potentials reflect the same impedance shift. Both cVEMP (saccular) and oVEMP (utricular) show enhanced responses in SCDS — lower thresholds for cVEMP (often <65 dB nHL, vs the usual 75–95 dB nHL norm) and higher amplitudes for oVEMP. The Bárány criteria accept either an enhanced cVEMP, an enhanced oVEMP, or a low-frequency negative bone-conduction threshold as the physiologic test required for diagnosis.2

In a hearing test, a patient with SCDS shows two unusual findings on the affected side. The first is a hearing loss at low pitches that looks like a problem with the middle ear (the kind seen in glue ear or otosclerosis). The second is the give-away: when sound is delivered through the skull bone instead of through the air, the patient hears better than normal. Skipping this bone-conduction check is how SCDS gets misdiagnosed as otosclerosis and operated on inappropriately.

The Bárány Society accepts "low-frequency negative bone conduction thresholds on pure tone audiometry" as one of three acceptable physiologic-test findings for the diagnosis of SCDS.2"Negative" here means below 0 dB HL — better than the audiometric zero defined for a young, otologically normal population. In SCDS, this supranormal threshold typically appears at 250, 500, and sometimes 1000 Hz on the affected side, with normal bone conduction at higher frequencies and on the unaffected side.

Practical audiometric tips. Use insert phones for bone masking to avoid over-masking and missing the supranormal threshold. Calibrate bone-conduction at -10 or -15 dB HL — many clinical audiometers default to a 0 dB floor and will not register the diagnostic finding. Confirm with a Weber test (which should lateralise to the dehiscent ear despite the apparent "conductive" loss — the opposite of what one might expect for true middle-ear pathology).

Three audiological discriminators separate SCDS from its principal mimic, otosclerosis. First, supranormal bone conduction (SCDS) vs. normal or notched bone conduction (otosclerosis). Second, normal acoustic reflexes (SCDS) vs. absent reflexes (otosclerosis — the stapedial reflex requires a mobile stapes). Third, enhanced VEMP responses (SCDS) vs. normal or absent VEMPs (otosclerosis). A patient with apparent conductive hearing loss whose stapedial reflexes are present is, until proven otherwise, a third-window patient — image with high-resolution CT in the Pöschl plane before considering any stapes surgery.6,7

Doctors need three things to make the diagnosis: the characteristic symptoms (sound or pressure causing vertigo, hearing one's own pulse or eye movements), an abnormal balance or hearing test showing the third-window effect, and a CT scan that confirms the missing bone. Without all three, the diagnosis is not made.

The Bárány Society's 2021 diagnostic criteria require the presence of all three of:2

1. At least one symptom consistent with third-window pathophysiology:
   • Hyperacusis to bone-conducted sound (autophony, audible eye movements, audible footfalls)
   • Sound-induced vertigo and/or oscillopsia time-locked to the stimulus
   • Pressure-induced vertigo and/or oscillopsia time-locked to the stimulus
   • Pulsatile tinnitus
2. At least one physiologic test or signindicating a third mobile window:
   • Eye movements in the plane of the affected superior canal time-locked to sound or pressure
   • Low-frequency negative bone-conduction thresholds on pure-tone audiometry
   • Enhanced VEMP responses (low cVEMP threshold or elevated oVEMP amplitude)
3. High-resolution CT with multiplanar reconstruction in the plane of the superior canal (Pöschl view) and orthogonal to it (Stenvers view) consistent with a dehiscence.

And — explicitly — "not better accounted for by another vestibular disease or disorder."

Two technical points about the imaging deserve attention. Slice thickness must be <1 mm (ideally 0.625 mm or less); thicker slices over-call dehiscence because of volume averaging.7 The Pöschl reformat — parallel to the plane of the superior canal — is the best view for identifying the defect; the Stenvers reformat (perpendicular) is used to confirm. Avoid making the diagnosis on axial slices alone, where bone thickness can be ambiguous.

Other third-window pathologies share the audiometric signature: posterior canal dehiscence, lateral canal dehiscence (associated with chronic otitis media or cholesteatoma), large vestibular aqueduct syndrome (LVAS), and dehiscence of the cochlea into the carotid canal or internal auditory canal. The differential matters because the surgical approach differs by location. In ambiguous cases, electrocochleography may show an elevated SP:AP ratio (>0.4 — non-specific but supportive); the Tullio sign on video-oculography in the plane of the affected canal under sound stimulus is highly specific and worth seeking actively.

Many patients with mild SCDS manage with reassurance and by avoiding the triggers — wearing earplugs in loud environments, avoiding heavy lifting, learning to keep bowel movements gentle (to avoid Valsalva-type strain). For patients whose symptoms are severe enough to interfere with work or sleep, surgery to plug or resurface the affected canal resolves symptoms in about 90% of patients, although a small risk of hearing loss exists.

Management is stratified by symptom severity. Mild symptoms are managed conservatively: counselling and identification of triggers, hearing protection in loud environments, stool softeners and posture advice for Valsalva-sensitive patients. Anti-tinnitus strategies (masking, cognitive behavioural therapy) help with the pulsatile-tinnitus component. Most patients with mild SCDS live well with the condition.

Surgical repair is offered when symptoms are disabling. Two anatomical targets, two surgical approaches:

• Middle fossa craniotomy: the original Minor approach. Provides direct visualisation of the dehiscence and allows precise plugging, capping (cartilage or fascia overlay), or resurfacing. Requires temporal lobe retraction and a 2–3 day hospital stay.1
• Transmastoid approach: the dehiscence is accessed through the mastoid and plugged blind. No craniotomy; shorter hospital stay; lower recurrence rate in the Schwartz multi-institutional series compared with middle fossa.8

Across both approaches, vertigo symptoms resolve in approximately 90% of patients; the systematic review by Gioacchini and colleagues reported a pooled success rate of 94%.9 Hearing outcomes are mostly stable with a small risk (≈5–10%) of high-frequency sensorineural loss. Surgery should be reserved for patients with disabling symptoms because the risk-benefit calculation favours observation in mild cases.

The conversation about surgery is best framed around three axes: which symptom is dominant, how disabling it is, and what hearing-preservation tolerance the patient has. Patients whose primary complaint is autophony alone are generally less benefited by surgery than those with disabling Tullio or chronic disequilibrium — autophony resolves in ≈75% post-operatively, whereas vertigo resolves in >90%. Patients with bilateral anatomic dehiscence on imaging warrant particular caution: operate on the worst side first and reassess at 6 months; bilateral surgery is occasionally necessary but doubles the risk of hearing loss.4

A failed stapedectomy in a patient with an apparent otosclerotic audiogram is the classic SCDS late presentation. If the patient still has third-window symptoms after middle-ear surgery, image the temporal bones and check VEMPs before considering anything further on the stapes — there is a real risk of a perilymph fistula being created on a dehiscent canal.6

Perilymph is the salty fluid that fills the bony outer chambers of the inner ear. Normally, it is sealed in by two membranes — one in each of the two windows that connect the inner ear to the middle ear. A perilymph fistula is a tear or weakness in one of these membranes that allows the fluid to leak.

The condition most often follows a definite trigger: a blow to the head, a diving or flying mishap, very heavy lifting, an explosive sneeze, or an ear operation. Symptoms vary widely — sudden hearing loss in one ear, ringing, dizziness, a feeling of imbalance worsened by exertion or changes in pressure. Many patients describe their balance as "off" rather than vertiginous.

The frustration of PLF is that no test reliably confirms it. The diagnosis is made by combining the story, the pattern of symptoms, the audiogram, and a CT scan to rule out other things — and sometimes confirmed only at surgery, when the surgeon can see the leak. Most patients are managed conservatively first, with rest and pressure precautions, and operated on if symptoms persist.

Perilymph fistula is defined as an abnormal communication between the perilymph-filled inner ear and the air-filled middle ear, usually at the round or oval window membrane.2 First definitively described by Goodhill in 1971 as a cause of sudden sensorineural hearing loss following physical exertion,1 the concept has had a contested fifty-year history; the diagnostic difficulty is real and the literature contains both robust surgical series and published scepticism about whether spontaneous PLF exists at all.

Aetiologically, PLF falls into three main categories. Traumatic PLF is the best-defined — direct head trauma, temporal-bone fracture, ear surgery (especially stapedectomy, where iatrogenic oval-window fistula is a recognised complication), or penetrating injury. Barotraumatic PLF follows diving accidents (descent or ascent), explosive blast, or even Valsalva manoeuvres during weightlifting, sneezing, or vomiting. Spontaneous PLF — without a clear precipitant — is the most controversial category; Kohut and others have demonstrated congenital microfissures in the otic capsule that may predispose certain patients to leaks under minor stress.4

Symptoms cluster into three groups. Cochlear symptoms — sudden or fluctuating sensorineural hearing loss, tinnitus, aural fullness. Vestibular symptoms — chronic disequilibrium more often than discrete vertigo episodes, sometimes with positional or motion-provoked features. Third-window symptoms — Tullio (sound-induced vertigo),Hennebert (pressure-induced vertigo), and autophony — that overlap with SCDS but with the dehiscence at the membranous window rather than the canal roof.

The clinical core of PLF is the same impedance-mismatch physics that underlies SCDS — energy diverted through an abnormal mobile interface — but the anatomical lesion is fundamentally different. SCDS is a defect of bone; the membranous labyrinth itself is intact. PLF is a defect of membrane; the bony labyrinth is intact. The downstream consequence is similar (a low-impedance pathway for pressure transmission, producing third-window symptoms), but the implications for imaging and surgery diverge completely. CT cannot see the leak directly because the membrane itself is radiolucent — at best it shows indirect evidence (pneumolabyrinth: air within the inner ear, pathognomonic but uncommon).6

The lack of a confirmatory non-invasive test has driven two parallel diagnostic strategies. One is the pre-operative biomarker approach: cochlin-tomoprotein (CTP), a perilymph-specific protein detectable by ELISA in middle-ear lavage, has been validated in Japanese series as a specific marker for perilymph leak.5 The other strategy — older but still widely practised — is exploratory tympanotomy: directly visualising the round-window niche and oval-window area under a microscope, grafting both windows even when no overt leak is seen, and observing whether symptoms resolve.7,8

The controversy of the last fifty years has revolved around how aggressively to pursue exploration. The strongest surgical evidence is for early intervention in barotraumatic PLF after head/ear trauma — Park's 2012 series showed a doubling of serviceable-hearing recovery rates when surgery was performed within 10 days vs. later.7The weakest evidence is for spontaneous "PLF" in patients with vague chronic dizziness — here the false-positive rate at exploration is high and the symptom-outcome benefit small.

