Evolutionary Changes in the Cochleaand Labyrinth: Solving the Problem

of Sound Transmission to the BalanceOrgans of the Inner Ear

JOHN CAREY* AND NIVEE AMINDepartment of Otolaryngology-Head and Neck Surgery, Johns Hopkins University

School of Medicine, Baltimore, Maryland

ABSTRACTThis review article examines the evolutionary adaptations in the verte-

brate inner ear that allow selective activation of auditory or vestibular haircells, although both are housed in the same bony capsule. The problem ofseparating acoustic stimuli from the vestibular end organs in the inner ear hasrecently reemerged with the recognition of clinical conditions such as superiorcanal dehiscence syndrome and enlarged vestibular aqueduct syndrome. Inthese syndromes, anatomical defects in the otic capsule alter the functionalseparation of auditory and vestibular stimuli and lead to pathological activa-tion of vestibular reflexes in response to sound. This review demonstrates thatwhile the pars superior of the labyrinth (utricle and semicircular canals) hasremained fairly constant throughout evolution, the pars inferior (saccule andother otolith, macular, and auditory end organs) has seen considerable changeas many adaptations were made for the development of auditory function.Among these were a relatively rigid membranous labyrinth wall, a variablyrigid otic capsule, immersion of the membranous labyrinth in perilymph, aperilymphatic duct to channel acoustic pressure changes away from the ves-tibular organs, and different operating frequencies for vestibular versus audi-tory epithelia. Even in normal human ears, acoustic sensitivity of the labyrinthto loud clicks or tones is retained enough to be measured in a standard clinicaltest, the vestibular-evoked myogenic potential test. Anat Rec Part 288A:482–490, 2006. © 2006 Wiley-Liss, Inc.

There seems to be little ambiguity between hearingand equilibrium in everyday human experience. Yet allvertebrates have inherited a basic inner ear design thathouses the auditory and vestibular sensory epithelia ina common bony capsule, the labyrinth. Because of thisdesign, the potential for acoustic stimuli to act on thebalance sensors of the vestibular system is greater thancommonly appreciated. Recent findings from the clinicalrealm indicate that sound can have measurable andimportant effects on the vestibular sense organs in nor-mal humans. Moreover, in humans with certain defectsin the bony structure of the labyrinth, characteristicdisturbances of hearing and disabling disturbances ofbalance have been identified. These entities have re-newed interest in the structural components of the in-ner ear that subserve the functional separation of au-ditory and vestibular stimuli.

This article will review the evolution of the inner earwith attention to specializations that effectively separatesound and balance stimuli and direct their effects to theappropriate sensory structures. Two illustrative syn-dromes characterized by disruptions of hearing and bal-ance will also be described in which there is an abnormalopening in the labyrinth (superior semicircular canal de-

*Correspondence to: John Carey, Department of Otolaryngolo-gy-Head and Neck Surgery, Johns Hopkins University School ofMedicine, 601 North Caroline Street, Room 6255, Baltimore, MD21287. Fax: 410-955-0035. E-mail: jcarey@jhmi.edu

Received 29 December 2005; Accepted 29 December 2005DOI 10.1002/ar.a.20306Published online 20 March 2006 in Wiley InterScience(www.interscience.wiley.com).

THE ANATOMICAL RECORD PART A 288A:482–490 (2006)

© 2006 WILEY-LISS, INC.

hiscence syndrome) or an abnormally large size of a nor-mal opening (enlarged vestibular aqueduct syndrome). Fi-nally, the vestibular-evoked myogenic potential test, aclinical test based on the sound sensitivity of the normalsacculus, will be described.

Hair Cells, the Common Mechanosensors ofHearing and Balance

Both hearing and balance rely on the mechanosensoryhair cell. Hair cells are so named for tufts of stereociliathat project from their apical surfaces. Deflection of ste-reocilia is the common mechanism by which all hair cellstransduce mechanical forces (Fig. 1). Stereocilia within abundle are linked to one another by protein strands called“tip links” that span from the side of a taller stereociliumto the tip of its shorter neighbor in the array. The tip linksare believed to act as gating springs for mechanicallysensitive ion channels, meaning that the tip links literallytug at molecular gates in the stereocilia (Hudspeth, 1985;Pickles and Corey, 1992). These gates, which are cationchannels, open or close (or, more precisely, spend more orless time in the open state), depending on the direction in

which the stereocilia are deflected. When deflected in theopen or “on” direction, which is toward the tallest stereo-cilium, cations, including potassium ions from the potas-sium-rich endolymph, rush in through the gates, and themembrane potential of the hair cell becomes more positive(Fig. 1B–D). This in turn activates voltage-sensitive cal-cium channels at the basolateral aspect of the hair cell,and an influx of calcium leads to an increase in the releaseof excitatory neurotransmitters from hair cells. Theseneurotransmitters bind to and excite receptors at thepostsynaptic densities (psd) on the underlying primaryafferent neurons, increasing the firing rates of these neu-rons (Fig. 1D). Deflection of stereocilia in the “off” direc-tion decreases the firing rates of afferent neurons by de-creasing synaptic activity.

