Tell Me Your Vestibular Deficit, and I’ll Tell You How You’ll Compensate

BASIC AND CLINICAL ASPECTS OF VERTIGO AND DIZZINESS

Tell Me Your Vestibular Deficit, and I’ll TellYou How You’ll Compensate

Michel Lacour, Sophie Dutheil, Brahim Tighilet,Christophe Lopez, and Liliane Borel

Aix-Marseille Universite, UMR 6149 Universite de Provence/CNRS,Marseille Cedex 03, France

Most patients with unilateral vestibular loss exhibit a similar static and dynamicvestibular syndrome consisting of vestibulo-ocular, posturolocomotor, and perceptivedeficits. This vestibular syndrome recovers more or less completely and more or lessrapidly over time. One open question is whether recovery mechanisms differ accordingto vestibular pathology and/or patients. It is reported here (1) data from three differentcat models of unilateral vestibular loss reproducing vestibular pathology with sudden(unilateral vestibular neurectomy [UVN] model), gradual (unilateral labyrinthectomy[UL] model), or reversible (tetrodotoxine [TTX]) model) loss of vestibular function, and(2) clinical observations in a population of unilateral vestibular loss patients sufferingthe same pathology (Meniere’s disease). Animal models show that time courses andmechanisms of recovery depend on the type of vestibular deafferentation, and clinicalfindings show that Meniere’s patients compensate their postural and perceptive deficitsusing different vicarious processes. Taken together, results point to a more complex pic-ture of compensation after unilateral vestibular loss, which cannot be reduced either to acommon recovery mechanism or to a single process identical for all individuals. Thesefindings should guide physiotherapists in treatment and rehabilitation for vestibulardeficits.

Key words: vestibular pathology; unilateral vestibular neurectomy; unilaterallabyrinthectomy; tetrodotoxine; posturo-locomotor function recovery; neurogenesis;idiosyncratic vicarious processes; cat; Meniere’s disease

Introduction

The anatomical and functional organizationof the vestibular system is now well under-stood,1,47 and its contribution to a wide rangeof functions, from postural and oculomotor re-flexes to spatial representation and cognition,has recently been underscored.9

Afferent fibers contacting sensory type I andtype II hair cells have their cell bodies locatedin Scarpa’s ganglion. Axons from these bipo-lar cells group into different branches of thevestibular nerve and project onto four main

Address for correspondence: Michel Lacour. Aix-Marseille Universite,UMR 6149 Universite de Provence/CNRS, Pole 3C, Case B, Centre deSt. Charles, 3 Place Victor Hugo – 13331 Marseille Cedex 03, [email protected]

vestibular nuclei (VN) in the brain stem. Threetypes of Scarpa’s ganglion neurons can bedistinguished on the basis of their anatomi-cal and functional characteristics: neurons withchalice-like dendritic terminals innervating thecenter of the sensory epithelium (type I sen-sory cells), with an irregular resting dischargeand a fast adaptation to prolonged stimulation(phasic units); neurons with “en bouton” den-dritic terminals innervating the periphery of thesensory epithelium (type II sensory cells), witha regular resting discharge and no adaptationto sustained stimuli (static units); and dimor-phic neurons contacting both type I and typeII cells in the sensory epithelium and exhibitingmixed response properties (phasicotonic units).Second-order vestibular neurons located in theVN complex show resting discharge averaging

Basic and Clinical Aspects of Vertigo and Dizziness: Ann. N.Y. Acad. Sci. 1164: 268–278 (2009).doi: 10.1111/j.1749-6632.2008.03731.x C© 2009 New York Academy of Sciences.

268

Lacour et al.: Vestibular Deficit 269

40 impulses/s in many species. These vestibu-lar neurons receive convergent multimodal sen-sory inputs from the visual and somatosensorysystems and, additionally, from the contralat-eral VN (via the commissural pathways), thecerebellum and the cerebral cortex. The VNcomplex projects to ocular and spinal motoneu-rons and influences the stabilization and orien-tation of gaze, of the head and of the body inspace. The VN also projects to various corticalareas implicated in self- as well as extrapersonalperception and, consequently, vestibular inputscontribute to high-level functions like internalspatial representation.

The vestibular syndrome observed after le-sion in the periphery involves successive lev-els of damage, from basic reflex problems tohigher-order deficits.12,13,23 Patients with uni-lateral vestibular loss suffer primarily frombalance and gait problems related to an im-paired vestibulospinal reflex, oscillopsia corre-sponding to impairment of the vestibulo-ocularreflex, and perceptive deficits, including ver-tigo and spatial disorientation. This syndromecompletely ameliorates over time, in a pro-cess described in the literature as “vestibularcompensation.”12,13,23,45

While neural plasticity mechanisms andtheir relation to functional recovery have beenthoroughly investigated in many species,14

many open questions remain. Why, for ex-ample, can the time course of recovery beso different between patients? Are distinctrecovery mechanisms involved in differentvestibular pathologies? Why do some uni-lateral vestibular-loss patients remain poorlycompensated while others do not? Are theredifferent ways to compensate a given pathol-ogy, with some being more efficient thanothers?

