occlusion effect of earmolds with different venting systems · pdf fileall four types of...

J Am Acad Audiol 16:237–249 (2005)

237

*Department of Audiology, Justus-Liebig University Giessen, Germany; †GN Resound, Taastrup, Denmark

Prof. Dr. Jürgen Kiessling, Dept. of Audiology, Justus-Liebig University Giessen, Feulgenstrasse 10, D-35385 Giessen,Germany; Phone: 49-641-99-43790; Fax: 49-641-99-43799; E-mail: [email protected]

Occlusion Effect of Earmolds withDifferent Venting Systems

Jürgen Kiessling*Barbara Brenner*Charlotte Thunberg Jespersen†Jennifer Groth†Ole Dyrlund Jensen†

Abstract

In this study the occlusion effect was quantified for five types of earmolds withdifferent venting. Nine normal-hearing listeners and ten experienced hearingaid users were provided with conventional earmolds with 1.6 and 2.4 mm circularventing, shell type earmolds with a novel vent design with equivalent cross-sectional vent areas, and nonoccluding soft silicone eartips of a commercialhearing instrument. For all venting systems, the occlusion effect was measuredusing a probe microphone system and subjectively rated in test and retestsessions. The results for both normal-hearing subjects and hearing aid usersshowed that the novel vents caused significantly less occlusion than thetraditional vents. Occlusion effect associated with the soft silicone eartip wascomparable to the nonoccluded ear. Test-retest reproducibility was higher forthe subjective occlusion rating than for the objectively measured occlusion.Perceived occlusion revealed a closer relationship to measured occlusion inthe ear in which the measured occlusion effect was higher (“high OE” ear) thanin the “low OE” ear. As our results suggest that subjective judgment of occlusionis directly related to the acoustic mass of the air column in the vent, theamount of perceived occlusion may be predicted by the vent dimensions.

Key Words: Acoustic mass, earmold, hearing instruments, measured occlusion,novel venting system, occlusion effect, perceived occlusion, prediction ofocclusion effect

Sumario

En este estudio se cuantificó el efecto de oclusión de cinco tipos de moldescon diferentes sistemas de ventilación. A nueve sujetos normo-oyentes y a diezusuarios con experiencia en auxiliares auditivos, se les proporcionaron moldesconvencionales con ventilaciones circulares de 1.6 y 2.4 mm, moldes de tipoconcha con un novedoso diseño de ventilación con áreas transversalesequivalentes de ventilación, y puntas no oclusivas de silicón suave de uninstrumento auditivo comercial. En todos los sistemas de ventilación, se midióel efecto de oclusión utilizando un sistema de sonda con micrófono,clasificándolos subjetivamente en sesiones de evaluación y re-evaluación. Losresultados, tanto para los sujetos normo-oyentes como para los usuarios deauxiliares auditivos, mostraron que las ventilaciones novedosas causaron unefecto de oclusión significativamente menor que los sistemas tradicionales.El efecto de oclusión asociado a las puntas de silicón suave fue comparable

JJoouurrnnaall ooff tthhee AAmmeerriiccaann AAccaaddeemmyy ooff AAuuddiioollooggyy/Volume 16, Number 4, 2005

238

It is widely known that only a smallproportion of candidates for amplificationactually own and wear hearing aids. The

extremely low penetration for hearing aids isobserved worldwide and can be ascribed to anumber of factors, including stigma of hearingaid use, availability of services, and economicaspects. In addition to these factors, wearingcomfort and poor sound quality are commonreasons for not acquiring hearing instruments(Stephens et al, 1990; Davies et al, 1991;Wilson et al, 1993; Saunders and Cienkowski,1996; van den Brink et al, 1996; Cienkowskiand Pimentel, 2001; Jerram and Purdy, 2001;Davis, 2003).