Walking through the pathway shows what makes PLF different from the other vestibular conditions. There is no single test that says "yes, this is PLF." Doctors put together history (was there a triggering event?), examination, hearing tests, and scans to decide whether the diagnosis is likely enough to act on — sometimes by resting and waiting, sometimes by operating.

The Fitzgerald 1992 clinical criteria remain a useful framework for the bedside workup, although they predate modern imaging and biomarker testing:3

• A history of a precipitating event (head trauma, ear surgery, barotrauma, or strenuous exertion);
• Sudden, fluctuating, or progressive sensorineural hearing loss in the affected ear;
• Vestibular symptoms — vertigo, disequilibrium, or positional symptoms — temporally related to the precipitant or to subsequent exertion;
• Provocative manoeuvres that reproduce symptoms (Valsalva, tragal pressure, positional change);
• Exclusion of other causes.

Modern workup adds high-resolution CT (Pöschl plane — primarily to exclude SCD, occasionally to demonstrate pneumolabyrinth or window pathology) and 3T inner-ear MRI (which may show signal changes in the labyrinth fluid spaces). VEMP testing is sometimes useful: unlike SCDS, VEMPs in PLF are typically normal or reduced rather than enhanced, helping distinguish the two third-window syndromes.6

Cochlin-tomoprotein testing — sampling middle-ear fluid and assaying for the perilymph-specific protein CTP — has emerged as the most specific available pre-operative test. The Ikezono ELISA had a reported sensitivity of around 65% and specificity around 95% in a definite-PLF cohort.5 Availability outside Japan remains limited.

Three pragmatic points anchor PLF diagnosis in clinic. First, the history is the test. A sudden audiovestibular event temporally tied to a defined mechanical insult (barotrauma, head trauma, surgery, Valsalva) carries the weight; the same audiometric and vestibular profile in a patient without a precipitating event has a much lower pre-test probability of PLF. Second, the imaging is exclusionary, not confirmatory: if CT shows SCD, the diagnosis is SCDS unless proven otherwise; if MRI shows vestibular schwannoma, that explains everything. Negative imaging does not rule PLF in. Third, when the picture fits and conservative management has failed, exploratory tympanotomy is both diagnostic and therapeutic, and is the only way to be confident about ruling it in or out.

The decision threshold for exploration deserves explicit consideration. Indications strong enough to proceed at one to two weeks: sudden SNHL with documented progression or fluctuation, unequivocal barotraumatic history with persistent symptoms, post-operative deterioration after stapes or middle-ear surgery. Indications weak enough to defer or decline: chronic non-specific imbalance without triggering event, audiogram unchanged despite weeks of symptoms, and any patient where the differential is more consistent with vestibular migraine, Ménière's, or PPPD.2

The hearing test in PLF is variable. The most common pattern is a one-sided loss that is worse for high pitches, sometimes with a small low-pitch gap from the third-window effect. Unlike SCDS, however, bone conduction is normal rather than supernormal — this distinction is one of the few audiometric clues that separates the two third-window conditions.

PLF audiograms are heterogeneous. Four patterns recur in the literature:2,7a pure high-frequency SNHL (similar to noise-induced hearing loss), a pancochlear SNHL (similar to sudden SNHL of any cause), a fluctuating low-frequency SNHL (a Ménière's mimic), and a low-frequency air-bone gap (the third-window pattern). Audiometric findings alone cannot diagnose PLF — but they can support the pre-test probability and they can help distinguish PLF from SCDS (normal vs. supranormal bone conduction respectively).

Serial audiometry over weeks is more informative than any single recording. Progressive deterioration despite conservative management is a strong indication to escalate to exploration; spontaneous improvement argues for continued observation.

Two audiometric scenarios warrant heightened suspicion for PLF specifically. First, post-stapedectomy deterioration after an initial improvement — the classic iatrogenic perilymph leak from a displaced prosthesis or a torn oval-window membrane. Second, sudden SNHL within hours of a defined barotraumatic event (diving descent, explosive sneeze, weightlifting) — these patients should be considered for early exploration if conservative management does not yield rapid improvement.7,8

Acoustic reflex testing is normal in PLF (unlike otosclerosis), and VEMP responses are typically normal or reduced (unlike SCDS where they are enhanced). The combination of an apparent conductive component on the audiogram with normal reflexes and normal-to-reduced VEMPs supports a membranous third-window pathology without a bony defect — that is, PLF rather than SCDS.

Treatment depends on how strong the suspicion is and how bad the symptoms are. Mild and well-tolerated symptoms, particularly without confirmed hearing loss, are usually managed conservatively — strict rest, avoiding heavy lifting, sneezing through an open mouth, sleeping with the head elevated, and waiting for the membrane to heal itself. Many patients improve over one to two weeks.

If symptoms persist or get worse, the surgeon may recommend an exploratory operation through the ear canal under microscope, looking for the leak and grafting the window with the patient's own tissue. Recovery from surgery is straightforward; vertigo and balance symptoms usually improve, although hearing recovery is less predictable.

Conservative management is the first step in most cases. Bed rest with head elevation (30–45°), stool softeners, avoidance of straining, lifting, blowing the nose, and Valsalva, and — where the precipitant was barotrauma — eustachian-tube measures (decongestants, nasal steroids). A typical conservative trial runs 1–2 weeks; review at that point with repeat audiometry guides escalation.

Exploratory tympanotomy with sealing of the oval and round windows is the definitive surgical option. The approach is transcanal under local or general anaesthesia. The surgeon elevates the tympanomeatal flap, inspects the round-window niche and oval-window region under magnification (sometimes with the patient performing Valsalva to provoke a leak), and grafts both windows with fat, fascia, perichondrium, or fibrin glue — most surgeons graft both windows even when no overt leak is seen, because microfistulae are difficult to visualise and the graft itself is well-tolerated.7,8

Outcomes are best documented for barotraumatic and post-traumatic PLF. Park's 2012 series reported a mean post-operative hearing gain of 27 dB, with serviceable hearing recovery in 57% when operated on within 10 days vs. 33% after. Vertigo symptoms resolve in ≈85% of operated patients regardless of timing. Heinrich 2012 reported visible perilymph fistulae at exploration in 18.8% of sudden-SNHL patients overall, with hearing improvement after sealing whether or not a leak was seen.7,8

Three management decisions in PLF deserve careful framing. First, when to operate. The strongest evidence is for early operation (within 7–10 days) in barotraumatic PLF with documented hearing loss. The weakest evidence is for elective exploration in long-standing non-specific dizziness; outside a research setting these patients are better served by conservative management and a vestibular-migraine or PPPD workup. Second, what to seal. Both windows are grafted in most contemporary series even when no leak is seen at exploration; the graft itself is innocuous and misses fewer microfistulae than a selective approach. Third, the conversation with the patient. PLF surgery is one of the few situations in otology where the procedure is partly diagnostic — the patient should understand that a normal-appearing exploration does not exclude the diagnosis (microfistulae are hard to see), and the post-operative course is the real arbiter of whether the diagnosis was correct.

For pure spontaneous PLF in a patient without a triggering event, the case for surgery is much weaker — most contemporary practice in Europe and North America defers exploration in this group, with vestibular rehabilitation, conservative measures, and reconsideration of the differential as primary management. Japanese centres with access to cochlin-tomoprotein testing have published the most thorough recent diagnostic frameworks.5,6

The brainstem and the back of the brain — the cerebellum — receive their blood supply from two arteries that run up the back of the neck and join into a single artery called the basilar. A stroke in this circulation produces vertigo, imbalance, slurred speech, double vision, or weakness on one side of the face or body. Sometimes only vertigo, making it easy to mistake for an inner-ear problem.

This is the diagnosis behind every dizzy patient's emergency presentation. Most are not strokes. But missing a stroke is the worst possible outcome — both for the patient (because secondary prevention prevents a much bigger second stroke) and because some posterior strokes, particularly basilar artery occlusion, need urgent thrombolysis or clot retrieval that has a narrow time window.

The bedside examination — the HINTS test, plus a careful check of gait, hearing, and the rest of the neurology — done by a trained clinician within 24 hours of symptom onset, is the most reliable way to separate the inner-ear from the brainstem cause. An MRI in the first 24–48 hours misses up to 1 in 8 small posterior strokes; the bedside examination misses far fewer.

Posterior circulation strokes — affecting the brainstem, cerebellum, and posterior thalamus — account for approximately 20% of all ischaemic strokes, and are misdiagnosed at twice the rate of anterior circulation events.1 The Tarnutzer 2017 meta-analysis found that roughly 9% of cerebrovascular events are missed at the initial ED visit, with the misdiagnosis rate rising to 35% when the presenting complaint is dizziness or vertigo.2

The clinical challenge has three roots. First, the presentation overlaps with common benign vestibular disorders: vestibular neuritis, vestibular migraine, BPPV. Many posterior strokes — particularly small PICA-territory cerebellar infarcts — present with isolated vertigo indistinguishable from peripheral vestibulopathy on the vertigo axis alone.4 Second, the standard neurological screen is often unrevealing — there is no facial droop, no hemiparesis, no aphasia in many posterior strokes, and the NIH Stroke Scale (designed for cortical strokes) scores them as low. Third, diffusion-weighted MRI, the standard stroke imaging modality, has a false-negative rate of 12–20% in posterior strokes within the first 48 hours of onset, falling to nearly zero by day 7.

The combined effect is that frontline assessment requires a different toolset for the dizzy patient than for the hemiparetic patient. The HINTS examination, gait assessment, bedside hearing, and a structured central nystagmus look — all woven into a single 5-minute examination — outperforms early MRI for ruling stroke in or out at the bedside.

The diagnostic framework that has emerged from twenty years of vestibular medicine research (Newman-Toker, Edlow, Kerber, Tarnutzer) reframes acute dizziness around the timing and triggers of symptoms rather than the patient's description of the dizziness itself. Three clinical syndromes are differentiated: acute vestibular syndrome (AVS) — continuous vertigo lasting hours to days, in which the differential is neuritis vs. central infarction; episodic vestibular syndrome triggered (s-EVS) — short positional or exertional episodes, where the differential is BPPV vs. central positional vertigo or third-window pathology; andspontaneous episodic vestibular syndrome (s-EVS)— recurrent attacks without a clear trigger, where the differential includes vestibular migraine, Ménière's, posterior-circulation TIA, and panic disorder. Each syndrome calls for a different examination strategy.9

The GRACE-3 emergency department guideline, published in 2023, is the current authoritative document for the workup of acute dizziness in the ED. Its central recommendations: characterise the timing pattern first; use HINTS for the AVS group; do not over-rely on CT or early MRI for ruling out posterior stroke; and arrange expedited outpatient assessment with vascular imaging (CTA or MRA) for patients with suspected TIA.9

The vascular substrate matters. Vertebral artery dissection is the leading cause of posterior stroke in patients under 50, often following neck trauma (chiropractic manipulation, head turning, sports). Atherothrombosis is the leading cause in older patients with vascular risk factors. Top-of-basilar syndrome from cardioembolism is the emergency that justifies aggressive imaging and thrombectomy referral.5,10

The five territories above are the "named places" of posterior circulation stroke. Each one produces a slightly different mix of vertigo, balance, and other neurological signs. The two most important to know are PICA (the most common cerebellar stroke that can present with isolated vertigo) and AICA (the great mimic of vestibular neuritis because it also affects the inner ear).