The first objective of this chapter is to as-sess whether recovery-time courses and mech-anisms are strongly dependent on vestibularpathology. This point is illustrated by investiga-tions in three experimental models of unilateralvestibular loss, corresponding to pathologieswith sudden, gradual, and reversible vestibu-

lar deficits. Clinical observations in Meniere’spatients are reported in a second part, whereinit is shown that although patients suffer fromthe same disease and are treated surgically inthe same way, they recover by using differentstrategies or reference frames.

Vestibular Pathologies andCorresponding Animal Models

Unilateral vestibular-defective patients com-plain of similar symptoms (vertigo, oscillop-sia, dizziness, and instability) despite vestibu-lar pathologies of different origins. Indeed,many pathology conditions can affect periph-eral vestibular function,10 and clinical classi-fication based on degenerative disorders, ac-quired disorders of known cause, and acquireddisorders of unknown (or multiple) causes wasproposed by Schuknecht.39 The basic patho-physiologic mechanisms of vestibular deficitsinclude the loss of sensory hair cells or neuronalcells, alteration of the hydrodynamics and mo-tion mechanics of the inner ear, and paralysis ofsensorineural function. The severity of the ini-tial deficits as well as the recovery-time course ofvestibular compensation are believed to differaccording to the various underlying patholo-gies. Furthermore, since mechanisms of plas-ticity after brain injury are found to be stronglydependent on the nature of the lesion (totalversus partial, acute versus chronic, fast versusslow starting), one can assume that recoverymechanisms should likewise differ.

Gradual and Partial Loss of UnilateralVestibular Function

With advancing age come impairments ofmost sensorimotor and cognitive functions,and these include degenerative changes in thevestibular system. In fact, age-related degener-ative changes are the most common cause ofgradual and partial loss of vestibular function.Reductions in the sensory hair-cell populationsaveraging 20% for the maculae and 40% for

270 Annals of the New York Academy of Sciences

the crista ampullaris were reported in individ-uals over 70 years, albeit with significant in-terindividual variations.34,35 This finding wasconfirmed by Merchant and colleagues30 intheir normative data on sensory hair-cell countin the vestibular organs of individuals frombirth to 100 years of age. The authors founda similar significant loss of type II cells with in-creasing age in the maculae and the crista ofall three canals, and a significant loss of typeI cells at a rate greater in the cristae than inthe maculae. A parallel reduction in the num-ber of vestibular nerve fibers, averaging 37%,was reported in individuals between 75 and85 years of age compared with individuals be-low 35 years of age, again with pronouncedinterindividual variations.3 Furthermore, andcorresponding to changes in their vestibular af-ferent pattern, the vestibular nuclei showed aneuronal loss averaging 3–5% per decade be-tween the ages of 40 and 93 years, as well as adecreased volume of the VN complex.28,41

Acoustic neuroma may affect vestibularnerve fibers and afferent vestibular inputs tothe VN complex in a fashion similar to age-related changes. A tumor’s size and its growthspeed are two factors contributing to the alteredtransmission of vestibular input to the VN. Pro-gressive and partial loss of vestibular functionis found in the case of small-size tumors oreven larger vestibular schwannomas of the cere-bellopontine angle that fill the internal audi-tory canal. In this latter case, a slow-growingschwannoma induces progressive pressure atro-phy of the cochlear and vestibular nerve trunks.In contrast, a faster and more complete lossof vestibular function is expected with largeracoustic neuromas exhibiting fast growth, or asa consequence of removing the tumor duringsurgery.

In spite of age-related degenerative changesthat start as early as 40 or 50 years ofage,28 the critical impact of these changeson balance function is seen about 30 yearslater.2,46 This strongly suggests that reductionin the vestibular input may be compensatedfor by other sensory cues, including visual

and proprioceptive inputs.10,29 Sensory substi-tution processes have been reported in bothvestibular-defective patients and animal mod-els of vestibular deafferentation.23 Consistentwith this phenomenon are reports of individu-als free of signs of vestibular deficit, or show-ing few symptoms for a long time, in spite ofvestibular schwannoma. Because the centralnervous system (CNS) is able to gradually ad-just, slow tumor-induced degenerative changesin these patients can be compensated for on-line. Reweighting of remaining sensory cues isvery likely involved, and this does not excludelocal structural reorganizations within the VN.

Ototoxic drugs (aminoglycosides like gen-tamicin or streptomycin) and trauma of thehead (fracture of the temporal bone, otiticbarotrauma, whiplash-associated disorders) areother known causes of peripheral vestibular dis-orders inducing gradual loss of sensory cells andprimary vestibular neurons. Progressive loss ofsensory hair cells, for instance, is seen withstreptomycin, and the amount of sensory cellloss depends on drug dosage and the durationof treatment, with a higher sensitivity for thehair cells of the semicircular canals comparedto those in the utricule and saccule.