Mechanosensation, however, evolved well before haircells appeared. Examples include the mechanosensory bris-tles of the fruit fly Drosophila melanogaster, the microtubu-late mechanosensors of the nematode Caenorhabditis el-egans, and even the membrane tension-sensing proteins inEscherichia coli (Kernan and Zuker, 1995). Further up theevolutionary tree, the most primitive chordates, ascidians(sea squirts), developed mechanosensory cells in the adultforms that were grouped into gelatinous cupular organsstrikingly similar to those of the vertebrate acousticolatera-lis systems (Bone and Ryan, 1978). All of these invertebratemechanosensors are actually specialized neurons with cili-ated endings surrounded by supporting cells.

With the evolution of vertebrates came the appearanceof neural crest cells and neurogenic placodes (Shimeld andHolland, 2000). Neurogenic placodes are paired ectoder-mal thickenings in the developing head that give rise tosensory cells and support cells. Hair cells and the associ-ated supporting cells of the inner ear are derived from theotic placodes. Like the sensory cells of other placodes, haircells are not neurons themselves, but rather are connectedto the brain by separate afferent and efferent neurons.

The vertebrate hair cell that evolved has proved to be aremarkably effective mechanosensor for both hearing andbalance. The stereociliary transduction apparatus is ableto detect movements smaller than even those created byBrownian motion of the stereocilia themselves (Denk andWebb, 1992) and with latencies of only 10 �sec (Corey andHudspeth, 1979). An ensemble of basolateral K� channelskeeps the resting potential of the hair cell near �70 mV(Holt and Eatock, 1995). Variations in this ensemble of K�

channels may determine the frequency responses of dif-ferent hair cells (Rusch and Eatock, 1996). This is anespecially important flexibility in hair cells, as the fre-quencies of stereociliary motion relevant to equilibriumsensation range from 0 to 20 Hz (Grossman et al., 1988),while those that are relevant to hearing go up to tens ofkHz. Voltage-gated Ca2� channels supply the rapid influxof calcium needed to trigger neurotransmitter release(Fuchs and Evans, 1990). Finally, specialized ribbon syn-apses coordinate the rapid release of vesicles and evensimultaneous release of multiple vesicles to enable high-frequency transmission necessary for hearing high-fre-quency sound (Glowatzki and Fuchs, 2002).

The sensitive, fast, and flexible properties of hair cellsmay have reduced any evolutionary pressure to developcompletely separate organs for hearing and balance. Butthe molecular ancestry of inner ear development may alsoexplain why the hair cell serves this double duty in ver-tebrates. The array of transcription factors that governthe differentiation of hair cells and associated structures

Fig. 1. Sensory transduction by vestibular hair cells. A: At rest, thereis some baseline release of excitatory neurotransmitter from the hair cellonto the underlying afferent neuron ending. B: Hair cells are depolarizedwhen the stereocilia are deflected in the “on” direction (toward thekinocilium in black). C: This occurs because the stretched tip linksmechanically open cationic channels in the stereocilia membranes. Theinflux of potassium ions raises the hair cell’s membrane potential. D: Theincreased membrane potential activates voltage-sensitive calcium chan-nels in the basolateral membrane of the cell. Synaptic release of exci-tatory neurotransmitter increases, and receptors in the postsynaptic(psd) density on the afferent increase its membrane potential, which inturn increases afferent firing rate.

483SOUND TRANSMISSION TO BALANCE ORGANS

in vertebrates has remarkable homologues at almost ev-ery branch point to those that govern the development ofciliated mechanosensory neurons and their supportingstructures in invertebrates (Fritzsch and Beisel, 2001).The pressure to conserve these transcription factor fami-lies, including zinc-finger proteins, Delta/Notch signaling,and basic helix-loop-helix (bHLH) proteins, may havebeen very strong. Thus, expansion of ancient families oftranscription factors that once led to single invertebratemechanosensors may have destined the vertebrates to asingle solution as well.

EVOLUTION OF VERTEBRATE LABYRINTHAND ADVANTAGE OF MIXED EQUILIBRIUM

AND AUDITORY ROLESThe inner ear likely first appeared prior to the Ordovician

period, during which invertebrate sea-dwelling creatureswere the predominant life form (Baird, 1974). It has longbeen assumed that the inner ear developed as an infolding ofthe lateral line system of fishes (Wilson and Mattocks, 1897).This system consists of neuromasts, patches of hair cellslocated superficially or in fluid-filled canals along the lateralaspects of the head and body. The hair cells of the lateral linesense movement of water over the surface of the fish orpressure gradients, which develop between the ends of thelateral line canals. It seems logical that closure of thesecanals would form the inner ear, but more recent evidencesuggests that the inner ear and lateral line systems devel-oped independently and at different times from a more gen-eral superficial neuromast system (Maisey, 2001).