Sound quality is one of the majorcomplaints of individuals who do wearhearing aids, specifically the unnaturalnessof one’s own voice and the disturbance ofother self-generated sounds. Such complaintsare due to the blocking of the ear canal by anearmold or hearing aid shell, creating theso-called occlusion effect (French-SaintGeorge and Barr-Hamilton, 1978; Dempsey,1990; Warland and Tonning, 1993; Brooks,1994). The occlusion effect is the buildup oflow-frequency sound pressure in the residualear canal that occurs when body-conductedsound is trapped (Studebaker and Zachman,1979; Grover and Martin, 1979; Dillon, 1991;Carle et al, 2002). Hearing aid users oftendescribe the subjective effect of occlusion as

sounding like they are talking in a barrel. Inaddition, sounds like chewing, swallowing,throat clearing, and even breathing can beperceived as being so loud that they interferewith the user hearing external sounds.

The degree of perceived occlusion isrelated to the amount of sound pressurebuildup in the ear canal but also to thehearing loss. While wearers with low-frequency hearing threshold levels greaterthan 40 dB are unlikely to be bothered by theocclusion effect (May and Dillon, 1992; Dillon,2001), most hearing aid fitting practices willstill see a great many patients who mayexperience difficulties with occlusion. Analysisof audiometric data from a dispensingpractice’s database of 700 patients revealedthat 59% of the patients demonstratedhearing losses ≤45 dB below 500 Hz(Laureyns, unpublished study), indicating aconsiderable potential for problems withocclusion. There is also evidence to suggestthat having to deal with occlusion-relatedcomplaints is daily fare for hearinginstrument fitters. For example, a survey of4421 hearing aid wearers conducted by Dillonand his colleagues (Dillon et al, 1999) showedthat 27% of the respondents found occlusionto be problematic. Similar findings werereported by Kochkin (2000). Otherinvestigations (French-Saint George andBarr-Hamilton, 1978; MacKenzie et al, 1989;

al del oído no ocluido. La reproducibilidad test/retest fue mayor para laestimación subjetiva de la oclusión que para la oclusión objetivamente medida.La oclusión percibida reveló una relación más cercana con la oclusión medida,en el oído donde el efecto de oclusión medido fue mayor (oído con “alto” OE)que en el oído con “bajo” OE. Dado que nuestros resultados sugieren que eljuicio subjetivo de oclusión está directamente relacionado con la masa acústicade la columna de aire en la ventilación, la magnitud de la oclusión percibidapuede predecirse de las dimensiones de la ventilación.

Palabras Clave: Masa acústica, molde auditivo, instrumento auditivo,oclusión medida, sistema novedoso de ventilación, efecto de oclusión, oclusiónpercibida, predicción del efecto de oclusión

OOcccclluussiioonn EEffffeecctt ooff EEaarrmmoollddss/Kiessling et al

239

Kuk, 1991; Brügel et al, 1992) as well asclinical practice show that the occlusion effectdetermines wearing comfort and consumersatisfaction to a great extent.

What tools are available to the hearingaid fitter to resolve occlusion-relatedcomplaints? During the past decade, manyinvestigators have reported on theeffectiveness of deep fittings (usuallycompletely-in-the-canal instruments) inwhich the otoplastic seals in the bony portionof the ear canal (Killion et al, 1985; Killionet al, 1988; Staab and Finlay, 1991; Staab,1997). Self-generated sounds such as ownvoice mainly cause vibration in thecartilaginous part of the ear canal, as theimpedance mismatch between bone and airhinders efficient input into the bony part.The deep seal works in minimizing occlusioneffect because it reduces vibration of the earcanal walls caused by the body-conductedsound. Another potential advantage of deepcanal fittings is that tendency to acousticfeedback can also be reduced. Drawbacks tothis approach are that they may causediscomfort to the wearer and necessitatecostly and inconvenient remakes. Spaceconstraints in the ear canal may also precludesuch deep fittings.