Three points worth committing to memory. PICA infarction is the most common cerebellar stroke; Lee 2006 showed that 11% of cerebellar infarcts present with isolated vertigo, mostly from medial-PICA territory.4These are the patients most easily mistaken for vestibular neuritis — and the patients in whom HINTS pays its biggest dividend.

AICA infarction involves the labyrinth because the labyrinthine arteryis a terminal branch of AICA. The result is a stroke that can present like neuritis on every component of HINTS except the hearing — hence the "HINTS-Plus" modification that adds a bedside hearing check.8

Lateral medullary (Wallenberg) syndrome — usually from PICA or vertebral artery occlusion — is the named brainstem syndrome of vestibular medicine. Its clinical fingerprint (ipsilateral Horner, ipsilateral facial sensory loss, contralateral body sensory loss, dysphagia, hoarseness, ataxia) is so distinctive that HINTS becomes irrelevant — a Wallenberg patient is diagnosable on standard neurological examination alone.

The two cerebellar territories that warrant specific operational concern beyond diagnosis: large SCA infarction carries the highest risk of cerebellar oedema and brainstem compression within 24–72 hours, and may need decompressive suboccipital craniectomy.11Top-of-basilar embolic syndrome from cardioembolism (often atrial fibrillation, often poorly anticoagulated) presents with bilateral oculomotor and visual deficits, altered consciousness, and prominent vertigo — the window for endovascular thrombectomy is 4.5–24 hours depending on imaging selection. Any acute vertigo presentation with bilateral signs, oculomotor abnormalities, or altered consciousness should bypass standard HINTS and trigger immediate vascular imaging.10

The card above is a quick reference to the most common abnormal eye-movement patterns that indicate a brain (rather than inner-ear) cause. The most important takeaway: any purely vertical nystagmus (downbeat or upbeat) or any purely torsional nystagmus is a central sign that warrants urgent imaging.

Three nystagmus patterns deserve specific attention in the acute setting. Downbeat nystagmus localises to the cervicomedullary junction or vestibulocerebellum — the differential includes Chiari I malformation (often missed in young patients), cerebellar degeneration, posterior fossa stroke, and toxin exposure (lithium, anticonvulsants). It is worsened by downgaze and often increased by lying supine.

Periodic alternating nystagmus (PAN) — horizontal jerk nystagmus that reverses direction every 60–90 seconds — is highly specific for nodulus-uvula lesions or cervicomedullary pathology. It is easily missed if the examiner does not watch for several minutes. PAN is one of the few nystagmus patterns with a specific treatment: baclofen.

Pure torsional nystagmus (no horizontal or vertical component) is always central, almost always midbrain or medullary, and warrants urgent imaging regardless of the rest of the examination.

Two practical observations on bedside nystagmus discrimination. First, the corollary of "central nystagmus rules in central" is not "peripheral-looking nystagmus rules out central". A unidirectional horizontal-torsional nystagmus that obeys Alexander's law can still be cerebellar — the cerebellum can mimic peripheral patterns convincingly. Use the broader examination (gait, head impulse, hearing) to disambiguate. Second, the suppression-by-fixation test — using Frenzel goggles or a covered eye to remove visual fixation — separates peripheral nystagmus (suppressed by fixation) from central nystagmus (not suppressed). It's a five-second additional test that adds substantial discriminative power.12

Doctors examining a dizzy patient look for several specific things to decide whether the cause is the inner ear (most common) or the brain (much rarer but much more serious). The most important: can the patient walk or stand unsupported? Inner-ear vertigo patients are usually able to, even if they look unsteady. Patients with a cerebellar stroke often cannot stand at all without falling.

Other things to check: the HINTS examination (described in the Vestibular Neuritis module), hearing on each side, whether the patient's face moves normally, and whether there is any difference in feeling between the two sides of the body.

A structured 5-minute bedside examination for the dizzy patient covers the following. History: onset (sudden vs. gradual), duration (seconds vs. hours vs. days), triggers (positional, exertional, spontaneous), and associated symptoms (hearing, speech, swallowing, vision, weakness). The triggered/spontaneous distinction determines which examination sub-protocol applies.9

HINTS for the AVS patient — see the Vestibular Neuritis module for the full technique. Three steps; any one central finding makes the pattern central.6

Gait: the most useful single bedside discriminator after HINTS. A patient who cannot stand unsupported with eyes open is not a vestibular neuritis patient — that is a cerebellar truncal ataxia until proven otherwise. Romberg testing distinguishes proprioceptive ataxia (eyes-closed worse) from cerebellar ataxia (worse in both).

Bedside hearing: rub fingers near each ear and ask the patient to say which side they hear it better. Acute asymmetric hearing in an AVS patient is a red flag for AICA infarction.7,8

Cranial nerves and limbs: a screening examination — facial movement and sensation, palate elevation, tongue protrusion, limb power, sensation, and coordination (finger-to-nose, heel-to-shin). Any focal abnormality redirects the patient to a stroke pathway regardless of HINTS.

Three nuances of bedside discrimination at the clinician level. First, the head impulse test should be performed with the patient's head pre-positioned 30° forward of neutral (to align the lateral semicircular canals with the horizontal plane) and the patient instructed to fixate on the examiner's nose with eyes locked. A high-velocity, low-amplitude rotation (10–20° at >1000°/s) is required — a slow rotation produces false negatives. Covert saccades (compensatory saccades that occur during the head movement itself, invisible to the naked eye) require vHIT to detect; their presence is a peripheral finding, their absence in the right context is reassuring.

Second, the bedside head-shaking test (vigorous head-shaking 20 times then observation for nystagmus) adds supplementary information: a perverted (cross-coupled) vertical or torsional response after horizontal shaking is a central sign. Third, the test of skew should be done with alternate cover testing, watching for vertical refixation as fixation is transferred from one eye to the other — easily missed if the patient is looking down or if cover times are too short.

A patient who cannot sit unsupported but has otherwise preserved limb coordination, sensation, and strength is exhibiting truncal ataxia — a cerebellar-vermis or paravermal finding that points directly to PICA territory infarction. This sign is more useful than any single HINTS component and is missed if the patient is examined only in the recumbent position. Always test sitting balance with arms folded across the chest.

Early CT scans of the head are good at finding bleeds but miss almost all acute ischaemic strokes in the posterior circulation. Early MRI is better but still misses about 1 in 8. This is why the bedside examination is so important: a normal early scan does not rule out a posterior stroke. Patients with persistent symptoms or red flags often need repeat imaging at 48–72 hours.

The imaging hierarchy in suspected posterior circulation stroke:

• Non-contrast CT head — first-line in any acute stroke presentation. Sensitivity for acute ischaemic stroke is poor in the first 24 hours; sensitivity for posterior fossa lesions is even worse because of beam hardening artefact from the petrous bones. CT primarily excludes haemorrhage.
• CT angiography (CTA) — head and neck — evaluates the vertebral and basilar arteries for stenosis, dissection, or occlusion. Should be performed in any patient with red flags or any concern for vascular cause, irrespective of CT head findings.
• MRI with diffusion-weighted imaging (DWI) — the imaging modality with the highest sensitivity for acute ischaemic stroke. False-negative rate 12–20% in posterior strokes within 48 hours. Repeat at 5–7 days if clinical suspicion remains and initial DWI is negative.2
• MR angiography (MRA) — used as the vascular component when MRI is the chosen modality.

The 2019 AHA/ASA stroke management guideline calls for vascular imaging in any patient with suspected posterior-circulation TIA — particularly because vertebral artery dissection requires anticoagulation rather than antiplatelet therapy.10

The clinician's most important imaging decision is not what to order but what to do with a normal scan. A patient with persistent AVS, a central-looking HINTS, and a negative initial MRI should be admitted for monitoring and repeat imaging at 48–72 hours; the false-negative rate falls dramatically by day 7. Discharge with outpatient follow-up is appropriate only when the clinical examination is reassuring, HINTS is peripheral, gait is preserved, and there are no red flags. The cost of an unnecessary admission is far lower than the cost of a missed posterior stroke that returns with completed cerebellar infarction or basilar occlusion.1

One specific scenario deserves separate attention. A young patient (under 50) with acute vertigo and neck pain — even without a clear traumatic precipitant — should have CTA or MRA of the vertebral arteries to exclude dissection. Vertebral dissection accounts for up to 25% of strokes in young adults; the typical clinical picture is occipital or posterior neck pain preceding vertigo by hours to days, often after minor trauma or even routine activities like hairdressing or chiropractic manipulation.5

Migraine is a brain condition usually associated with severe headaches. But not every migraine attack involves a headache, and many people with migraine experience attacks of vertigo — sometimes only vertigo — that come and go in the same pattern as their headaches once did. This is vestibular migraine.

Vestibular migraine is the most common cause of recurrent vertigo that is not from the inner ear. It affects roughly 1% of the general population and up to 10% of patients with migraine. Attacks last anywhere from five minutes to three days; some patients have only a few attacks per year, others have several per month.

The diagnosis is made by recognising the pattern of attacks, not by any specific test. Examination between attacks is usually normal. The good news is that vestibular migraine responds to the same preventive medications used for headache migraine — beta-blockers like propranolol, topiramate, and others. Many patients also identify and avoid triggers (sleep deprivation, stress, certain foods, hormonal changes) and find this alone helps substantially.