The labyrinth of extant cyclostomes may give some cluesas to the morphology of the early vertebrate labyrinth. Thisancestral inner ear was likely a simple parameningeal ves-icle that communicated with the surface of the head andincorporated a common macula, or patch of hair cells, withneuromast organs (Lewis et al., 1985a). The labyrinth of thehagfish consists of a simple torus with ampullae at twodilations in the wall (Lowenstein and Thornhill, 1970).These appear to be the equivalents of anterior and posteriorcanals. There is a single macula covering the medial andpart of the ventral wall in the base of the torus. Electrophys-iological recordings identified rotational and tilt responses,but not vibrational responses, perhaps corresponding to thelack of a saccule or macula neglecta. For the purposes of thisarticle, it is important to note that the otic vesicle in thehagfish is completely encapsulated except for the foraminafor the sensory nerves.

Lowenstein et al. (1968) demonstrated that the lampreyinner ear is also encased in cartilage except for the foram-ina for its sensory nerves. As in the hagfish, they foundthat in the lamprey the labyrinth has only two ampullaecorresponding to the vertical canals. However, in the lam-prey, they found that the ampullae communicate withciliated chambers that are unique among all the verte-brates and still mysterious in their function. They alsofound that the lamprey’s macula appeared more differen-tiated than that of the hagfish, and various subdivisionswere proposed that are analogous to the separate macularend organs in higher vertebrates. Robust responses tovibration were found in some macular units (Lowenstein,1970). Of note, the cyclostomes represent offshoots prior tothe gnathostomes in the vertebrate line (Fig. 2). It is thuspossible that the encapsulation of the otic vesicle that isrepresented in higher organisms is unrelated to thatfound in the cyclostomes (Lewis et al., 1985a).

In the sharks, skates, and rays (elasmobranches), theotic labyrinth is enclosed in cartilage for the most part(Lowenstein et al., 1964). However, surprisingly, the oticduct not only penetrates the cartilage but even opens tothe surface of the head. The semicircular canals and themajor parts of the otic labyrinth are more fully differenti-ated in the chondrichthyes than in the cyclostomes. Inchondrichthyes, the saccule assumes it role as the organi-zational center of the labyrinth. Its broad base is locatedin the ventromedial part of the capsule while its bodytapers to communicate with the otic duct on the dorsome-dial aspect of the capsule. Elasmobranches also have asmall lagena, another otolith organ that is a posteriorevagination of the saccule.

The osteichthyes contain an otic capsule that is primar-ily cartilaginous but incorporates bone in varying degrees(Lewis et al., 1985a). Parts of the labyrinth occupy bonyfossae rather than bony canals, and cartilage serves topartition the area of the semicircular canals. This area isthe pars superior of the labyrinth, and it remains rela-tively constant throughout the evolution from the gnatho-stomes onward. Pars superior consists of a single otolithorgan, the utricle, and three semicircular canals. The utri-cle has hair cells mostly oriented horizontally and sensi-tive to linear accelerations in the horizontal plane. Thesemicircular canals are inertial sensors of rotation.

In contrast to the constancy of pars superior, the parsinferior has many variations in the rest of the vertebrateline, and it appears that these variations are adaptations forauditory sensation (Lewis et al., 1985a) (Fig. 3). In variousteleost lines, for example, auditory sensitivity can be foundin the saccule, lagena, utricle, or a rarer end organ, themacula neglecta. In some of the teleost lines, the sacculeand/or lagena sense ambient pressure changes directly froma membrane-covered opening in the lateral otic capsule. Butin the ostariophysan fishes, an air bladder and coupled We-berian ossicles form a pressure-to-displacement transducer.Pressure changes in the ambient water produce displace-ment of perilymph surrounding the saccule. In amphibians,sensation of substrate and water vibration is retained in thesaccule, while sensation for airborne vibration is developedin two acoustical end organs, the amphibian papilla and thebasilar papilla. Acoustic sensitivity in reptiles is retained inpart in the saccule, while experimentation with and elonga-tion of the basilar papilla lead to a more elaborate auditorysensor with tonotopic organization. Elongation and coiling ofthe basilar papilla and the development of the cochlear ductand loss of the lagena produce the mammalian cochlea, andthe saccule becomes primarily a gravitoinertial sensor.

It appears, then, that there was not evolutionary pressureto separate acoustic and equilibrium sensation in the verte-brate ear. On the contrary, the natural experiments with thestructures of the pars inferior demonstrate a potential ad-vantage in keeping the full range of mechanosensory organsin a common otic capsule. As evolution exposed vertebratesto varying ambient conditions, changes could occur in theaccessory structures surrounding the hair cells of the innerear so that an end organ’s sensitivity could be adapted to thechanging needs of communication and motion.

SPECIALIZATIONS OF LABYRINTH TOPREVENT AMBIENT PRESSURE CHANGESFROM AFFECTING EQUILIBRIUM ORGANSThe hair cells responsible for equilibrium sensation are

not housed separately from those responsible for hearing

484 CAREY AND AMIN

or vibration sensation. However, several structural spe-cializations have evolved within the vertebrate labyrinththat might help minimize the effect of ambient pressurechanges on the balance organs. In the following sections,the evolution of these specializations will be considered.