An alternative way to reduce theocclusion effect is by use of appropriateventing. This pragmatic approach reducesocclusion effect by allowing self-generatedsounds to escape from the ear canal, and alsoprovides greater physical comfort to thewearer. The main disadvantage of venting isthat acoustic feedback is more likely to occuras vent size is increased, thereby limitingthe amount of useable gain for the fitting.This in turn limits the severity of hearinglosses that can be fit with large vents. Earcanal size may also limit the dimensions ofthe vent.

One important advance in hearinginstrument technology has been thedevelopment of feedback suppressionsystems. Some of these systems providehigher maximum stable gain by the use offeedback cancellation filters (Dyrlund andBisgaard, 1991; Murray and Hanson, 1992;Dyrlund et al, 1994; Henningsen et al, 1994;Greenberg et al, 2000). While this addedfeedback margin may be considered as meansto protect hearing instrument wearers fromever experiencing feedback, it may also beutilized to provide greater venting than thedesired gain settings would otherwise allow.

The availability of this processing has sparkedrenewed interest in venting solutionsdesigned to reduce occlusion.

As widening of the vent diameter islimited by the dimensions of the ear canal, anovel venting system with an extremely shortvent length has been developed under theworking hypothesis that the occlusion effectis mainly determined by the acoustic massand therefore proportional to the effectivevent length. The purpose of this study was toinvestigate the occlusion effect and perceivedocclusion associated with this novel ventdesign and compared to a conventional ventdesign. Specifically, we sought to address thefollowing questions:

• Does this novel venting system give areduced sense of occlusion and/or areduced objectively measuredocclusion effect for normal-hearingand hearing-impaired individualscompared to traditional earmolds withthe same cross-sectional vent diameterarea?

• What is the measured occlusion effectand perception of occlusion with acommercially available soft siliconeeartip for open fittings?

• How reproducible and reliable areobjective measurements andsubjective judgments of occlusion?

• Is there a relationship betweenobjective and subjective occlusion?

• Can the amount of perceived occlusionbe predicted?

MMEETTHHOODD

SSuubbjjeeccttss

The study was carried out with a testgroup of ten hearing-impaired listeners withmild to moderate sloping sensorineural lossand a reference group consisting of ninenormal-hearing individuals. The hearing-impaired subjects were experienced hearingaid users who were accustomed to occlusionof the ear canal by either in-the-ear hearinginstruments or earmolds for at least threeyears. The type of venting they were used towas ranging from traditional 2 mm vents toalmost open fittings. The inclusion criteria forthe test group were hearing threshold levelsbetter than 30 dB at frequencies up to 1 kHz,and between 50 and 90 dB from 4 to 8 kHz.


240

Participants in the reference group were notaccustomed to ear occlusion. Figure 1 showsthe average pure-tone hearing loss of both thetest and the reference groups. In this figurethe data for right and left ears were pooled,as hearing losses were highly symmetrical.

EEaarrmmoollddss

Ear impressions were taken past thesecond bend using a two-component siliconematerial (Heba-Form) applied by a double-cartridge impression gun. One ear impressionper ear was used for all types of customearmolds needed for this study. All subjectswere provided with bilateral earmolds asdescribed below and depicted in Figure 2.All four types of custom earmolds for eachsubject were manufactured to terminate atthe same depth in the ear canal.

• Two pairs of conventional earmoldswith parallel vents, one with 1.6 mmand one with 2.4 mm diameter(Figure 2, left)

• Two pairs of shell type earmolds witha novel vent design, one with a cross-sectional vent area equivalent to acircular vent with a diameter of 1.6mm, and one with a cross-sectionalvent area equivalent to a circular

vent with a diameter of 2.4 mm(medium and large sized FlexVentTM,Figure 2, center)

• One pair of soft silicone eartips usedwith a commercial behind-the-earhearing instrument (ReSoundAIRTM,Figure 2, right).