Vestibular migraine is now recognised as one of the most common vestibular disorders. Population prevalence estimates range from 0.98% (the seminal Neuhauser 2006 study) to 2.7% in different methodologies.4 Approximately 10% of all patients with migraine experience migrainous vertigo at some point, and approximately 20% of patients presenting to vestibular clinics with recurrent vertigo have vestibular migraine as their final diagnosis.3,5

The diagnostic criteria, jointly developed by the Bárány Society and the International Headache Society in 2012 and updated in 2022, define two categories: definite vestibular migraine (all four criteria met) and probable vestibular migraine (a relaxation that allows diagnosis when either the migraine history or the migraine features during attacks are absent).1,2 Vestibular migraine appears in the appendix of ICHD-3 as a developing diagnosis — formal inclusion in the main body of ICHD requires further accumulated evidence, but the appendix entry is the authoritative classification for clinical use.6

Three features make vestibular migraine particularly important to recognise. First, it is common — far more common than Ménière's, far more common than vestibular schwannoma, far more common than most other episodic vestibular diagnoses combined. Second, it is treatable — well-evidenced prophylactic agents (propranolol, topiramate) substantially reduce attack frequency. Third, misdiagnosis matters: vestibular migraine patients inappropriately treated as Ménière's with intratympanic gentamicin lose vestibular function unnecessarily.8

The pathophysiology of vestibular migraine remains the subject of active research. Three converging hypotheses shape the current understanding. The cortical spreading depression mechanism — well-established for migraine aura — likely contributes to the perceptual disturbance during vestibular migraine attacks, with spreading depression affecting central vestibular processing areas (parietoinsular vestibular cortex, the vestibular nuclei). The trigeminovascular activation theory implicates the same neuropeptides (CGRP, substance P) that drive headache, with vestibular manifestations attributed to CGRP-mediated modulation of the vestibular nuclei and inner ear vasculature — the rationale for trialling anti-CGRP monoclonal antibodies in refractory vestibular migraine. The ion-channel dysfunction theory draws on the familial hemiplegic migraine genetics (CACNA1A, ATP1A2) and proposes that vestibular migraine shares aspects of episodic ataxia type 2, where calcium-channel mutations produce both migraine phenomena and vestibular episodes.

Clinically, the lifespan epidemiology is distinctive. Vestibular migraine has the broadest age distribution of any vestibular disorder — onset can occur in childhood (as benign paroxysmal vertigo of childhood, the migraine equivalent of recurrent vertigo with normal examination, often evolving into typical migraine in adolescence) or in later adulthood (after menopause, women whose typical migraines have remitted may develop vestibular migraine). The female-to-male ratio is approximately 3:1, mirroring migraine generally.7

The longitudinal pattern matters for diagnosis: many patients describe a personal history of childhood motion sickness, a peak of headache migraine in their 20s and 30s, and the emergence of vestibular attacks in their 40s and 50s — sometimes as headaches diminish. This temporal relationship — vertigo replacing headache rather than accompanying it — is the most common reason patients fail to recognise their own diagnosis. Many do not connect the two until specifically asked.

For a doctor to diagnose "definite" vestibular migraine, a patient needs to meet all of the following:

• At least five episodes of vertigo of moderate or severe intensity, each lasting 5 minutes to 72 hours;
• A history of migraine (the headache kind);
• Migraine features (headache, light/noise sensitivity, or visual aura) during at least half of the vertigo episodes;
• No better alternative explanation.

If the patient has the episodes but not the migraine history, or has the episodes and the history but the episodes don't have migraine features, the diagnosis is called "probable" instead of definite. Either way, the treatment is the same.

The full diagnostic criteria — the version applied by the criteria checker above — are:1,2

Vestibular migraine (definite):

1. At least 5 episodes of vestibular symptoms of moderate or severe intensity, lasting 5 minutes to 72 hours;
2. Current or past history of migraine with or without aura according to ICHD criteria;
3. One or more migraine features with at least 50% of the vestibular episodes:
   • Headache with at least 2 of: unilateral location, pulsating quality, moderate or severe pain intensity, aggravation by routine physical activity;
   • Photophobia and phonophobia;
   • Visual aura.
4. Not better accounted for by another vestibular or ICHD diagnosis.

Probable vestibular migraine requires criteria 1 and 4 above plus EITHER criterion 2 (migraine history) OR criterion 3 (migraine features during attacks) — but not necessarily both.

Three operational points. First, criterion 3's photophobia-and-phonophobia requirement is conjunctive, not disjunctive — both must be present in an attack to count it. Second, the duration window (5 min–72 h) does heavy diagnostic work: attacks < 5 min suggest BPPV or TIA; > 72 h suggest AUVP, stroke, or PPPD. Third, the exclusion criterion is the rate-limiting step in most clinic practice: many vestibular migraine diagnoses are made only after Ménière's has been excluded by audiometry showing no persistent low-frequency SNHL.

The criteria's strictness is deliberate — the Bárány/IHS framers chose specificity over sensitivity to preserve the diagnosis's research value and to discourage diagnostic creep. In clinical practice this produces a meaningful population of patients with "vestibular migraine — clinically diagnosed but criteria not met". These patients are typically managed identically to definite or probable patients; the criteria threshold matters more for research enrolment, clinical trials, and billing than for treatment decisions.

Three classification subtleties deserve mention. Patients with migraine and BPPV occurring together is common (migraineurs have a higher prevalence of BPPV than non-migraineurs), and should be diagnosed as both — the BPPV is positional and short-lived, the migrainous vertigo is spontaneous and longer. Patients with both vestibular migraine and Ménière's disease (criteria for each met independently) are increasingly recognised — overlap rates of 30–40% in series that apply both sets of criteria. And benign recurrent vertigo of adults (Slater 1979) — recurrent vertigo without migraine history or features — is a probable vestibular migraine presentation by current criteria, and responds to migraine prophylaxis in most cases.

Vestibular migraine looks similar to a few other vestibular conditions, and getting the diagnosis right matters because the treatments differ. The most important distinction is from Ménière's disease. Both cause episodes of vertigo lasting hours, but Ménière's also causes hearing loss that builds up over time on one side, with ringing and a feeling of pressure in the same ear. Vestibular migraine doesn't damage hearing.

The Ménière's vs vestibular migraine differential is the most clinically consequential decision in episodic vestibular medicine. Symptom overlap is substantial — both cause spontaneous episodic vertigo lasting hours, both can produce tinnitus and aural fullness, both can be accompanied by headache. The cleanest discriminators are:

• Audiometric evidence of progressive low-frequency SNHLin Ménière's — present in essentially every confirmed case, absent in vestibular migraine. A series of audiograms across attacks is more informative than any single one.9
• Attack duration distribution: Ménière's attacks are 20 minutes to 12 hours; vestibular migraine attacks span 5 minutes to 72 hours. Attacks > 12 h favour vestibular migraine; < 20 min favours vestibular migraine or BPPV.
• Migraine features (headache, photo/phonophobia, aura) during attacks favour vestibular migraine; their absence does not exclude it.
• Caloric paresis on testingfavours Ménière's; vestibular migraine patients have largely normal caloric responses interictally.
• Response to migraine prophylaxisis a diagnostic test in itself — patients diagnosed as Ménière's who respond to propranolol or topiramate often have vestibular migraine.8

Other differentials worth considering. BPPV is positional and short-lived (< 1 min). Posterior circulation TIA is brief, recurrent, and accompanied by other neurological symptoms (visual deficits, dysarthria, ataxia) — particularly important to exclude in patients over 50 with vascular risk factors. Vestibular paroxysmia produces very brief (seconds) attacks attributed to neurovascular cross-compression and responds dramatically to carbamazepine. PPPD produces persistent rather than episodic dizziness, although it can be triggered by an initial vestibular migraine episode.

The clinical reality is that vestibular migraine and Ménière's coexist more often than the textbooks suggest. The Radtke 2012 follow-up of patients diagnosed with vestibular migraine found that 13% had also developed Ménière's criteria by 9-year follow-up, often with progressive low-frequency SNHL accumulating across attacks.7Migraine is roughly twice as prevalent in patients with Ménière's as in the general population, suggesting genuine pathophysiological overlap rather than coincidence.8

Two clinical strategies emerge. First, serial audiometry in any patient with episodic vertigo whose attacks last hours — the cleanest objective separator. Second, in ambiguous patients, a 3-month trial of migraine prophylaxis (propranolol or topiramate) is both diagnostic and therapeutic — responders are very likely to have a migrainous component to their disease regardless of whether they also have Ménière's. Avoid intratympanic gentamicin in any patient where vestibular migraine remains in the differential — irreversible vestibular destruction in a migraine patient with preserved peripheral function is a permanent iatrogenic injury.

Treatment has two parts. First, identifying and avoiding personal triggers — common ones include not enough sleep, stress, certain foods (caffeine, alcohol, aged cheeses, processed meats with nitrates), skipping meals, and hormonal changes. Some patients find that managing these alone reduces attacks substantially.

Second, if attacks are still frequent or disabling, preventive medication. The most effective are propranolol (a blood pressure pill that also prevents migraine) and topiramate (an anti-seizure medicine, also used for migraine). Both are taken daily; effects build up over weeks. Acute attacks can be managed with anti-nausea medications and rest in a dark, quiet room — much like a regular migraine.

Vestibular migraine management follows the same broad structure as headache migraine: lifestyle modification, acute attack management, and prophylaxis for frequent or severe disease.

Lifestyle and trigger management is first-line and well-evidenced for migraine generally, though specific to-vestibular-migraine evidence is weaker. A migraine diary identifies patient-specific triggers across sleep, stress, dietary, hormonal, environmental (weather changes, barometric pressure), and sensory (bright lights, strong odours) domains. Regularising sleep timing and ensuring adequate hydration have the strongest informal evidence base.

Acute attack treatment mirrors typical migraine: an antiemetic (prochlorperazine, metoclopramide, ondansetron) and a triptan if headache is part of the attack. Triptans have been tested specifically in vestibular migraine with limited evidence of benefit but no contraindication. Benzodiazepines provide rapid symptomatic relief and are reasonable for severe attacks, although chronic use should be avoided.

Prophylaxis is indicated for attacks occurring more than 2–3 times per month, attacks of long duration, or attacks substantially affecting quality of life. The 2025 systematic review by Almohammed and colleagues identified propranolol and topiramate as first-line agents, with moderate-quality evidence supporting both.11 The Salviz 2016 RCT compared propranolol vs venlafaxine head-to-head and found equivalent vestibular outcomes, with venlafaxine showing additional benefit in the emotional/depressive dimension.10

Anti-CGRP monoclonal antibodies (erenumab, fremanezumab, galcanezumab) — transformative for headache migraine prophylaxis — show promising early evidence in vestibular migraine but the trial data is limited. Reasonable for refractory cases where two or three conventional agents have failed.11

Three practical points in selecting and titrating prophylactic therapy. First, choose by patient profile — propranolol for the hypertensive patient with anxiety; topiramate for the overweight patient (modest weight loss benefit) but avoid in patients with cognitive concerns or kidney stones; venlafaxine where depressive symptoms accompany the vestibular symptoms; flunarizine (where available — not in the US) for refractory cases or patients who tolerate the others poorly.