Animal Model

The animal model corresponding to grad-ual and partial loss of unilateral vestibularfunction employs a unilateral labyrinthectomy(UL) to achieve deafferentation. This lesion ofthe peripheral vestibular system, which sparesScarpa’s ganglion, causes progressive degen-erative changes characterized—in the cat—byslow transganglionic degeneration of vestibularnerve fibers. This process continues for severalyears after UL.37 Additionally, Scarpa’s neu-rons continue to feed VN neurons with affer-ent tonic input related to resting discharge oftheir axons.19 The UL model, therefore, repro-duces a partial and gradual deafferentation ofthe VN cells. Deafferentation of the VN com-plex is more functional than structural withinthis model (Fig. 1).


Figure 1. Animal models of vestibular patholo-gies. Three animal models of unilateral vestibular lossare illustrated: unilateral labyrinthectomy (UL: up-per panel), unilateral vestibular neurectomy (UVN:middle panel), and transtympanic injection oftetrodotoxin (TTX: lower panel). The UL, UVN, andTTX models reproduce gradual, sudden, and transientunilateral loss of vestibular function, respectively, andare supposed to mimic more or less accurately dif-ferent vestibular pathologies or diseases (see text).(Abbreviation: SG, Scarpa’s ganglion.)

Sudden and Total Loss of UnilateralVestibular Function

The best way to cause a sudden and com-plete unilateral loss of vestibular function is tosurgically cut the vestibular nerves on one side.This procedure induces a total structural (lossof anatomic connections) and functional (lossof all vestibular information) deafferentation ofthe ipsilateral VN complex. One example ofthis type of deafferentation is Meniere’s patientswho undergo a curative vestibular neurotomy,while a second one consists of those patientstreated surgically for acoustic neuromas andvestibular schwannomas, since the vestibularnerves are cut when the tumor is removed.

Among the most common causes of fast-onset total (or subtotal) destruction of periph-eral vestibular function are ischemic necrosis,head trauma, and bacterial and viral labyrinthi-tis. Another classic example of acute unilateral

loss of vestibular function is vestibular neuritis.Postmortem exams in patients with vestibularneuritis may reveal a complete atrophy of thevestibular nerve branches on the affected sidewith no accompanying morphological changesto the ipsilateral vestibular end organs, whereasboth vestibular nerve and sensory hair cells ap-pear normal on the unaffected side.

Animal Model

The cat model of unilateral vestibularneurectomy (UVN) reproduces very well atotal and sudden loss of vestibular function.Postganglion axotomy due to UVN associatesfunctional deafferentation of the VN complex(UL model) with structural (anatomic) deaf-ferentation, since all primary vestibular affer-ents degenerate (Fig. 1). Fast Wallerian degen-eration after UVN contrasts with the relativelyslow transganglionic degeneration of vestibularnerve fibers after UL.22 In this case, however,one important difference in between the animalmodel and the corresponding human vestibu-lar pathology is the absence of compensatoryprocesses prior to UVN in the cat.

Transient Impairment of UnilateralVestibular Function

Patients with Meniere’s disease show dra-matic vestibular symptoms during their crises,but often have few objective findings betweenattacks. A typical clinical picture involves re-curring episodes of severe vertigo with nauseaand vomiting, fluctuating hearing loss, and tin-nitus. It was hypothesized that overaccumula-tion of endolymph results in recurring parox-ysmal contamination of the perilymph withneurotoxic (K+) endolymph.38 Periodic en-dolymphatic membrane ruptures, with subse-quent transient potassium palsy of the vestibu-lar nerve fibers, were also postulated as thecausative factor for vertigo attacks and posturalinstability in Meniere’s disease.

Benign paroxysmal positional vertigo(BPPV), the most common cause of peripheralvertigo induced by vestibular end-organ


disorder, also can be seen as a reversible andpartial dysfunction of the peripheral vestibularsystem. The clinical picture of this disorderis violent vertigo and intense nystagmusassociated with head movement in the planeof the affected semicircular canal, or withchanges in head position. Patient history isusually quite typical, and a positive responseto the Hallpike maneuver test is virtuallydiagnostic. Cupulolithiasis has been proposedas a pathophysiological mechanism,36 butthere are now convincing arguments insteadfor canalolithiasis and a “heavy clot-induced”endolymph flow mechanism as the causativefactor. Lying on the affected side, for instance,moves free-floating endolymph particlesand produces a particle-induced endolymphcurrent that displaces the cupula away fromthe utricle. The result is an increase in theresting discharge rate of first-order vestibularneurons in Scarpa’s ganglion, which createsthe static vestibular asymmetry in the restingactivity between the two VN complexes andresults in the BPPV symptoms of nystagmusand vertigo attack. Avoidance of specific ag-gravating head movements, however, preventsthis transient impairment and allows patientsto remain symptom free. BPPV may resolveitself spontaneously over weeks or months, andcan be treated effectively with physiotherapythat employs the liberatory maneuver40 ormodified otolith repositioning procedures.