A Relatively Inelastic Membranous LabyrinthOne such specialization may be the wall of the mem-

branous labyrinth itself. Derived from cartilage, it hasbeen assumed that this wall is nearly inelastic to thepressure changes caused by acoustic stimulation (Lewiset al., 1985b). However, this may not be true of portionsof the wall, such as that of the saccule, which demon-strates acoustical sensitivity in so many vertebrates.Indeed, investigators have successfully utilized piezo-electric indenters on the membranous canals to mimicthe endolymph displacement caused by head accelera-tion (Dickman et al., 1988). These experiments chal-lenge the notion of an inelastic wall in these portions ofthe membranous labyrinth. However, mechanical in-dentation forces in these experiments may be much

larger than those expected for sound stimuli that mightnormally reach the canals. The elasticity of this mem-branous canal wall has not been specifically measuredin the indentation experiments (Rabbitt et al., 1995).Finally, even if the wall of the membranous labyrinth issufficiently elastic to prevent sound from affecting thebalance organs under physiologic circumstances, it maynot, as we shall see, hold under pathologic circum-stances that shunt sound pressure toward the vestibu-lar end organs.

A Rigid Otic CapsuleA second specialization that may prevent ambient pres-

sure transmission to the balance organs is the encapsula-tion of the labyrinth in a rigid otic capsule. This featureappeared early in the evolution of vertebrates. The petro-myzont inner ear is completely encased in cartilage exceptfor the medial foramen for the acoustic nerve and associ-ated blood vessels (de Burlet, 1934). However, otic capsulerigidity was not uniformly retained in subsequent evolu-tion. The medial wall of the labyrinth is often open in both

Fig. 2. A simplified vertebrate evolutionary tree. Vestibular receptors likely evolved very early and gaverise independently to the lateral lines of fishes and the inner ear sensory epithelium. Throughout vertebrateevolution, the pars superior has been fairly well conserved, while considerable change has occurred in thepars inferior, mainly as adaptations for auditory sensation. Figure based on Myers (2001).

485SOUND TRANSMISSION TO BALANCE ORGANS

cartilaginous and bony fishes. In fact, the seahorses andpipefish lack any otic capsule, with their membranouslabyrinths simply suspended within their cranial cavities(de Burlet, 1934). Moreover, some cyclostomes have aninner ear that still communicates with the surface of thehead. Such a communication would clearly allow access ofambient pressure changes to the inner ear balance organs.It must be emphasized, however, that ambient pressurechanges with such arrangements will not produce en-dolymph flow across the vestibular sensory epithelia. Zeroor single openings in the labyrinth or arrangements inwhich equal pressures are exerted over all openings wouldnot create local pressure gradients or inappropriate en-dolymph flow across acceleration-sensing hair cells in theinner ear.

There may in fact be an advantage to the incompleteencapsulation of the labyrinth in fish. These aquatic ver-tebrates must deal with changing ambient water pres-sure, which occurs when changing depth. Openings in thelabyrinth that transmit ambient pressure changes providea mechanism to sense that depth. It has recently beendemonstrated that horizontal semicircular canal afferentneurons in the elasmobranch dogfish carry signals relatedto hydrostatic pressure (Fraser and Shelmerdine, 2002).Yamauchi et al. (2002) previously showed that the semi-circular canal ampulla in the toadfish is sensitive to dila-tional pressure in the absence of endolymph flow. In fact,the ambient pressure fluctuations expected for these fishin their natural range is 106 times greater than the dila-tional pressure threshold for their semicircular canal af-ferents. The vestibular end organs in fish may therefore beexcellent depth sensors in addition to acceleration detec-tors. Once again, the tremendously useful features of thehair cell appear to be adapted to multiple mechanosensoryfunctions.

The Surrounding PerilymphIf ambient pressure is transmitted to the vestibular

end organs for depth sensation, it must occur withoutcausing endolymph flow, or the animal will have aninappropriate sense of motion or change in orientation.Yamauchi et al. (2002) point out that the complete im-mersion of the membranous labyrinth in a continuousperilymphatic enclosure appears to be the solution tothis problem. The complete enclosure in a fluid mediumensures that pressure is equally distributed around themembranous labyrinth, preventing the development oflocal transmembrane dilational pressure gradients.

Fig. 3. Evolution of the perilymphatic duct system and pars inferior ofthe vertebrate inner ear. A. The amphibian inner ear has a simple per-ilymphatic duct (pd) that begins with the perilymphatic sac (ps). Motionof the columella in the fenestra vestibuli (fv) sets the fluid in the duct intomotion and moves the elastic membrane at the fenestra ovalis (fo). Theelastic perilymphatic duct approximates the extension of the sacculusthat contains the basilar papilla and transmits pressure changes to thisorgan. B. In the reptile inner ear the perilymphatic duct is more elongatedand intimately associated with the elongated lagena. Again, this is amechanism by which perilymphatic motion caused by ambient vibrationand/or pressure changes is transmitted to an organ specialized to sensethese stimuli. C. In birds, crocodylids, and mammals the perilymphaticduct becomes more elongated and, in mammals, differentiates intoscala vestibuli (sv) and scala tympani (st). These envelop the cochlearduct, containing the auditory sensory organ of Corti.