Second, set realistic expectations. Most patients on effective prophylaxis see a 50% reduction in attack frequency, not complete remission. A trial period of at least 8–12 weeks at therapeutic dose is required before judging response — premature discontinuation is one of the most common reasons for treatment failure. The Dizziness Handicap Inventory (DHI) is a validated quantitative outcome that's worth tracking.

Third, vestibular rehabilitation is an underused adjunct. Patients with persistent inter-ictal imbalance, motion sensitivity, or PPPD-like features benefit from a structured VRT programme tailored to migraine sensitivity (avoiding aggressive habituation that can trigger attacks). The combination of pharmacological prophylaxis and graded VRT outperforms either alone in observational series.

Refractory vestibular migraine — failing > 3 conventional agents at adequate dose and duration — warrants reconsideration of the diagnosis before escalating to CGRP monoclonal antibodies or onabotulinumtoxinA. The most common reason for apparent treatment refractoriness in our experience is an alternative or overlapping diagnosis (PPPD, comorbid Ménière's, or medication-overuse from frequent triptan use).

Multiple sclerosis is a long-term disease in which the body's own immune system mistakenly attacks the insulating coverings of nerve fibres (myelin) in the brain and spinal cord. The damaged areas — called "plaques" or "lesions" — interfere with the nerves' ability to carry signals quickly. Symptoms depend on exactly where in the brain or spinal cord the lesions form, and can include weakness, numbness, vision problems, balance problems, and dizziness.

Vertigo is the first symptom in roughly 5% of patients, and occurs at some point in around 50% of patients across the disease course. The vertigo is usually a sign that the brainstem (the part of the brain that connects to the spinal cord) has been involved. In a young person — particularly a young woman — sudden vertigo with double vision is one of the classic presentations that triggers the suspicion of MS.

Modern MS treatment is highly effective. Once the diagnosis is established by MRI, lumbar puncture, and the clinical pattern, disease-modifying therapies can dramatically reduce relapse frequency and slow disability accumulation. Specific vestibular symptoms are managed in parallel — with medications during attacks, with vestibular rehabilitation between attacks.

Multiple sclerosis is the most common chronic inflammatory disease of the central nervous system, affecting 2.5–3 million people worldwide. Peak onset is in the third decade, with a female-to-male ratio of around 3:1 in relapsing-remitting disease.1 Geographical gradients — higher incidence at higher latitudes — and family-history clustering support a combination of genetic susceptibility and environmental triggers (vitamin D status, Epstein-Barr virus, smoking).

Vestibular and oculomotor symptoms are common across the disease course. Vertigo presents as the index symptom in approximately 5–10% of MS patients and occurs at some point in around 50%.5Three patterns recur: isolated central vestibular syndromes from brainstem plaques, oculomotor abnormalities (most distinctively INO), and chronic disequilibrium from cumulative cerebellar or brainstem involvement. Vertigo in MS is almost always central — peripheral vestibular symptoms in a known MS patient should prompt the same differential you would apply in any other patient (BPPV is common, vestibular migraine is common, Ménière's coexists occasionally).

The Lublin 2014 classification recognises four clinical phenotypes:4 relapsing-remitting MS (RRMS, the commonest at presentation), secondary progressive MS (SPMS, into which most RRMS eventually evolves), primary progressive MS (PPMS, with progressive disability from onset, typically older men), and the clinically isolated syndrome / radiologically isolated syndrome (CIS/RIS, the earliest detectable forms).

Vestibular MS deserves attention as a distinct presentation pattern because the diagnosis is often substantially delayed. A young patient presenting with acute vertigo lasting days — the classic acute vestibular syndrome — is often initially labelled vestibular neuritis, particularly if HINTS is applied carelessly. The clinical clues that should redirect toward central disease in this population: incomplete HINTS pattern (e.g. abnormal head impulse but with central nystagmus features), gait beyond what neuritis would explain, associated visual symptoms (optic neuritis preceding or accompanying), and atypical recovery. MRI with gadolinium contrast is the rate-limiting investigation — order it liberally in the young dizzy patient with any central features.6

Three brainstem regions are particularly important for vestibular MS. The medial longitudinal fasciculus — covered in detail in the next section — produces INO when demyelinated. The vestibular nuclei in the dorsolateral medulla produce central vestibular syndromes when involved, often with hemiataxia or sensory findings. The cerebellar peduncles — particularly the inferior cerebellar peduncle — produce ataxia and downbeat nystagmus when demyelinated. Plaques in any of these regions are commonly silent until specifically tested for, and benefit from a structured oculomotor examination (smooth pursuit, saccades, optokinetic response, head impulse, nystagmus, alignment) in every follow-up.

When you look to one side, both eyes have to move together — one outward, one inward. The brain coordinates this using a narrow bundle of nerve fibres connecting the two sides of the brainstem (the medial longitudinal fasciculus, or MLF). In MS, a patch of demyelination here breaks the coordination: the outward-moving eye works fine, but the inward-moving eye lags or fails to cross the midline. The patient often notices double vision when looking to one side, or describes "blurred vision".

The classic test that confirms this is brainstem-based rather than caused by a weak eye muscle: ask the patient to focus on a target moved close to their nose. The same eye that failed during lateral gaze now moves inward normally. This is preserved convergence — the diagnostic fingerprint of internuclear ophthalmoplegia.

INO results from a lesion of the medial longitudinal fasciculus, the heavily myelinated tract running through the dorsomedial brainstem tegmentum from the abducens nucleus (pons) to the contralateral oculomotor nucleus (midbrain). The MLF carries the signal that yokes the ipsilateral lateral rectus (abduction) to the contralateral medial rectus (adduction) during conjugate horizontal gaze.5

Clinical features of unilateral INO:

• Failure of adduction of the eye on the side of the lesion when looking toward the contralateral side.
• Abducting nystagmus in the contralateral eye on the same lateral gaze attempt.
• Preserved convergence — when the patient focuses on a near target, both eyes adduct normally. The medial rectus subnucleus is intact; the defect is in the connecting pathway, not the effector muscle.
• Skew deviation (vertical misalignment) is present in a substantial minority.
• Vertical pursuit and vestibular eye movements are often abnormal — looking for these adds diagnostic value.

The aetiological split by age is consistent across studies: in patients under 45, MS accounts for the majority of INO cases (often bilateral); in older patients, ischaemic stroke (typically unilateral) is the dominant cause.7 Bilateral INO in a young patient is almost pathognomonic for MS — particularly when accompanied by exotropia in primary gaze (the WEBINOsyndrome, "wall-eyed bilateral INO").

The bedside discrimination of INO has several refinements worth practising. Subtle INO is detected most reliably with rapid horizontal saccades rather than smooth pursuit — ask the patient to switch fixation rapidly between two widely separated targets (your two index fingers, or the tip of your nose and a distant point) and watch for slowness or fatigue of the adducting eye. Quantitative infrared oculography can detect adduction lag of just a few milliseconds — useful in clinical trials and in tracking longitudinal disease activity.6

Differential diagnosis of INO is dominated by two conditions and a handful of mimics. Demyelinating disease (MS) in young patients, often bilateral. Brainstem stroke (particularly small pontine lacunes) in older patients, typically unilateral. Less common but worth considering: Arnold-Chiari malformation, Wernicke encephalopathy, brainstem tumour (glioma, metastasis), Lyme disease, neurosyphilis, Whipple disease, and drug toxicity (tricyclics, opioids, phenothiazines). A myasthenic pseudo-INO can mimic demyelinating INO and improves with edrophonium or rest — worth screening with acetylcholine-receptor antibodies in the right clinical context.5

The functional consequences of INO vary. Some patients report only intermittent diplopia or vague visual blurring; others have disabling oscillopsia when reading or driving. Treatment of the underlying MS with high-dose corticosteroids in acute exacerbations often produces substantial improvement; chronic INO can be managed symptomatically with Fresnel prisms or, where necessary, strabismus surgery. The persistence or recovery of an INO is a useful clinical marker of disease activity that can be followed across visits.

Beyond internuclear ophthalmoplegia, MS can cause several other patterns of dizziness and abnormal eye movements, depending on which part of the brainstem or cerebellum is affected. Most of these are central — meaning the problem is in the brain rather than the inner ear — and can usually be distinguished from inner-ear causes by careful bedside examination and MRI.

Beyond INO, the vestibular and oculomotor features of MS cluster in three main groups:

• Central vestibular syndromes: acute or subacute central vertigo from plaques in or around the vestibular nuclei. The HINTS examination typically shows one or more central features (normal head impulse, direction-changing or vertical nystagmus, skew). MRI with gadolinium shows the enhancing plaque, often in the dorsolateral medulla or middle cerebellar peduncle.
• Central nystagmus patterns: downbeat nystagmus (from cervicomedullary or vestibulocerebellar plaques), pendular nystagmus (characteristic of MS, often asymmetric between the eyes, may improve with memantine or gabapentin), periodic alternating nystagmus, and gaze-evoked nystagmus.
• Cerebellar findings: truncal ataxia, limb ataxia, scanning dysarthria, saccadic dysmetria — the classic Charcot triad of nystagmus, intention tremor, and scanning speech describes late-stage cerebellar MS.

Optic neuritis is the most common cranial-nerve manifestation of MS and may precede or accompany vestibular symptoms. Recognising optic neuritis in a dizzy patient is an important diagnostic clue — and the McDonald 2024 criteria now formally include the optic nerve as a fifth topographic site for dissemination in space, reflecting its central role in early MS diagnosis.3

Two oculomotor features deserve specific clinician-level attention beyond INO. Pendular nystagmus is a distinctive MS-associated nystagmus — typically a small-amplitude, sinusoidal oscillation that is often asymmetric or even monocular, sometimes with vertical and horizontal components dissociated between the eyes. It localises to brainstem-cerebellar circuits and can respond to gabapentin or memantine. Its presence is strongly suggestive of MS in the appropriate clinical context.6

Saccadic dysmetria and saccadic pursuit are subtle but highly localising. Hypermetric saccades (overshoot followed by corrective saccade) point to cerebellar involvement. Saccadic intrusion into smooth pursuit — a pursuit movement broken up by small saccades — is one of the earliest detectable oculomotor signs of MS and is captured well by quantitative eye-tracking.6

Vestibular MS deserves a separate consideration for rehabilitation. Patients with persistent central vestibular deficits — particularly cerebellar — often respond poorly to standard peripheral-vestibular rehabilitation. A targeted programme emphasising gaze stabilisation, balance training in progressively challenging contexts, and habituation to motion-provoking stimuli is more effective. Coordination between the MS neurologist and a vestibular physiotherapist with experience in central vestibular disorders is the practical optimum.