Animal Model

One means of reproducing transient uni-lateral loss of vestibular function in an ani-mal model is to block primary vestibular af-ferents’ resting discharge with tetrodotoxine(TTX). Transtympanic TTX injection func-tionally deafferents the VN complex on theinjected side, and creates postural and ocu-lomotor syndromes similar to those observedafter UL.11 TTX vestibular syndrome is re-versed once the drug has cleared the systemseveral days after TTX injection, and the timecourse of this recovery is dose-dependent. Al-though the TTX model mimics transient uni-

lateral loss of vestibular function, its mechanismis not closely analogous to known vestibularpathologies.

Animal Models of VestibularPathologies Do Not Recover

Similarly

Here we report data collected from our threeexperimental models of unilateral vestibularloss. We investigated the time course of pos-ture recovery in TTX, UL, and UVN cats us-ing behavioral tests and immunohistochemicalmethods to measure reactive neurogenesis (cellproliferation and differentiation) as a possibleunderlying recovery mechanism within the VNusing. Readers may refer to our original paperson behavior15,42,43,48 and immunohistochem-istry15,44 for more details.

The Time Course of Recovery Dependson the Type of Vestibular Loss

After TTX injection, UL, or UVN, all ani-mals exhibited the typical vestibular syndromeconsisting of ocular nystagmus, postural asym-metry, and falls toward the lesioned side. Pos-tural syndrome was qualitatively similar in allthree cat models of unilateral vestibular loss.Initial postural signs comprised head tilt andfalls toward the side of the lesion, and postu-ral asymmetry with an increase in support sur-face. The recovery time course from the initialdeficits were, however, quantitatively differentbetween cat models.

Figure 2 illustrates the mean time coursesof posture recovery recorded in the TTX, UL,and UVN cats. TTX-treated cats showed lowerinitial deficits and faster recovery compared tothe others. On average, the postinjection nor-malized support surface was 3.3 times largerin this group than control measurements takenbefore TTX treatment, and animals fully recov-ered within one week. The UL cats exhibitedsimilar initial postural deficits, but took longer(around two weeks) to fully recover. UVN cats


Figure 2. Posture recovery depends on the typeof vestibular loss. Time course of posture recovery inthe three experimental models of unilateral vestibularloss in cats: unilateral vestibular neurectomy (UVN:open triangles), unilateral labyrinthectomy (UL: opensquares), and unilateral transtympanic injection oftetrodotoxin (TTX: open circles). Support surface (onthe ordinates) was evaluated in cm2 and normalizedwith respect to the prelesion values referred to unityfor each cat as a function of the postdeafferentationtime (abscissae, in weeks). Significant differences(P < 0.001) were found in the recovery profiles be-tween UVN and UL animals, and between UL andTTX cats. (Modified from Dutheil et al.15)

exhibited the largest postural deficits of thethree groups in the days following vestibular le-sion (5 times the control value) and the longesttime period for recovery. On average, a returnto normal posture required six weeks or morein the UVN group.

Recovery Mechanisms Dependon the Type of Vestibular Loss

Neurogenerative capacity of the adult cen-tral nervous system may help promote endoge-nous repair of the adult CNS after stroke orinjury.20 At least part of the behavioral recov-ery process after vestibular lesion has been at-tributed to plastic events in the deafferentedVN.14,23 The question of whether cell prolif-eration and differentiation, a repair processinvolved in CNS structures after deafferenta-tion,21 was also present in our different modelsof vestibular deafferentation was, therefore, ofinterest.

We provided the first demonstration of reac-tive neurogenesis in the deafferented VN in our

UVN model. A high number of dividing cellswith typical 5-bromo-2′-deoxyuridine (BrdU) -immunoreactive nuclei were found exclusivelyin the deafferented VN, with a peak in cell pro-liferation at 3 days after UVN.44 Most of thenewly generated cells survived up to one monthafter UVN and gave rise to a variety of celltypes. Confocal analyses revealed three cell lin-eages: a low proportion of microglial cells (OX42/BrdU-immunoreactive cells), high propor-tions of astrocytes [glial fibrillary acidic pro-tein (GFAP) /BrdU-immunoreactive cells], andnewly generated neurons [neuronal nucleus(NeuN) /BrdU-immunoreactive cells]. All thenewly generated neurons consisted of small-and medium-sized GABAergic neurons. Thefunctional role of reactive neurogenesis wasfurther investigated by injecting the UVN catswith antimitotic drugs, which completely blockcell proliferation. Results showed antimitotic-induced dramatic delay of compensation forboth ocular motor and posture deficits,16 sug-gesting that the newly generated GABAergicneurons play an important role in the recoveryprocess.

In contrast, newly generated neurons are to-tally absent in the TTX and UL cats. Figure 3Aillustrates BrdU-immunostaining in the deaf-ferented medial VN of representative UVN,UL, and TTX cats examined 3 days aftervestibular deafferentation. Whereas intense cellproliferation peaked at 3 days in the UVN an-imals, no cell proliferation was observed in theUL and TTX cats (Fig. 3B). We can conclude,then, that reactive neurogenesis is a plasticitymechanism restricted to the UVN model.