RRaattiioonnaallee ffoorr NNoovveell VVeenntt DDeessiiggnn

An effective vent allows low-frequencysound to escape from the ear canal, therebyreducing the occlusion effect. Venteffectiveness is determined by the inertia ofthe air column in the vent. The inertia canbe described in terms of acoustic mass, whichis proportional to the effective vent length(geometric length plus end correction) andinversely proportional to the cross-sectionalarea of the vent. Therefore, a vent should beas short and as wide as possible to achieveminimum occlusion effect. In the extremecase it becomes an open fitting. Theseconsiderations led to the novel design (Figure2, center). The FlexVent is used in an earshellproduced in the same manner as shells for in-the-ear hearing instruments. A 1 mm ventplate (available in red for right ears and bluefor left ears), with a hole in the middle, isattached to the canal part of the shell

FFiigguurree 11.. Mean pure-tone audiogram (right and left ears pooled) with standard deviations of the test group(10 hearing aid users with sensorineural hearing loss) and the reference group (9 normal-hearing listeners).HI = hearing impaired. NH = normal hearing.


241

earmold. The FlexVent system consists of achangeable vent insert (also 1 mm thick)carrying the sound tube. The vent insert isavailable in two sizes: medium (with a cross-sectional vent area equivalent to a circularvent with a diameter of 1.6 mm) and large(with a cross-sectional vent area equivalentto a circular vent with a diameter of 2.4 mm).The sound tube is attached to the vent insertand is placed in the vent plate; that is, thepart of the hole in the plate not filled out bythe insert constitutes the vent. The size of theinsert can be chosen according to the neededgain. This solution results in a constant ventdiameter along with a full vent length of 1mm. A previous investigation with this typeof earmold revealed significant improvementsin perceived and measured occlusioncompared to traditional earmolds withequivalent cross-sectional vent area andtermination of ear mold length in the earcanal (Jespersen and Asgaard, 2003). TheFlexVent was tested to verify this finding.

SSoofftt SSiilliiccoonnee EEaarrttiipp

The soft silicone eartip was included tocorroborate previous findings suggesting thatocclusion effect was almost eliminated withthis coupling to the ear canal (Kiessling et al,

2003; Groth and Søndergaard, 2004). Thiseartip system consists of a dome-shaped,medical grade silicone earplug with multiplevents like a sieve and is connected to anextremely thin sound tube. It is snapped ontoa behind-the-ear (BTE) housing as a singleunit. The earplugs and tubes come in avariety of standard sizes, which allows fittingof the majority of adult ears. The retentionof the soft silicone eartip has been proven tobe sufficient in a previous study (Kiessling etal, 2003).

PPrroocceedduurree

Subjects were invited to the clinic forthree appointments. At the first session, acase history was recorded and otoscopy, pure-tone audiometry, tympanometry, and bilateralear impressions were performed. Then, a setof traditionally vented earmolds and FlexVentearmolds were manufactured, and appropriatesoft silicone eartips were selected.

In the second session, the occlusion effectwas measured objectively, and subjectiveratings of naturalness of own voice weredetermined for each type of earmold. Allmeasurements and subjective evaluationswere performed in a sound-treatedaudiometric test booth. Objective occlusion

FFiigguurree 22.. Types of earmolds used: traditional earmold with venting (left), shell type earmold with flexible vent(center), nonoccluding soft silicone eartip of a commercial hearing aid (right).