486 CAREY AND AMIN

However, it is interesting to note that many of theosteichthyes do not have such an arrangement of com-plete immersion of the membranous labyrinth; rather,there is soft tissue surrounding the pars superior struc-tures in many cases. (Baird, 1974). Perhaps more im-portant than the complete immersion in perilymph hasbeen the development of a perilymphatic duct systemthat appropriately directs the endolymph flow resultingfrom ambient pressure changes.

Perilymphatic Duct System to Direct PressureGradients Originating at Fenestra Vestibuli

The fenestra vestibuli, or oval window, first appeared inthe chondrichthyes. In the amphibians, however, it be-comes associated with a perilymphatic ductal system(Harrison, 1902) (Fig. 3). Just deep to the oval window isa perilymphatic sac. This sac approximates the wall of thesaccule, where the membranous labyrinth wall thins outso that only two cell layers separate endolymph and per-ilymph. An extension of this perilymphatic duct runs upover the saccule and then down to a second mobile windowin the labyrinth, the foramen perilymphaticum, a mem-branous opening into the intracranial or meningeal space.This is the configuration in urodeles. In anurans, there isan additional fenestra at the foramen perilymphaticuminferius, which is an extracranial opening. This configu-ration is the first of the oval/round window type.

The perilymphatic duct directs pressure changes fromthe oval window to a second mobile window. In its pas-sage, the duct is associated with auditory sensory epithe-lia, such as the basilar papilla in urodeles. In reptiles, theperilymphatic duct system runs between the stapes in theoval window to the round window. The perilymphatic ductloops around another elongating duct protruding from thesaccule, the lagena. The auditory epithelium associatedwith this configuration is again the basilar papilla. Theelongation of the duct systems continues in birds andcrocodylids so that the loops of the perilymphatic ductbecome scala vestibuli and scala tympani, and the elon-gating lagena becomes the cochlear duct. The basilar pa-pilla becomes an expanded tonotopic membrane. Finally,in mammals, the ducts coil to form the cochlea, and thebasilar papilla is specialized to form the organ of Corti(Lewis et al., 1985a).

In the human cochlea, the oval and round windows areseparated by the rigid osseous spiral lamina and the basi-lar membrane. The remainder of the labyrinth is normallycompletely encased in bone, except for the medial open-ings for the neurovascular bundles. Sound pressure ap-plied to the oval window by stapes motion is thus shuntedthrough scala vestibuli, around the helicotrema, and backalong the scala tympani to the elastic mobile membrane atthe round window. Shearing of the stereocilia of cochlearinner hair cells is caused by the interaction of the flexiblebasilar membrane and the overlying tectorial membranealong this route. In contrast, vestibular hair cells areisolated from this sound-induced movement of perilymphby virtue of their location in the closed bony labyrinth.

Accessory Structures to Tune End OrganResponses to Particular Frequency Ranges

The end organs in which hair cells are embedded serveto convert endolymph motion into shearing of the stereo-cilia relative to the cuticular plates of the hair cells. Thedifferent properties of the various accessory structures

associated with hair cells confer on the vestibular andauditory end organs ideal frequency ranges for operation.The tonotopic organization of the cochlea is a clear exam-ple of this for auditory hair cells. O’Leary et al. (1974)demonstrated that the crista of the semicircular canal alsohas regions of frequency specificity. The hydrodynamics ofthe semicircular canal are such that it functions best inthe frequency range of 0.012 to 27 Hz (Rabbitt et al.,2004). This ideal operating range for the semicircular isclearly much lower than the range of sound frequenciesthat best stimulate the human cochlea. Thus, anotherreason sound entering the inner ear may have little effecton vestibular sensation is that it is not in the effectivefrequency range for the vestibular end organs. However,infrasound of sufficient intensity can cause both auditorypercepts and disorientation (von Gierke and Parker,1974). It is important to recognize that even when thestimulus is not in the ideal operating range for a givenhair cell system, it does not mean that the system will notrespond to sufficiently intense stimuli.

Pathological Conditions in Which PressureAffects Human Balance Organs

When the bony capsule overlying a semicircular canal isdisrupted, one of the major protective mechanisms thatprevent sound energy from stimulating the canal is lost.Tullio (1929) demonstrated this in his seminal experi-ments with sound as a stimulus for the labyrinth in pi-geons after he fenestrated their semicircular canals. Heobserved that this caused eye and head nystagmus in theplane of the fenestrated semicircular canal. Huizinga(1935) proposed that the fenestra created a third “mobilewindow” in the labyrinth, in addition to the oval andround windows. This window opens another route forsound pressure dissipation in the labyrinth. This newroute is along the affected canal, so endolymph movesthrough the semicircular canal under the influence ofsound or other pressure changes applied to the oval orround windows. The fenestrated superior canal exposed toloud sound encodes the resulting endolymph flow as itwould a head rotation in the plane of the affected canaland toward the affected side. The vestibular system gen-erates a compensatory eye movement that is just theopposite of this perceived head movement.