Conclusions

Taken together, our results demonstrate thatboth recovery profile and underlying recov-ery mechanisms differ according to the typeof vestibular deafferentation. Transient (theTTX model) and gradual (the UL model)loss of vestibular function are compensatedfaster than acute and sudden vestibular loss(the UVN model), and do not involve reactive


Figure 3. (A, B) Recovery mechanisms dependon the type of vestibular loss. Cell proliferation inthe deafferented medial vestibular nuclei (MVN) inthe three animal models of unilateral vestibular lossin cats: unilateral vestibular neurectomy (UVN), uni-lateral labyrinthectomy (UL), and unilateral transtym-panic injection of tetrodotoxin (TTX). Cell prolifera-tion in response to vestibular deafferentation wasperformed by injection of 5-bromo-2′-deoxyuridine(BrdU), a marker of cell mitotic activity. (A) Photomi-crographs of BrdU immunostaining in the medial VNof UVN, UL, and TTX of cats observed 3 days aftervestibular deafferentation. Note the intense cell pro-liferation in the UVN cats (arrows) contrasting withthe total lack of newly generated cells in both theUL and TTX animals. (Scale bar: 50 μm. 4V: fourthventricle.) (B) Quantification of the number of BrdU-positive nuclei in the medial VN. Compared to controlcats (C), there is no change in the TTX and UL ani-mals, whatever the postdeafferentation time (in days:abscissae). In contrast, there is a peak of cell prolif-eration at 3 days (D + 3) after UVN (Modified fromDutheil et al.15)

neurogenesis as a probable key recovery mech-anism. In addition, microglial and astroglial re-actions are present after gradual and suddenloss, but not after transient vestibular loss. Itappears, therefore, that the repertoire of brainplasticity mechanisms implicated in the recov-

ery process after unilateral vestibular loss ex-hibits a graded response, depending on whethervestibular deafferentation is functional or struc-tural. Blockade of vestibular afferent activityby transtympanic tetrodotoxin injection is verylikely too short-acting to elicit microglial andastroglial reactions, as was already reported inthe rat UL model.11

We have confirmed in the UL cat model thesame central vestibular glial reaction found inUL rats,11 but the role of this reaction in recov-ery remains to be more precisely determined.This reaction most probably is due to centraltissue inflammation following the peripherallesion. Growth factors and cytokines releasedby activated glial cells may promote the sur-vival of deafferented vestibular neurons in theVN complex, as suggested in other nerve-injuryparadigms.32,33 Reactive GABAergic neuroge-nesis, observed only in the UVN model, re-mains an open question, but it is likely bene-ficial for behavioral recovery, since preventingcell proliferation and differentiation induces astrong delay in vestibular compensation.16,44

Interindividual Differences in theRecovery Process for a Given

Vestibular Pathology

Here we report observations from patientssuffering Meniere’s disease and having under-gone the same surgical treatment (unilateralvestibular neurotomy) to alleviate their vertigo.We have clearly demonstrated that Meniere’spatients use several different sensory strate-gies to compensate for their postural deficits,and that representation of their vertical orien-tation, which is necessary for orienting theirbody in space, is based on different referenceframes. Readers may refer to our original pa-pers6–9,23–27 for more details.

Sensory Strategies as IdiosyncraticVicarious Processes

Using static posturography for investigatingposture control in Meniere’s patients examined


in eyes open (EO) and eyes closed (EC) con-ditions one week after being submitted to acurative UVN, we showed a bimodal distribu-tion of individual values of the mean percent-age difference in sway area between conditions.The whole population was split into two signif-icantly different subgroups: one with patientsswaying less in the EO condition (54% of theMeniere’s population), and the other with pa-tients swaying less in the EC condition (46%of the Meniere’s population) This partition re-vealed two different types of unilateral vestibu-lar defective patients—visual (V) and nonvisual(NV)—distinguished by their employment ofdifferent sensory strategies for controlling theirposture.

These strategies underscore the conceptof the subjects’ behavioral repertoire, com-posed of vicarious processes, compensating thevestibular loss, with idiosyncratic selection oforientation reference frames depending on thesubjects’ internal spatial representation. Theidea of differential weighting of visual cues forstatic balance control, though, had been re-ported earlier in Meniere’s patients4,5 and pa-tients suffering from vestibular neuritis.18 Onthe basis of dynamic posturographic record-ings, Nashner and colleagues31 and Black andNashner5 suggested as well that vestibular pa-tients must select an alternative orientation ref-erence based either on vision or on a supportsurface, that is, proprioception. Moreover, per-ceptive changes reported among unilateralvestibular-loss patients in judgments of visualvertical pointed again to differential reweight-ing and selection of allocentric versus ego-centric reference frames (see the followingsection).