242

measurements were performed with the GNOtometrics VisibleSpeech probe microphonesystem. After the probe tube was placed in theear canal, a reference response of the subject’sown voice was recorded for the nonoccludedear. For this purpose, the subject was askedto vocalize the vowel /ee/ at about 75 dB SPLfor a couple of seconds until the measurementsystem had averaged a stable frequencyresponse. The stability of the sound pressurelevel produced by the subjects was monitoredby the VisibleSpeech unit. Then, the sameprocedure was repeated for each type ofearmold connected to a BTE hearinginstrument in the turned-off position. Theresult of this measurement shows thedifference in frequency response compared tothe nonoccluded condition. Finally, thenonoccluded condition was measured againto verify that the position of the probe tubewas unchanged; that is, the frequencyresponse of the open ear canal was 0 ± 3 dB.The procedure was repeated on the otherear. In this study, we quantified the occlusioneffect in the following way: the three-frequency average (250, 500, and 750 Hz) ofthe real ear sound pressure level for the testcondition minus the three-frequency averageof the real ear sound pressure level for thenonoccluded condition.

To obtain a subjective measure of theperceived occlusion, the participants were askedto rate the naturalness of their own voice whilereading a text sample aloud. They wereencouraged to read the text as many times asneeded to make their judgments. For thisexperiment, the participants were blinded as faras possible concerning the conditions, which wereabsolutely rated on a 10-point scale (0: verynatural sound of own voice … 9: very unnaturalsound of own voice). For the rating procedure,earmold configurations were paired to achieve acertain contrast between related conditions as truepair comparisons were impossible due to theobjective of the study: 1.6 mm traditionalvent/medium FlexVent, 2.4 mm traditionalvent/large FlexVent, and open soft siliconeeartip/nonoccluded ear canal. The sequence of thepaired conditions as well as the sequence withineach pair was randomized to avoid order effects.

The third visit took place about twoweeks after the second appointment. At thissession, both the objective occlusionmeasurements and the subjective occlusionratings were repeated to check test-retestreliability of the data.

SSTTAATTIISSTTIICCSS

The Wilcoxon-Signed-Rank test wasemployed to test for significance between

relevant pairs of conditions for both measuredand rated occlusion, and two-sided p-valueswere calculated. Pearson correlationcoefficients and coefficients of determinationr2 were calculated to describe, respectively,test-retest reliability and relationshipsbetween measured and related occlusion.

RREESSUULLTTSS

MMeeaassuurreedd OOcccclluussiioonn

The objective occlusion measure variedfrom -8 to 25 dB over all subjects andconditions. Mean values and standarddeviations for test and retest are shown inFigure 3. Because this objectively measureddata was not dependent on the presence ordegree of hearing loss, data for hearing-impaired and normal-hearing listeners werepooled. The average measured occlusion fortraditional earmolds with 1.6 mm circularventing was approximately 12 dB whereasFlexVents with an equivalent cross-sectionalvent area resulted in only about 5 dB averagemeasured occlusion. A similar relationshipwas found for larger vents with 2.4 mm cross-sectional vent areas, with a measuredocclusion effect of about 10 dB for traditionalearmolds and 3 dB for the FlexVent. The softsilicone eartip produced an occlusion effectcomparable to the nonoccluded ear. TheWilcoxon test revealed a significant difference(p < 0.01) between the following pairs of testconditions: medium FlexVent versustraditional circular 1.6 mm vent, largeFlexVent versus 2.4 mm circular traditionalvent, and soft silicone eartip versus largeFlexVent. There was no significant differencein measured occlusion between the softsilicone eartip and nonoccluded conditions.

RRaatteedd OOcccclluussiioonn

Figure 4 shows average results andstandard deviations for naturalness of ownvoice ratings (rated occlusion plottedupwards) for both the reference (normalhearing [NH]: upper panel) and the test group(hearing impaired [HI]: lower panel). Forboth populations, the FlexVents resulted in


243

significantly (p < 0.01) reduced perceivedocclusion compared to traditionally ventedearmolds with equivalent cross-sectional ventareas. Also for both groups, perceivedocclusion effect associated with the softsilicone eartip was significantly (p < 0.01)lower than for any other venting system

tested here and yielded comparable resultsto the nonoccluded condition.