Superior Canal Dehiscence SyndromeMinor et al. (1998) first described a syndrome of vertigo

and nystagmus caused by loud sounds or changes in in-tracranial or middle ear pressure. Observation of the nys-tagmus induced by these loud sounds applied to the af-fected ear demonstrated brief vertical-torsional eyemovements with slow phases moving upward and rollingcontralateral to the stimulated ear. This is the nystagmusexpected for excitation of the ipsilateral superior semicir-cular canal. High-resolution CT scans of the temporalbones in these individuals showed that they had sponta-neous dehiscences of one or both superior semicircularcanals (Fig. 4).

Evidence to date suggests that superior canal dehis-cence (SCD) is a developmental anomaly: it frequentlyoccurs bilaterally in the absence of any erosive lesions orintracranial hypertension, and the location of dehiscenceat the floor of the middle cranial fossa is the last portion ofthe temporal bone to ossify during development (Carey etal., 2000). It does not, however, appear to be due simply to

487SOUND TRANSMISSION TO BALANCE ORGANS

the development of the labyrinth in an abnormally highposition in the skull or to an abnormally rotated orienta-tion of the labyrinth (Potyagaylo et al., 2005). Instead, thedefect may be in the process of ossification at the locationabove the superior canal.

Patients may have auditory symptoms as well as ves-tibular symptoms; in some cases, they have only auditorysymptoms (Minor, 2005). These include the sensations ofhearing one’s voice, pulse, or even eye movements unusu-ally well in the affected ear. These symptoms arise froman unusual condition of conductive hypercusis: bone-con-ducted sound energy is admitted to the ear better thanair-conducted sound energy (Minor et al., 2003). This ef-fect has been successfully modeled by the introduction of athird mobile window in the labyrinth (Rosowski et al.,2004).

The symptoms in some patients with superior canaldehiscence can be controlled by avoiding loud sounds andpressure changes in the affected ear. For those patientswho are debilitated by their symptoms, plugging of thesuperior canal performed through a middle cranial fossaapproach has been effective in achieving resolution oftheir vestibular symptoms and signs (Minor, 2005).

Enlarged Vestibular Aqueduct SyndromeAnother inner ear malformation that creates an abnor-

mal opening in the bony labyrinth in humans is the en-larged vestibular aqueduct. Valvassori and Clemis (1978)identified enlargement (� 1.5 mm diameter) of the vestib-ular aqueduct on computerized tomography scans in 50patients. There were other associated ear abnormalities in60%, including enlarged vestibule (14), enlarged vestibuleand lateral semicircular canal (7), enlarged vestibule and

hypoplastic cochlea (4), and hypoplastic cochlea (4). Theauthors concluded that the large aqueduct presumablyrepresents an arrested phase of inner ear development.They found that bilateral involvement is twice as commonas unilateral with a female-to-male predominance of 3:2.

Most cases of enlarged vestibular aqueduct syndromeare associated with congenital sensorineural hearing loss.However, some patients have a conductive component totheir hearing loss as well (Arjmand and Webber, 2004;Sheykholeslami et al., 2004). This conductive componentlikely represents a similar third mobile window effect seenin superior canal dehiscence.

Vestibular-Evoked Myogenic Potentials:Manifestation of Pressure Effects on NormalHuman Balance Organs

As noted previously, sufficiently loud sound can stimu-late the vestibular end organs in the inner ear. In the caseof the human ear, the saccule seems to be the most sen-sitive organ to sound stimulation. This may represent aninherent sound sensitivity retained by the saccule in ver-tebrate evolution. However, the primate saccule has beenfound to be no more sound-sensitive than the other ves-tibular end organs (Young et al., 1977). In a guinea pigmodel, sufficiently loud sounds applied to the ear excitevestibular nerve units with other responses typical of sac-cular afferents (Murofushi and Curthoys, 1997).

A useful clinical test has developed from these findings.Vestibular-evoked myogenic potentials (VEMPs) are tran-sient decreases in flexor muscle electromyographic (EMG)activity that can be evoked by loud acoustic clicks or tonesapplied to the ear. Activation of multiple saccular unitssimultaneously is probably interpreted by the central ves-tibular system as a sudden loss of postural tone. As aresult, the system responds with an increase in extensormuscle activity and a decrease in flexor muscle activity.

The EMG activity averaged over multiple acoustic stim-uli from a tonically contracting flexor muscle will demon-strate a biphasic short-latency relaxation potential. TheEMG activity can be recorded in many different flexormuscles, but the sternocleidomastoid responses have beenbest described (Colebatch and Halmagyi, 1992). Becausethe saccule is the only end organ that mediates VEMPresponses, absence of VEMP responses may indicate sac-cular dysfunction. However, transmission of the VEMPacoustic stimulus is very sensitive to any cause of conduc-tive hearing loss in the middle ear, and VEMPs are usu-ally absent in the presence of conductive hearing loss.