Vertical Perception as Selection ofDifferent Orientation Reference Frames

Visual-field dependency was assessed be-fore unilateral vestibular loss using the rod-and-frame test. Subjective visual vertical (SVV)judgment was tested with a frame tilted 20◦ to-ward the ipsilesional or the contralesional side.

Once again, patients were easily divided intotwo well-defined and significantly different sub-populations. Some patients exhibited a largeframe effect (field-dependent patients [FD]),while the remaining patients showed a weakerframe effect (field-independent patients [FI]).The group to which a patient belongs is basedon a differential weighting of visual cues in SVVperception, with some relying more on an allo-centric frame of reference (FD) and others onegocentric or gravitational frames of reference(FI).

After unilateral vestibular loss, subjective vi-sual vertical perception differed greatly be-tween FD patients and FI patients. For ipsile-sional tilt of the frame, FD patients had largedeviations of the SVV in the direction of theframe tilt. As a rule, FI patients had weakerSVV deviations. The greatest difference be-tween the two groups occurred for a frametilted toward the contralesional side. FD pa-tients tilted the perceived vertical in the di-rection of the frame tilt, while FI patients al-ways tilted the SVV toward the lesioned side.Such changes, which occurred maximally oneweek after vestibular loss, remained for up to3 months (Fig. 4). Patients reported that frametilt toward the treated side was more disturbing.Additionally, when patients were required to setthe subjective vertical in front of a large disk ro-tating in the frontal plane, SVV changes due tounilateral vestibular loss were similar to thosedescribed with the tilted frame, both for the FDand FI patients. We therefore hypothesize thatsimilar perceptual deficits can be observed dur-ing manipulation of static and dynamic visualreferences.

When vertical and horizontal referenceswere provided, SVV was improved in both sub-populations, suggesting that all patients rely onthe same allocentric strategy. They all rely onvisual references they know to be stable and in-dicative of verticality (wall, door, and windowsedges). This likely explains why the recoverytime course of postural deficits is shorter inlighted environments with vertical and hori-zontal visual references than in darkness.


Figure 4. Visual-field dependency after vestibu-lar loss. Comparison between subjective visual verti-cal perception with tilted frame (static visual condi-tion) and with optokinetic stimulation in the frontalplane (dynamic visual condition) for ipsilesional(filled histograms) and contralesional (open his-tograms), and for field-dependent (FD) and field-independent (FI) patients tested 3 months after uni-lateral vestibular loss. Note similar subjective visualvertical perception for both static and dynamic vi-sual conditions, but significant differences betweenFD and FI patients.

Conclusions

We have summarized our view of the postu-ral and perceptive syndromes underscoring theleading role of internal spatial representation.9

We suggest that vestibular-induced changes af-ter unilateral vestibular loss are based on a dy-namic mental representation of space elabo-rated on the basis of the available sensory cues.This mental representation is updated continu-ously by taking into account the environmentalcontext and postural constraints, as well as thesubject’s psychological status. That compensa-tion of similar vestibular deficits may show in-terindividual differences between patients suf-fering from the same pathology can be moreeasily understood in the light of this concep-tion. Brain selection of new reference framesamong the preexisting ones operates differentlyaccording to the subjects, illustrating the reper-toire of various fast-adaptive processes. Thison-line fast selection and reweighting of ref-erence frames is likely favored by the overlap of

the neural substrates for egocentric, allocentric,and geocentric reference frames demonstratedin neuroimaging investigations in healthysubjects.17,49

Concluding Remarks

The process of compensation following uni-lateral vestibular loss has generally been consid-ered as a monolithic concept referring to sim-ilar recovery mechanisms, albeit complex andplural ones, in all vestibular pathologies andfor all patients. We do not view the compensa-tion process in that light. Instead, we proposetwo distinct determinants of the mechanismsvestibular compensation.

First, we suggest that the type of vestibularlesion determines the nature of the induced re-covery mechanisms. The CNS response differs,depending on whether vestibular deafferenta-tion is sudden, gradual or transient, partial ortotal. Fast-adaptive processes (reweighting ofremaining sensory inputs, switching referenceframes, shifting sensorimotor strategies) arepredominantly involved after functional deaf-ferentation, and do not necessitate a deep reor-ganization of the neuronal networks, whereasmechanisms of neuroplasticity (neurogenesis,axonal sprouting) are required for compensat-ing structural deafferentation. Different “soft-like” or “hard-like” recovery processes withshort or long time constants, respectively, aretherefore differentially implicated, dependingupon the nature of the vestibular pathology inquestion.

Second, we propose that different fast-adaptive processes observed for a givenpathology depend on the subjects’ behav-ioral repertoire, internal spatial representationand perceptive style. Idiosyncratic selection ofreference frames and sensorimotor strategies isthe rule rather than the exception in vestibularcompensation. What has been seen until nowas random variability should instead be consid-ered as vicarious processes related to a subject’sown history.


Of course, our models for central compensa-tion have some limitations regarding vestibularpathology, since they cannot reproduce exactlywhat happens in a particular disease. However,we believe that rehabilitation of patients withvestibular deficits should be guided by thesefindings, and that programs should considerboth the nature of a patient’s disease and theparticular compensation strategies employedby individual patients.