The diagnosis of MS combines the pattern of clinical episodes, MRI evidence of multiple lesions in characteristic locations, and sometimes a spinal-fluid test (lumbar puncture) showing immune-system activity in the brain. The current rule (the "McDonald criteria", most recently updated in 2024) allows a diagnosis to be made earlier than previously possible, because effective treatment started early protects the brain.

Treatment has three parts: treating acute flare-ups (usually with a short course of high-dose steroids), preventing future flare-ups with disease-modifying medicines, and managing symptoms — including specific management of vertigo, double vision, and balance difficulties.

The McDonald criteria, originally published in 2001 and most recently revised in 2024, are the operational framework for MS diagnosis. The core principle is demonstrating dissemination in space (lesions in multiple CNS locations) and dissemination in time (lesions developing or symptoms occurring at multiple time points). The 2017 revisions allowed CSF oligoclonal bands to substitute for dissemination in time.2

The 2024 revisions — published in Lancet Neurology in September 2025 — introduce several important updates:3

• Optic nerve added as a fifth topographic site for dissemination in space (alongside periventricular, cortical/juxtacortical, infratentorial, and spinal locations).
• Dissemination in time may be waived in patients fulfilling dissemination in space when additional biomarkers (CSF kappa free light chains, the central vein sign, or paramagnetic rim lesions) are present — enabling earlier diagnosis from a single MRI in many cases.
• Radiologically isolated syndrome (incidental MS-like lesions on MRI in asymptomatic patients) can now be formally diagnosed as MS under specific criteria.
• Stricter thresholds for patients over 50 and those with vascular risk factors, mitigating the risk of misdiagnosis from age-related white-matter changes.

Management has three components. Acute relapse is treated with high-dose intravenous methylprednisolone (typically 1g daily for 3–5 days); plasma exchange is reserved for severe steroid-refractory attacks. Disease-modifying therapy is stratified by disease activity and patient preference, ranging from injectable platform agents (interferon-β, glatiramer) through oral agents (teriflunomide, dimethyl fumarate, fingolimod, siponimod, ozanimod) to monoclonal antibody infusions (natalizumab, ocrelizumab, ofatumumab, rituximab, alemtuzumab) — progressively more effective and more immunosuppressive.8 Symptomatic management targets specific deficits including spasticity (baclofen, tizanidine), neuropathic pain (gabapentin, pregabalin), fatigue (modafinil, amantadine), bladder dysfunction, and the vestibular and oculomotor symptoms discussed in this module.

Two diagnostic mistakes deserve specific attention. The first is over-diagnosis of MS in patients with non-specific white-matter changes on MRI — particularly older patients with vascular risk factors, in whom small vessel disease produces lesions that meet location criteria but not biological MS criteria. The 2024 revisions explicitly address this with stricter thresholds in older and vascular-risk populations.3 The second is delayed diagnosis in patients presenting predominantly with vestibular symptoms — a young patient with isolated acute vertigo, no other neurological signs, and a normal CT is often discharged with a label of vestibular neuritis. Order MRI with gadolinium liberally in young dizzy patients, particularly when HINTS is incomplete or shows mixed features.

For vestibular-MS symptom management specifically: acute central vertigo from active demyelination responds to high-dose corticosteroids; chronic positional imbalance benefits from vestibular rehabilitation tailored to central deficits; persistent pendular nystagmus may respond to memantine or gabapentin; chronic INO with disabling diplopia can be managed with Fresnel prisms or strabismus surgery in selected cases. The neurology and neuro-ophthalmology services should remain engaged throughout — vestibular medicine in MS is best practised as a multidisciplinary endeavour.

Paediatric MS deserves brief mention as a distinct population: it accounts for approximately 3–5% of MS cases, presents with a higher relapse rate but slower disability accumulation than adult-onset disease, and requires age-appropriate disease-modifying therapy.9

The cerebellum is the "little brain" sitting under the back of the skull. Its job is to fine-tune movement: when you reach for a cup, the cerebellum makes sure your hand stops at the cup rather than past it. When you walk, it makes sure each step lands smoothly under your centre of gravity. When you turn your head, it makes sure your eyes stay locked on the world.

A damaged cerebellum makes movements clumsy. Reaching becomes wobbly (intention tremor). Walking becomes wide-based and unsteady (ataxia, from the Greek for "without order"). Speech becomes slurred and broken (dysarthria). And the eyes develop characteristic wandering patterns that a trained clinician can recognise.

Cerebellar disease can present with vertigo, but the vertigo is usually less dramatic than the unsteadiness — patients feel off-balance even when sitting still. This is the key clinical difference from inner-ear vertigo, where patients often feel fine when still and worse with movement.

The causes are diverse: inherited conditions that run in families, long-term alcohol use, autoimmune diseases, side effects of cancer (paraneoplastic), and reactions to certain drugs. Some cerebellar disorders are treatable; many are not, but supportive care and rehabilitation help most patients live well.

The cerebellum is functionally divided into three zones, each connected to a different brain network and producing different deficits when injured. The vestibulocerebellum (flocculus, nodulus, parts of the uvula) calibrates the vestibulo-ocular reflex and is the substrate of central vestibular compensation; lesions here produce ocular abnormalities such as downbeat nystagmus and impaired smooth pursuit. The spinocerebellum (vermis and paravermal hemispheres) controls trunk and limb coordination; lesions here produce truncal and gait ataxia, with the most florid ataxia coming from vermal injury. The cerebrocerebellum (lateral hemispheres) handles motor planning and supports cognitive and affective processing; lesions here produce limb dysmetria, dysarthria, and the cerebellar cognitive-affective syndrome.2,1

The clinical fingerprint of a cerebellar lesion has three pillars. Ataxia — wide-based gait, truncal instability, limb dysmetria, and intention tremor — is the core syndrome. Cerebellar dysarthria is scanning, irregular in rhythm, and explosive in volume. And the cerebellar nystagmus zoo — downbeat, gaze-evoked, rebound, periodic alternating, square-wave jerks, ocular flutter — provides eye-movement signatures that often allow more precise lesion localisation than imaging alone.12

Aetiologically, cerebellar disease falls into seven broad categories that every clinician should be able to rank-order by frequency in their own practice setting: vascular (covered in the posterior circulation stroke module), hereditary (spinocerebellar ataxias, Friedreich's, episodic ataxias), toxic (alcohol, phenytoin, lithium, chemotherapy), immune (MS, gluten ataxia, anti-GAD ataxia), paraneoplastic (anti-Yo, anti-Hu, anti-Tr, anti-Ri), structural (tumour, Chiari, abscess), and idiopathic late-onset cerebellar ataxia.

The cerebellar functional anatomy that matters at the bedside is finer-grained than the three-zone classification suggests. Within the vestibulocerebellum, the flocculus tunes the angular VOR gain and smooth pursuit — its failure produces gain-reduced or gain-inverted VOR and saccadic pursuit. The nodulus stores velocity information from the semicircular canals and controls the duration of post-rotational responses; its failure produces periodic alternating nystagmus (PAN) and abnormally prolonged post-rotatory responses. The uvulacontributes to tilt suppression of the VOR — when the uvula is injured, tilting the head 90° fails to reduce the duration of caloric or post-rotational responses (a finding sometimes called "dumping" failure).12

Spinocerebellar lesions produce different ataxias depending on their position. Anterior vermis lesions (classically alcoholic cerebellar degeneration) produce pronounced gait ataxia with relatively preserved limb coordination — the syndrome of the wide-based heel-toe-walking ex-drinker. Posterior vermislesions (medulloblastoma in children, paraneoplastic ataxia in adults) produce truncal ataxia with preserved gait per se but inability to sit unsupported. Lateral hemisphere lesions produce ipsilateral limb dysmetria with relatively spared gait — the classical "arm-on-the-side" ataxia of cerebellar stroke.

The clinical-anatomical correlations are not as crisp as temporal-bone–era textbooks suggest, but the framework still organises bedside interpretation. The most consequential principle: any patient with disproportionate gait instability relative to the severity of vertigo needs cerebellar localisation considered, even when the head impulse is abnormal. Concurrent peripheral and central pathology (e.g. PICA-territory infarct adjacent to a labyrinthine infarct) is well-described and easily missed.3

When the cerebellum is damaged, the eyes do strange things. The brain's normal "hold steady" instruction to the eye muscles breaks down, and the eyes drift in stereotyped patterns. A trained clinician can often guess where the lesion is just by watching the eyes — long before imaging is available.

Some patterns are highly specific to cerebellar disease: eyes that beat downward when the patient looks straight ahead (downbeat nystagmus); eyes that wander rhythmically from side to side over several minutes (periodic alternating nystagmus); little jerky "hops" that occur even without provocation (square-wave jerks).

Other patterns are less specific but still suggest a central cause when seen with imbalance: nystagmus that changes direction depending on which way the patient looks (gaze-evoked nystagmus); nystagmus that briefly reverses when the eyes return to the centre after sustained lateral gaze (rebound nystagmus).

Downbeat nystagmus is the most common central nystagmus, and one of the most localising. The fast phase beats vertically downward, typically maximal in downward and lateral gaze. The mechanism is loss of cerebellar inhibition of the anterior semicircular canal pathways, allowing unopposed downward drive. The largest case series (n = 117) found the commonest causes to be cerebellar degeneration (40%), MS (10%), drugs (lithium, phenytoin, carbamazepine; 10%), and Chiari malformation (10%); a significant minority remained idiopathic.13

Periodic alternating nystagmus (PAN) is rare but pathognomonic of cerebellar (specifically nodular) pathology. The horizontal nystagmus changes direction every 90–120 seconds — right-beating for two minutes, briefly null, then left-beating for two minutes, and so on. Almost always associated with cerebellar degeneration; baclofen is the established symptomatic treatment.

Gaze-evoked nystagmus appears or worsens on eccentric gaze, with the fast phase in the direction of gaze. It reflects failure of the neural integrator — the brainstem-cerebellar circuit that holds the eyes in eccentric position. Cerebellar disease is one cause; others include drugs (alcohol, sedatives, anti-epileptics), myasthenia, and brainstem lesions.