In fact, the preservation of VEMP responses in the faceof conductive hearing loss implies an abnormally lowacoustic impedance of the labyrinth, such as occurs insuperior canal dehiscence syndrome (Minor et al., 2003) orwith enlarged vestibular aqueduct syndrome (Nakashimaet al., 2000).

SUMMARYEvolution of the vertebrate labyrinth has kept the ves-

tibular and auditory sensory organs within the same cap-sule. As a result, sound energy can, under certain circum-stances, affect the vestibular system. Several labyrinthinespecializations work to minimize the effects of sound onthe vestibular end organs. These specializations include arelatively inelastic membranous labyrinth wall, a variablyrigid otic capsule, immersion of the membranous laby-rinth in perilymph, a perilymphatic duct to channel pres-

Fig. 4. High-resolution CT scan images from a patient with superiorsemicircular canal dehiscence. The inset image shows the affectedsuperior canal along the long axis. The main image is a CT reconstruc-tion perpendicular to the plane of the canal. In these views, loss of boneover the superior semicircular canal is apparent. This dehiscence leadsto abnormal vestibular sensitivity to acoustic stimuli.

488 CAREY AND AMIN

sure changes from the oval window to the round window,and different operating frequencies for vestibular versusauditory epithelia. Despite these specializations, suffi-ciently loud sound can stimulate the normal human sac-culus. This gives rise to the useful test of vestibular-evoked myogenic potientials. When there is an abnormalopening in the labyrinth, e.g., at the superior canal orvestibular aqueduct, vertigo can result from loud soundsor pressure changes. In addition, conductive hypercusis orconductive hearing loss can result.

LITERATURE CITEDArjmand EM, Webber A. 2004. Audiometric findings in children with

a large vestibular aqueduct. Arch Otolaryngol Head Neck Surg130:1169–1174.

Baird IL. 1974. Anatomical features of the inner ear in submamma-lian vertebrates. In: Keidel WD, Neff WD, editors. Handbook ofsensory physiology: auditory system. New York: Springer-Verlag. p159–212.

Bone Q, Ryan KP. 1978. Cupular sense-organs in ciona (Tunicata-Ascidiacea). J Zool 186:417–429.

Carey JP, Minor LB, Nager GT. 2000. Dehiscence or thinning of boneoverlying the superior semicircular canal in a temporal bone sur-vey. Arch Otolaryngol Head Neck Surg 126:137–147.

Colebatch JG, Halmagyi GM. 1992. Vestibular evoked potentials inhuman neck muscles before and after unilateral vestibular deaffer-entation. Neurology 42:1635–1636.

Corey DP, Hudspeth AJ. 1979. Response latency of vertebrate haircells. Biophys J 26:499–506.

de Burlet HM. 1934. Zur vergleichenden Anatomie und Physiologiedes perilymphatischen Raumes. Acta Otolaryngol (Stockh) 13:153–187.

Denk W, Webb WW. 1992. Forward and reverse transduction at thelimit of sensitivity studied by correlating electrical and mechanicalfluctuations in frog saccular hair cells. Hear Res 60:89–102.

Dickman JD, Reder PA, Correia MJ. 1988. A method for controlledmechanical stimulation of single semicircular canals. J NeurosciMethods 25:111–119.

Fraser PJ, Shelmerdine RL. 2002. Dogfish hair cells sense hydrostaticpressure. Nature 415:495–496.

Fritzsch B, Beisel KW. 2001. Evolution and development of the ver-tebrate ear. Brain Res Bull 55:711–721.

Fuchs PA, Evans MG. 1990. Potassium currents in hair cells isolatedfrom the cochlea of the chick. J Physiol 429:529–551.

Glowatzki E, Fuchs PA. 2002. Transmitter release at the hair cellribbon synapse. Nat Neurosci 5:147–154.

Grossman GE, Leigh RJ, Abel LA, Lanska DJ, Thurston SE. 1988.Frequency and velocity of rotational head perturbations duringlocomotion. Exp Brain Res 70:470–476.

Harrison MS. 1902. On the perilymphatic spaces of the amphibianear. Int Monatschrift Anat Physiol 19:221–261.

Holt JR, Eatock RA. 1995. Inwardly rectifying currents of saccularhair cells from the leopard frog. J Neurophysiol 73:1484–1502.

Hudspeth AJ. 1985. The cellular basis of hearing: the biophysics ofhair cells. Science 230:745–752.

Huizinga E. 1935. On the sound reactions of Tullio. Acta Otolaryngol(Stockh) 22:359–370.

Kernan M, Zuker C. 1995. Genetic approaches to mechanosensorytransduction. Curr Opin Neurobiol 5:443–448.

Lewis ER, Leverenz EL, Bialek WS. 1985a. Comparative inner earanatomy. In: The vertebrate inner ear. Boca Raton, FL: CRC Press.p 13–94.

Lewis ER, Leverenz EL, Bialek WS. 1985b. The general componentsand roles of inner ears. In: The vertebrate inner ear. Boca Raton,FL: CRC Press. p 3–11.