Conflicts of Interest

The authors declare no conflicts of interest.

References

1. Angelaki, D.E. & K.E. Cullen. 2008. Vestibular sys-tem: the many facets of a multimodal sense. Annu.

Rev. Neurosci. 31: 125–150.2. Baloh, R.W., S. Corona, K.M. Jacobson, et al. 1998.

A prospective study of posturography in normal olderpeople. J. Am. Geriatr. Soc. 46: 438–443.

3. Bergstrom, B. 1973. Morphology of the vestibularnerve. II. The number of myelinated vestibular nervefibers in man at various ages. Acta Otolaryngol. (Stockh.)

76: 173–179.4. Black, F.O. 1982. Vestibular function assessment in

patients with Meniere’s disease: the vestibulo-spinalsystem. Laryngoscope 92: 1419–1436.

5. Black, F.O. & L.M. Nashner. 1984. Vestibulo-spinalcontrol differs in patients with reduced versus dis-torded vestibular function. Acta Otolaryngol. (Stockh.)

(Suppl. 406): 110–114.6. Borel, L., F. Harlay, J. Magnan & M. Lacour. 2001.

How changes in vestibular and visual referenceframes combine to modify body orientation in space.Neuroreport 12: 3137–3141.

7. Borel, L., F. Harlay, J. Magnan, et al. 2002. Deficitsand recovery of head and trunk orientation and sta-bilization after unilateral vestibular loss. Brain 125:880–894.

8. Borel, L., F. Harlay, C. Lopez, et al. 2004. Walkingperformance of vestibular-defective patients beforeand after unilateral vestibular neurotomy. Behav. Brain

Res. 150: 191–200.9. Borel, L., C. Lopez, P. Peruch & M. Lacour. 2008.

Vestibular syndrome: a change in internal spatial rep-resentation. Clin. Neurophysiol. In press.

10. Brandt, T. 1999. Vertigo. Its Multisensory Syndromes, 2nded. Springer. London.

11. Campos-Torres, A., M. Touret, P.P. Vidal, et al. 2005.The differential response of astrocytes within thevestibular and cochlear nuclei following unilaterallabyrinthectomy or vestibular afferent activity block-ade by transtympanic tetrodotoxin injection in therat. Neuroscience 130: 853–865.

12. Curthoys, I.S. 2000. Vestibular compensation andsubstitution. Curr. Opin. Neurol. 13: 27–30.

13. Curthoys, I.S. & G.M. Halmagyi. 1995. Vestibularcompensation: a review of the oculomotor, neural,and clinical consequences of unilateral vestibular loss.J. Vestib. Res. 5: 67–107.

14. Dieringer, N. 1995. Vestibular compensation: neuralplasticity and its relations to functional recovery af-ter labyrinthine lesion in frogs and other vertebrates.Prog. Neurobiol. 46: 97–129.

15. Dutheil, S., M. Lacour & B. Tighilet. 2008. Differentadult cat models of unilateral vestibular loss inducedifferential astroglial and neuronal responses withinthe vestibular nuclei and different behavioural recov-ery time-courses. Submitted.

16. Dutheil, S., J.M. Brezun, J. Leonard, et al. 2008.Infusion of cytosine-b-D-arabinofluranoside (Ara-C)blocks vestibular lesion-induced cell proliferation inthe vestibular nuclei and delays behavioural recoveryin the adult cat. Submitted.

17. Galati, G., G. Committeri, J.N. Sanes & L. Pizza-miglio. 2001. Spatial coding of visual and somaticsensory information in body-centered coordinates.Eur. J. Neurosci. 14: 736–746.

18. Gagey, P.M. & M. Toupet. 1991. Orthostatic posturalcontrol in vestibular neuritis: a stabilometric analysis.Ann. Oto-Rhino-Laryngol. 100: 971–975.

19. Jensen, D.W. 1983. Survival of function in the deaf-ferented vestibular nerve. Brain Res. 8: 376–378.

20. Kempermann, G. 2006. Stem Cells and Neuronal Devel-

opment in the Adult Brain. Oxford University Press. NewYork.

21. Kokaia, Z. & O. Lindvall. 2003. Neurogenesis afterischaemic brain insults. Curr. Opin. Neurobiol. 13: 127–132.

22. Kunkel, A.W. & N. Dieringer. 1994. Morphologicaland electrophysiological consequences of unilateralpre-versus post-ganglionic vestibular lesions in thefrog. J. Comp. Physiol. 174: 621–632.

23. Lacour, M. 2006. Restoration of vestibular function:basic aspects and practical advances for rehabilita-tion. Curr. Med. Res. Opin. 22: 1651–1659.

24. Lacour, M., J. Barthelemy, L. Borel, et al. 1997. Sen-sory strategies in human postural control before andafter unilateral vestibular neurotomy. Exp. Brain Res.

115: 300–310.25. Lopez, C., M. Lacour, J. Magnan & L. Borel. 2006.