TTeesstt--RReetteesstt RReelliiaabbiilliittyy

Average test (Figure 3, white columns)and retest (Figure 3, gray columns)

FFiigguurree 33.. Mean measured occlusion (three-frequency average at 250, 500, and 750 Hz) with standard devia-tions for different types of earmolds for all subjects pooled (n = 19). Test-retest results are given by white columnsand retest results by gray columns.

FFiigguurree 44.. Mean rated occlusion (on a scale 0 … 9) with standard deviations for different types of earmolds forthe normal-hearing (NH) listeners (upper panel) and the hearing-impaired (HI) subjects (lower panel). Test resultsare given by white and retest results by gray columns.


244

measurements showed similar results.However, individual test-retest reliability forthe measured occlusion was fairly poor asevidenced by the scattergram in Figure 5containing all data for hearing-impaired andnormal-hearing test persons over allconditions. The test-retest reliability isdescribed by a coefficient of determination of0.429 for a linear relation.

As was the case for the objectivelymeasured occlusion, the average results forthe subjective ratings were replicated at theretest session, as can be seen in Figure 4 bycomparing the white columns (test) with thegray columns (retest) for each condition. Forthe hearing-impaired participants, the test-retest reliability was given by a coefficient ofdetermination of 0.823 for a linear relation.For the normal-hearing group, the coefficientof determination for a linear relation wasdetermined to be 0.782.

RReellaattiioonn bbeettwweeeenn MMeeaassuurreedd aannddRRaatteedd OOcccclluussiioonn

The issue of how rated occlusion is relatedto the measured effect raises the question ofwhether the rated occlusion is predominantlydetermined by the ear in which the measuredocclusion effect is higher (“high OE” ear) orby the other (“low OE”) ear. One mightassume that the “high OE” ear determines the

perceived occlusion effect. To check thishypothesis, we plotted rated occlusion as afunction of measured occlusion in the “highOE” ear and as a function of measuredocclusion in the “low OE” ear. Figure 6 showsthis relationship for the ear where themeasured occlusion was greatest (“high OE”ear) in the test session. All data from hearing-impaired and normal-hearing listeners arepooled. The coefficient of determination forthe relationship between the “high OE” earand rated occlusion was 0.461, while it was0.331 for the “low OE” ear. Coefficients ofdetermination for a linear relationshipbetween both variables indicate that thecorrelation of rated occlusion with measuredocclusion in the “high OE” ear is slightlyhigher than the one for the “low OE” ear forboth test and retest measurements.

DDIISSCCUUSSSSIIOONN

We wanted to confirm previous findingsshowing that the FlexVents provide a

reduced sense of occlusion and reducedmeasured occlusion effect compared totraditional earmolds with equal cross-sectional vent area (Jespersen and Asgaard,2003). Our results clearly support that thisis the case. Average measures of occlusionwere 7 dB lower for the FlexVents than forthe traditional vents. These results were

FFiigguurree 55.. Test-retest scattergram for measured occlusion and all subjects pooled (n = 19). Test-retest reliabilityis characterized by a coefficient of determination in the order of 0.429.


245

reflected in the subjective ratings in thatboth the normal-hearing and hearing-impaired groups rated the FlexVents asproviding a more natural sounding quality ofown voice. None of the participants madeuse of the entire rating scale when ratingtheir own voice with the FlexVents, whilethere were maximum ratings forunnaturalness with the traditional earmolds.

A reason for including normal-hearingparticipants in this study was to represent a“worse case” scenario. It was hypothesizedthat the normal-hearing participants wouldbe more sensitive to the effects of occlusionby virtue of their hearing acuity. However, thesubjective ratings of the hearing-impairedand normal-hearing groups turned out to beremarkably similar. The present data supportthat normal-hearing persons perceive andrate occlusion effects in the same way asexperienced hearing aid users do. Thissuggests that the perceptual effects ofocclusion are not affected by being accustomedto occlusion. This is in agreement withHansen (1997), who found that hearing aidusers do not adapt to the occlusion effect.