Square-wave jerks are small horizontal saccades (typically 0.5–5°) that move the eyes off fixation and bring them back after a brief intersaccadic interval (~200 ms). A few square-wave jerks per minute are normal; frequent square-wave jerks (≥10/min) indicate cerebellar pathology or, less commonly, progressive supranuclear palsy. Ocular flutter and opsoclonus are pathological extensions of the same phenomenon — back-to-back saccades without an intersaccadic interval, occurring in one (flutter) or all (opsoclonus) directions. Opsoclonus in children classically signals neuroblastoma; in adults, paraneoplastic syndromes or post-infectious encephalitis.12

The cerebellar nystagmus repertoire functions, at the bedside, as a topographic lesion atlas. Floccular lesions produce gaze-evoked nystagmus, downbeat nystagmus, and impaired smooth pursuit and VOR cancellation. Nodular lesions produce periodic alternating nystagmus and failure of velocity-storage tilt suppression. Anterior vermis involvement classically yields gait ataxia with relatively normal eye movements (since the ocular control circuits sit posteriorly).

Distinguishing cerebellar from peripheral vertigo at the bedside rests on three independent observations. Direction-changing nystagmus on gaze is the strongest single sign — peripheral nystagmus respects Alexander's law (increases in the direction of fast phase, decreases in the direction of slow phase) but always beats in the same direction. A normal head impulse in a patient with persistent vertigo is central until proven otherwise — the basis of the HINTS exam. And truncal instability disproportionate to the degree of vertigo — the patient who cannot sit unsupported but reports only modest dizziness — is a cerebellar finding until proven otherwise.12

Symptomatic pharmacotherapy for cerebellar oculomotor signs is limited but real. 4-aminopyridine and 3,4-diaminopyridine potentiate Purkinje-cell firing and reduce downbeat nystagmus amplitude in many patients (number-needed-to-treat approximately 3 for symptomatic relief). Baclofen reliably suppresses periodic alternating nystagmus. Acetyl-DL-leucine has been tested in Niemann-Pick type C and idiopathic cerebellar ataxias with modest but reproducible benefits on scale outcomes.14

Strokes are the most acute cause of cerebellar trouble (covered in the posterior circulation module). The rest of this module covers the slow burners — diseases that produce gradual ataxia over months to years rather than the abrupt onset of vascular events.

The big categories: inherited (the family-history ataxias), toxic (chronic alcohol, certain medications), immune (an autoimmune attack on cerebellar cells), paraneoplastic (caused by an undiagnosed cancer elsewhere), and structural (tumours, Chiari malformation of the brainstem). Each category has its own clinical fingerprint and investigation pathway.

Hereditary ataxiasdivide into autosomal dominant (spinocerebellar ataxias, SCAs — over 40 subtypes described, most expanding-repeat or point mutations in cerebellar-relevant genes) and autosomal recessive (Friedreich's being the prototype, with ataxia plus cardiomyopathy and diabetes from GAA expansion in frataxin). SCA1, 2, 3 (Machado-Joseph), and 6 account for roughly 60% of dominant cases worldwide; subtype prevalence varies dramatically by population.4,5 Onset is typically in the third to sixth decade for SCAs and the first to second decade for Friedreich's.6

Episodic ataxias (EA1, EA2) are channelopathies producing attacks of cerebellar dysfunction lasting minutes (EA1, KCNA1 potassium-channel mutation) to hours (EA2, CACNA1A calcium-channel mutation, allelic with familial hemiplegic migraine type 1). EA2 responds dramatically to acetazolamide — a treatable cause not to miss.

Alcoholic cerebellar degeneration is the classical chronic-toxic ataxia: anterior vermis-predominant atrophy after years of heavy intake, producing gait ataxia with relatively preserved limb function. Originally described by Victor and Adams in 1959 as a distinct entity separate from Wernicke-Korsakoff syndrome.7 Modern data complicate the simple toxicity story — concurrent nutritional deficiency, gluten sensitivity, and age-related cerebellar atrophy may all contribute.8

Drug-induced ataxia is reversible and under-recognised. Phenytoin and carbamazepine at toxic levels reliably produce gait ataxia and downbeat nystagmus; lithium even at therapeutic levels can cause persistent cerebellar signs; chemotherapy agents (cytarabine, 5-fluorouracil, oxaliplatin) commonly produce acute or subacute cerebellar syndromes; amiodarone is an emerging cause. Stop the drug; watch for recovery.16

Paraneoplastic cerebellar degeneration is an immune-mediated cerebellar syndrome triggered by an occult cancer. Anti-Yo (ovarian, breast), anti-Hu (small-cell lung), anti-Tr (Hodgkin lymphoma), and anti-Ri (breast, lung) are the canonical antibodies. Onset is typically subacute over weeks; the cerebellar syndrome often precedes the cancer diagnosis. The 2021 PNS-Care criteria standardise diagnosis.10,11

Immune cerebellar ataxiain the absence of a known cancer includes anti-GAD-associated cerebellar ataxia (often with diabetes), gluten ataxia (anti-TG6, anti-gliadin; sometimes responds to gluten-free diet), Hashimoto's encephalopathy with cerebellar features, and isolated postinfectious cerebellitis (mostly paediatric, mostly self-limiting).9

The diagnostic algorithm in a new ataxia consultation is age-stratified. Acute or subacute onset (days to weeks) implies vascular, immune, paraneoplastic, toxic, or post-infectious — urgent imaging, broad bloods including CSF, paraneoplastic panel, and consideration of cancer screening. Chronic progressive ataxia (months to years) implies hereditary, alcoholic, or degenerative — extensive family history-taking, alcohol history, MRI to characterise atrophy pattern, and a graded approach to genetic testing starting with the regionally commonest SCAs.

Subtle features sharpen the differential. Saccadic slowing with normal range is highly suggestive of SCA2 (the most common SCA worldwide and the only SCA with cardinal saccadic slowing). Tendon areflexia plus Babinski signis the Friedreich's combination — a phenotype with virtually no other differential. Square-wave intrusions plus opsoclonus in a previously well adult should trigger paraneoplastic investigation regardless of cancer screening results — the cancer can lag the neurological syndrome by months.15,11

Management beyond aetiology-specific therapy (acetazolamide for EA2, gluten-free diet for gluten ataxia, immunotherapy for paraneoplastic disease, drug withdrawal for toxic ataxia) rests on three pillars: cerebellar rehabilitation with intensive coordination training and balance work (multiple trials demonstrate durable benefit on scale outcomes), symptomatic pharmacotherapy for downbeat nystagmus (aminopyridines) and ataxia (acetyl-DL-leucine in some indications),14 and quality-of-life management — speech therapy for dysarthria, swallow assessment for advancing disease, mobility aids selected to compensate rather than substitute, and psychological support.

The cerebellar cognitive-affective syndrome (CCAS) described by Schmahmann should be screened for in every patient with significant cerebellar disease — the executive, visuospatial, linguistic, and affective disturbances of CCAS are frequently the most disabling aspect of cerebellar pathology, frequently the most amenable to rehabilitation, and frequently missed when attention focuses on motor signs alone.1

Persistent postural-perceptual dizziness (PPPD) is a condition in which dizziness, unsteadiness, or a non-spinning sensation persists for three months or more, even after the original problem that caused it has resolved. Most patients describe feeling unsteady when standing or walking, disorientated in busy visual environments (supermarkets, crowded streets, watching screens), and worse with movement — particularly riding in cars, looking up at shelves, or turning their head quickly.

Crucially, PPPD is not "all in the head" in any dismissive sense. It's a recognised disorder of how the brain processes balance signals, often triggered by a definite event such as an inner-ear infection, a panic attack, or a mild head injury. The patient's balance system gets stuck in a hypervigilant state — like a smoke detector that won't reset after a small fire.

The good news is that PPPD responds to treatment: a combination of specialised physiotherapy (vestibular rehabilitation), a talking therapy called cognitive behavioural therapy (CBT), and sometimes a low dose of an antidepressant medicine such as sertraline. The antidepressant isn't used because the patient is depressed but because these medicines also help reset the balance system. Recovery is usually gradual but substantial.

PPPD is a functional vestibular disorder formally defined by the Bárány Society in 2017.1 The 2017 criteria unified four previously separate clinical entities — phobic postural vertigo (Brandt & Dieterich, 1986), space-motion discomfort, visual vertigo, and chronic subjective dizziness — under a single diagnostic framework grounded in a common mechanism: maladaptive plastic changes in postural control and visual-vestibular processing.3,2

The disorder is functional in the modern neurological sense — meaning that it produces real, measurable changes in brain function (the Indovina 2015 fMRI work showed insular hypoconnectivity in PPPD patients during vestibular stimulation) without an identifiable structural lesion.7"Functional" in this connotation explicitly does not mean psychogenic or psychosomatic; it means a problem with how the system is operating rather than with the hardware.

Epidemiologically, PPPD is precipitated by a definable event in essentially every case. The precipitant distribution is approximately:1,2

• Peripheral vestibular conditions (BPPV, vestibular neuritis, Ménière's) — 25%
• Vestibular migraine — 15–20%
• Psychiatric conditions (panic attack, generalised anxiety) — 15–25%
• Mild traumatic brain injury — 10–15%
• Dysautonomia or other medical illness — 5–10%
• Multiple sequential precipitants — disproportionately likely to develop PPPD

The Gabacorta 2022 finding that 22% of patients with two or more episodic vestibular conditions develop PPPD, versus 4% of those with a single condition, suggests that cumulative vestibular insults predispose to the maladaptive plasticity that underlies the disorder.6

The clinical importance of PPPD lies in recognising that many patients with apparently "treatment-refractory" vestibular disorders are actually presenting with two conditions: the original peripheral or central vestibulopathy that triggered PPPD, and the PPPD itself. Continued vestibular suppressants, repeated investigations, and escalating symptom-focused interventions worsen rather than help — they reinforce the maladaptive postural and visual-dependent strategies that underlie the chronic symptoms.

Two corollaries matter operationally. First, vestibular suppressant medications (benzodiazepines, meclizine, prochlorperazine, betahistine) are useful in the acute phase of vestibular illness but are actively harmful in the chronic phase — they prevent central compensation, perpetuate visual dependence, and confound the natural resolution of vestibular symptoms. Many PPPD patients arrive in tertiary clinics on chronic vestibular suppressants prescribed years earlier; withdrawal is often the first therapeutic step.2

Second, repeated negative investigations — multiple MRIs, repeated audiometry, repeated VNG — reinforce rather than reassure. The clinician's task is to make a positive diagnosis of PPPD, explain the model to the patient (the "balance system stuck in alarm mode" metaphor works in clinic), and direct the patient to evidence-based intervention. Investigation proportional to the clinical question is fine; investigation as anxiolytic is iatrogenic.

To meet the formal criteria for PPPD, a patient needs all five of the following:

• Dizziness, unsteadiness, or non-spinning vertigo on most days for 3 months or more;
• Symptoms made worse by all three of: standing up, moving (the patient or what they are looking at), and busy or moving visual environments;
• A precipitating event — usually a vestibular illness, a head injury, a panic attack, or another medical problem;
• Significant distress or impact on daily life;
• No other condition that better explains the symptoms.