Lowenstein O, Osborne MP, Wersall J. 1964. Structure and innerva-tion of the sensory epithelia of the labyrinth in the thornback ray(Raja clavata). Proc R Soc Lond B Sci 160:1–12.

Lowenstein O, Osborne MP, Thornhill RA. 1968. The anatomy andultrastructure of the labyrinth of the lamprey (Lampetra fluviatilisL.). Proc R Soc Lond B Biol Sci 170:113–134.

Lowenstein O. 1970. The electrophysiological study of the responses ofthe isolated labyrinth of the lamprey (Lampetra fluviatilis) to an-gular acceleration, tilting and mechanical vibration. Proc R SocLond B Sci 174:419–434.

Lowenstein O, Thornill RA. 1970. The labyrinth of Myxine: anatomy,ultrastructure and electrophysiology. Proc R Soc London B Biol Sci176:21–42.

Maisey JG. 2001. Remarks on the inner ear of elasmobranchs and itsinterpretation from skeletal labyrinth morphology. J Morphol 250:236–264.

Minor LB, Solomon D, Zinreich JS, Zee DS. 1998. Sound- and/orpressure-induced vertigo due to bone dehiscence of the superiorsemicircular canal. Arch Otolaryngol Head Neck Surg 124:249–258.

Minor LB, Carey JP, Cremer PD, Lustig LR, Streubel SO, Rucken-stein MJ. 2003. Dehiscence of bone overlying the superior canal asa cause of apparent conductive hearing loss. Otol Neurotol 24:270–278.

Minor LB. 2005. Clinical manifestations of superior semicircular ca-nal dehiscence. Laryngoscope 115:1717–1727.

Murofushi T, Curthoys IS. 1997. Physiological and anatomical studyof click-sensitive primary vestibular afferents in the guinea pig.Acta Otolaryngol 117:66–72.

Myers P. 2001. Vertebrata. Animal diversity Web site (http://an-imaldiversity.ummz.umich.edu/site/accounts/information/Verte-brata.html). Accessed 1 December 2005.

Nakashima T, Ueda H, Furuhashi A, Sato E, Asahi K, Naganawa S,Beppu R. 2000. Air-bone gap and resonant frequency in large ves-tibular aqueduct syndrome. Am J Otol 21:671–674.

O’Leary DP, Dunn RF, Honrubia V. 1974. Functional and anatomicalcorrelation of afferent responses from the isolated semicircularcanal. Nature 251:225–227.

Pickles JO, Corey DP. 1992. Mechanoelectrical transduction by haircells. Trends Neurosci 15:254–259.

Potyagaylo VL, Della Santina CC, Minor LB, Carey JP. 2005. Supe-rior canal dehiscence is not due to cephalic displacement of thelabyrinth. Ann NY Acad Sci 1039:498–502.

Rabbitt RD, Boyle R, Highstein SM. 1995. Mechanical indentation ofthe vestibular labyrinth and its relationship to head rotation in thetoadfish, Opsanus tau. J Neurophysiol 73:2237–2260.

Rabbitt RD, Damiano ER, Grant JW. 2004. Biomechanics of the semi-circular canals and otolith organs. In: Highstein SM, Fay RR, PopperAN, editors. The vestibular system. New York: Springer. p 153–201.

Rosowski JJ, Songer JE, Nakajima HH, Brinsko KM, Merchant SN.2004. Clinical, experimental, and theoretical investigations of theeffect of superior semicircular canal dehiscence on hearing mecha-nisms. Otol Neurotol 25:323–332.

Rusch A, Eatock RA. 1996. A delayed rectifier conductance in type Ihair cells of the mouse utricle. J Neurophysiol 76:995–1004.

Sheykholeslami K, Schmerber S, Habiby KM, Kaga K. 2004. Vestib-ular-evoked myogenic potentials in three patients with large ves-tibular aqueduct. Hear Res 190:161–168.

Shimeld SM, Holland PW. 2000. Vertebrate innovations. Proc NatlAcad Sci USA 97:4449–4452.

Tullio P. 1929. Das Ohr und die Entstehung der Sprache und Schrift.Berlin: Urban und Schwarzenberg.

Valvassori GE, Clemis JD. 1978. The large vestibular aqueduct syn-drome. Laryngoscope 88:723–728.

von Gierke HE, Parker DE. 1974. Infrasound. In: Keidel, Neff WD,editors. Handbook of sensory physiology. Berlin: Springer-Verlag. p585–624.

Wilson HV, Mattocks JE. 1897. The lateral sensory anlage in thesalmon. Anatomischer Anzeiger 13:658–660.

Yamauchi A, Rabbitt RD, Boyle R, Highstein SM. 2002. Relationshipbetween inner-ear fluid pressure and semicircular canal afferentnerve discharge. J Assoc Res Otolaryngol 3:26–44.

Young ED, Fernandez C, Goldberg JM. 1977. Responses of squirrelmonkey vestibular neurons to audio-frequency sound and headvibration. Acta Otolaryngol (Stockh) 84:352–360.

489SOUND TRANSMISSION TO BALANCE ORGANS