Visual field dependence-independence before and af-ter unilateral vestibular loss. Neuroreport 17: 797–803.


26. Lopez, C., M. Lacour, A.E. Ahmadi, et al. 2007.Changes of visual vertical perception: a long-termsign of unilateral and bilateral vestibular loss. Neu-

ropsychologia 45: 2025–2037.27. Lopez, C., M. Lacour, J. Leonard, et al. 2008. How

body position changes visual vertical perception afterunilateral vestibular loss. Neuropsychologia 46: 2435–2440.

28. Lopez, I., V. Honrubia & R.W. Baloh. 1997. Agingand the human vestibular nucleus. J. Vesibt. Res. 7:77–85.

29. Manchester, D., M. Woollacott, N. Zederbauer-Hylton & O. Marin. 1989. Visual, vestibular andsomatosensory contributions to balance control inthe older adult. J. Gerontol. 44: M118–M127.

30. Merchant, S.N., L. Velasquez-Villasenor, K. Tsuji,et al. 2000. Temporal bone studies of the human pe-ripheral system. Normative vestibular hair cell data.Ann. Otol. Rhinol. Laryngol. 181(Suppl.): 3–13.

31. Nashner, L.M., F.O. Black & C. Wall. 1984. Adapta-tion to altered support and visual conditions duringstance: patients with vestibular deficits. J. Neurosci. 2:536–544.

32. Raivich, G., M. Bohatschek, C.U. Kloss, et al. 1999.Neuroglial activation repertoire in the injured brain:graded response, molecular mechanisms and cuesto physiological function. Brain Res. Rev. 30: 77–105.

33. Raivich, G., L.L. Jones, A. Werner, et al. 1999.Molecular signals for glial activation: pro- and anti-inflammatory cytokines in the injured brain. Acta Neu-

rochir. Suppl. (Wien) 73: 21–30.34. Rosenhall, U. 1973. Degenerative patterns in the ag-

ing human vestibular neuro-epithelia. Acta Otolaryngol.

(Stockh.) 76: 208–220.35. Rosenhall, U. & W. Rubin. 1975. Degenerative

changes in the human vestibular sensory epithelia.Acta Otolaryngol. (Stockh.) 79: 67–80.

36. Schuknecht, H.F. 1969. Cupulolithiasis. Arch. Oto-

laryngol. 90: 765–778.37. Schuknecht, H.F. 1982. Behaviour of the vestibular

nerve following hemilabyrinthectomy. Ann. Otol. Rhi-

nol. Laryngol. 91(Suppl.): 16–32.38. Schuknecht, H.F. 1989. Meniere’s Disease. Vol. 1. Lip-

pincott. Philadelphia.39. Schuknecht, H.F. 1993. Pathology of the periph-

eral vestibular system. In The Vestibule-ocular Reflex and

Vertigo. J.A. Sharpe & H.O. Barber, Eds.: 269–277.Raven Press. New York.

40. Semont, A., G. Freyss & E. Vitte. 1988. Curring theBPPV with a liberatory maneuver. Adv. Otorhinolaryn-

gol. 42: 290–293.41. Tang, Y., I. Lopez & R.W. Baloh. 2002. Age-related

change of the neuronal number in the human me-dial vestibular nucleus: a stereological investigation.J. Vesibt. Res. 11: 357–363.

42. Tighilet, B., J. Leonard & M. Lacour. 1995. Betahis-tine dihydrochloride treatment facilitates vestibularcompensation in the cat. J. Vestib. Res. 5: 53–66.

43. Tighilet, B., S. Trottier, C. Mourre & M. Lacour.2006. Changes in the histaminergic system duringvestibular compensation in the cat. J. Physiol. (Lond.)

573: 723–739.44. Tighilet, B., J.M. Brezun, S. Gustave Dit Duflo, et al.

2007. New neurons in the vestibular nuclei complexafter unilateral vestibular neurectomy in the adultcat. Eur. J. Neurosci. 25: 47–58.

45. Vidal, P.P., C. de Waele, N. Vibert & M. Muhlethaler.1998. Vestibular compensation revisited. Otolaryngol.

Head Neck Surg. 119: 34–42.46. Whipple, R.H., L. Wolfson, C. Derby, et al. 1993. Al-

tered sensory function and balance in older persons.J. Gerontol. 48: 71–76.

47. Wilson, V.J. & G. Melvill Jones. 1979. Mammalian

Vestibular Physiology. Plenum Press. New York.48. Xerri, C. & M. Lacour. 1980. Compensation of

posturo-locomotor deficits after unilateral vestibularneurectomy in the cat. Role of sensorimotor activity.Acta Otolaryngol. (Stockh.) 90: 414–420.

49. Zaehle, T., K. Jordan, T. Wustenberg, et al. 2007. Theneural basis of the egocentric and allocentric spatialframe of reference. Brain Res. 1137: 92–103.

Documents

Tell Me Your Vestibular Deficit, and I’ll Tell You How You’ll Compensate