If a patient has all the features but only for six weeks, they don't yet meet the criteria — but they should be identified as at-risk and started on early intervention to prevent crystallisation into the full disorder.

The five PPPD criteria (Staab 2017) — exactly as applied by the criteria checker above — are:1

1. A. One or more symptoms of dizziness, unsteadiness, or non-spinning vertigo are present on most days (≥ 15 days/month) for ≥ 3 months. Symptoms typically last for prolonged periods (hours to all day) but may wax and wane in severity.
2. B. Persistent symptoms occur without specific provocation, but are exacerbated by:
   1. upright posture,
   2. active or passive motion without regard to direction or position, and
   3. exposure to moving visual stimuli or complex visual patterns.
   All three exacerbators must be present.
3. C. The disorder is precipitated by conditions that cause vertigo, unsteadiness, dizziness, or problems with balance, including acute, episodic, or chronic vestibular syndromes; other neurological or medical illnesses; or psychological distress.
4. D. Symptoms cause significant distress or functional impairment.
5. E. Symptoms are not better accounted for by another disease or disorder.

Three operational points. First, criterion B's three exacerbators are conjunctive— all three are required. Missing the visual component (motion-rich environments like supermarkets, scrolling on phones, busy patterns) is a common reason for missed diagnosis. Second, the duration threshold of 3 months separates PPPD from sub-threshold or evolving forms — patients with the right phenotype but only 4–8 weeks of symptoms benefit from early intervention but do not yet have the diagnosis. Third, the exclusion criterion is permissive — PPPD can and does coexist with vestibular migraine, BPPV, or Ménière's; co-diagnose where criteria for both are met.2

Three diagnostic refinements at the clinician level. First, the "motion intolerance" component of criterion B is broader than many clinicians initially appreciate — it includes self-motion (walking, head movement) and external motion (passenger in a car, escalators, watching moving objects). Asking specifically "is being a passenger in a car worse than driving?" often unlocks the visual-vestibular conflict pattern.

Second, visual dependence is the most diagnostically distinctive feature of PPPD and is captured by the visual-stimuli component of criterion B. The classic clinical questions: "is it worse in supermarkets?", "is it worse on patterned carpets?", "does scrolling on your phone make it worse?", "is it worse with crowds walking past you?". Affirmatives to two or more of these strongly suggest visual dependence and PPPD even before applying the full criteria framework.

Third, the criterion E exclusion is more permissive than criterion E in the vestibular migraine criteria. PPPD frequently coexists with an active vestibular disorder — vestibular migraine is the most common co-diagnosis, with rates of 30–50% in series that apply both criteria sets. Persistent symptoms that fit PPPD between vestibular migraine attacks are an additional disease, not a recharacterisation of the migraine. Both should be diagnosed and both should be treated.6

After an inner-ear or balance problem, the brain has to adapt to the new information it's getting. Normally, this adaptation happens automatically within weeks. In PPPD, the adaptation gets stuck in a particular pattern: the brain becomes over-reliant on what the eyes are seeing to maintain balance, and it stays in "high alert" even when there's no longer a real problem.

This explains the characteristic symptom pattern. Standing up activates the balance system, so it's worse than sitting. Movement — your own or things you're looking at — challenges the system, so it's worse. Busy visual environments overload the visual processing the brain has come to depend on, so they're worst of all.

The mechanism of PPPD is best understood as a maladaptive plastic change in postural control and multisensory integration following a precipitating vestibular, neurological, or psychiatric insult. Three interacting changes characterise the chronic state:

• Visual dependence: down-weighting of vestibular and proprioceptive inputs in favour of visual inputs for postural control. Adaptive in the acute phase (when vestibular signals are unreliable), maladaptive when persisted.
• High postural control gain: stiff, ankle-strategy postural control with increased muscle co-contraction. Visible on posturography as increased sway amplitude with paradoxical hypersensitivity to perturbation.
• Hypervigilance and threat appraisal: the brain's threat-evaluation systems (insula, amygdala) remain activated by vestibular signals long after the original threat has resolved. The fMRI evidence supports altered functional connectivity between the insula and central vestibular networks.7

The precipitating events that trigger this maladaptation fall into four broad categories with roughly equal contribution to the population of PPPD patients:1,2

• Vestibular: BPPV, vestibular neuritis, Ménière's, vestibular migraine — the trigger is a genuine vestibular event that resolves but leaves PPPD behind.
• Neurological: mild traumatic brain injury, stroke (particularly small posterior fossa events), demyelinating disease.
• Medical: orthostatic intolerance, POTS, autonomic dysfunction.
• Psychiatric: panic attack, generalised anxiety disorder, depressive episode — particularly when accompanied by autonomic symptoms that the patient experiences as dizziness.

Common features predicting evolution from acute illness to PPPD include high pre-morbid trait anxiety, prior anxiety or depressive disorders, female sex, and — most strongly — multiple sequential vestibular insults. The clinical implication is that early intervention in high-risk patients (those with these predictors and ongoing symptoms beyond 4–6 weeks post-trigger) may prevent crystallisation into full PPPD.6

The functional-not-psychogenic distinction deserves careful clinician-level framing. PPPD is mechanistically distinct from somatisation, conversion disorder, and anxiety-driven dizziness, even though it can coexist with all three and shares some risk factors. The functional model — maladaptive plasticity in a real neurological system, demonstrable on fMRI and posturography — gives the clinician a positive, mechanistic explanation to offer the patient: the brain has learned the wrong response to a real triggering event, and the treatment is unlearning, not symptom suppression.2

This framing matters for two reasons. First, it makes the rehabilitation rationale comprehensible to the patient — vestibular rehabilitation works by retraining the maladaptive postural and visual-dependent strategies, not by "exercising the inner ear". Second, it de-stigmatises the diagnosis. Patients who have been told their symptoms are anxiety-driven or psychogenic often arrive defensive and treatment-resistant; reframing the problem as a learnable-and-unlearnable functional disorder unlocks engagement with therapy.

PPPD treatment combines three approaches, usually in parallel rather than sequentially. Vestibular rehabilitation is a specialised physiotherapy programme that gradually exposes the patient to the movements and visual environments that trigger their symptoms, in a structured way that helps the brain relearn proper balance processing. Cognitive behavioural therapy (CBT) helps the patient understand the condition, reduce avoidance behaviours, and manage the anxiety that often accompanies chronic dizziness. Medication — usually a low-dose SSRI such as sertraline — helps reset the hyperactive balance-threat circuits and can be very effective, particularly when combined with the other two approaches.

Recovery is usually substantial but gradual — most patients see meaningful improvement over 3–6 months of consistent treatment. The condition is not a permanent disability; with the right intervention most patients return to good function, even if some residual sensitivity to busy environments persists.

PPPD treatment is multimodal, with the strongest evidence supporting a combination of vestibular rehabilitation therapy (VRT), cognitive behavioural therapy (CBT), and serotonergic pharmacotherapy. Each component targets a different facet of the maladaptive plasticity.

Vestibular rehabilitation is the cornerstone, with the longest evidence base. PPPD-targeted VRT differs from standard VRT in its progressive graded-exposure structure: starting from positions and environments the patient currently tolerates, gradually increasing motion and visual complexity, and explicitly including provocative head movements and visual-motion exposure. Standard one-size-fits-all VRT can paradoxically worsen PPPD by triggering threat-circuit responses; a therapist experienced in PPPD-specific approaches is materially more effective than a general vestibular physiotherapist.8

Cognitive behavioural therapy addresses the catastrophising, threat-monitoring, and avoidance behaviours that maintain PPPD. The Yu 2018 RCT compared sertraline alone with sertraline plus CBT and found substantially greater improvement in dizziness handicap (DHI), anxiety (HARS), and depression (HDRS) scores in the combined group — supporting CBT as a value-adding augmentation rather than a standalone treatment.4

Serotonergic pharmacotherapy uses low-dose SSRIs or SNRIs as first-line agents. Sertraline (50–200 mg/day), escitalopram (10–20 mg/day), and venlafaxine (75 mg/day) all have supporting evidence; paroxetine has the longest track record in this indication (Horii et al. 2007).5,4 The mechanism is believed to be modulation of serotonergic projections to the vestibular nuclei and to the insular-amygdalar threat-appraisal network, rather than a primary antidepressant effect. Treatment is typically continued for 12 months minimum after symptom resolution to allow stable replasticisation.

Three things to avoid in PPPD management. Vestibular suppressants (meclizine, cinnarizine, prochlorperazine, betahistine, benzodiazepines) prevent central compensation and perpetuate symptoms — withdraw them during initial management. Repeat investigations (multiple MRIs, repeated VNG, repeated audiometry) without clear new indication reinforce the threat-monitoring pattern. Symptomatic escalation — stronger antiemetics, sedatives, surgery for incidental findings — moves the patient in the wrong direction.2

Three clinician-level refinements on the standard management framework. First, the order of initiation matters. Starting all three modalities simultaneously is impractical for most patients; the pragmatic sequence is (1) clear diagnostic explanation and rationale, (2) referral to a PPPD-experienced vestibular physiotherapist, (3) commencement of SSRI at low dose with slow titration, (4) addition of CBT when the patient is engaged and ready. Setting expectations explicitly — 3–6 months to substantial improvement, not 3–6 weeks — protects against premature discontinuation, the most common reason for apparent treatment failure.

Second, the SSRI choice and dose matter. Start lower than standard depression doses — sertraline 25 mg daily for 5–7 days then 50 mg, escitalopram 5 mg then 10 mg — and titrate slowly. Patients with PPPD are often initially intolerant of activating side effects, which can mimic or worsen their dizziness in the first 1–2 weeks. Forewarning the patient about this and offering a slow titration improves persistence substantially.5

Third, refractory PPPD deserves a systematic re-look. The most common reasons for apparent treatment failure in our experience: an undiagnosed comorbid vestibular migraine that needs its own prophylactic treatment; persistent use of vestibular suppressants that perpetuate visual dependence; insufficient duration of VRT (less than 12 weeks at adequate intensity); and unrecognised comorbid generalised anxiety disorder that would itself benefit from anxiety-specific CBT rather than PPPD-focused CBT. Refractory PPPD warrants multidisciplinary review (neurology, vestibular physiotherapy, clinical psychology) before further pharmacological escalation.6

The most recent systematic review of conservative therapy in PPPD found that VRT or CBT combined with SSRI consistently outperformed single-modality treatment on the Dizziness Handicap Inventory, the Hospital Anxiety and Depression Scale, and the Hamilton scales — consolidating the multimodal approach as standard.6