Another reason for including normal-hearing participants was to ensure thatreliable, reproducible judgments were made.Grover and Martin (1979) performedsemantic differential tests for different typesof venting and reported that hearing-impaired

subjects were less able to make systematicjudgments on sound quality than a controlgroup consisting of normal-hearingindividuals. In contrast to their findings, ourdata show that the test-retest reliability forthe hearing impaired is marginally higherthan for normal-hearing listeners. Thediscrepancy with Grover and Martin (1979)may be explained by the fact that our hearing-impaired test persons had near normalhearing up to 1 kHz, whereas their subjectshad flat losses ranging from 30–70 dB overthe speech frequency range. Thus our subjectswould have been better able to discriminatelow-frequency vent-related effects becauseof their better hearing. Our results do notindicate that the hearing impaired are lessreliable in their judgments of occlusion thannormal-hearing individuals.

One of the research questions was relatedto the measured and perceived occlusionassociated with a commercially available softsilicone eartip for open fittings. This wasmotivated by earlier findings that patients fitwith this device had no occlusion-relatedcomplaints (Kiessling et al, 2003) and thatreal ear occluded gain for this ear tip wasnearly identical to real ear unaided gain(Groth and Søndergaard, 2004). The presentresults revealed no significant differences inmeasured occlusion between the eartip andnonoccluded conditions, nor were there any

FFiigguurree 66.. Scattergram for rated occlusion versus measured occlusion in the “high OE” ear for all subjects pooled(n = 19). Measured occlusion in the “high OE” ear explains about 46% of the variation of perceived occlusion.


246

differences between these conditions for thesubjective ratings. This supports that thesoft silicone eartip is indeed nonoccluding.

Although previous studies have notdemonstrated a clear relation between sensedand objective occlusion (May and Dillon,1992; Biering-Sørensen et al, 1994), it hasbeen put forth that objective measures ofocclusion effect are a useful clinical tool fortroubleshooting occlusion-related complaints(Mueller, 1992). For this to be true requiresthe ability to perform reliable measurements.We attempted to shed light on this issue byreplicating our measurements at a latersession. Since the same examiner performedall measurements, we expected the test-retestreliability to be high. Surprisingly, it turnedout to be only around 0.4. In contrast, the test-retest reliability of the subjective ratings ofnaturalness of own voice were much higher,which suggests that what patients reportabout their experience of occlusion may be abetter indication of vent effectiveness than asingle measurement of occlusion effect.Measurements of occlusion are subject to thesame sorts of procedural errors as other realear measurements, with shifting of the probetube arguably being the greatest source oferror. It is also the case that the soundpressure level in the ear canal for self-generated sounds is affected by other factorsthan vent effectiveness alone. For example,Carle and colleagues (2002) demonstratedthat increased middle ear impedance canproduce a greater occlusion effect. Although

this would be measurable, it probably wouldnot lead to increased occlusion-relatedcomplaints.

An interesting observation regarding theobjective measurements of occlusion was therather large individual variation in values forthe same condition. In some cases, negativevalues compared to the nonoccluded conditionwere attained. An obvious explanation forthis could be a shift in probe tube positionbetween measurements despite efforts tocontrol for this. However, this could also beaccounted for by individual resonances andthe fact that the net effect of low-frequencysound transmission into and out of the ventcan be either negative or positive (Dillon,1991).

Discussion of the reliability of objectivemeasurement of occlusion and subjectivejudgments leads to the question of whetherthere is a relationship between the two. Ourdata indicated that objective and subjectiveocclusion are positively correlated. It wasalso the case that the average rating ofnaturalness of own voice was more closelyrelated to the measured occlusion effect in theear where the occlusion effect was highest.The measured occlusion in the ear with thehigher occlusion effect explains about 46% ofthe variation of perceived occlusion. Thisfinding supports the assumption that theexperienced occlusion is mainly determinedby the degree of occlusion in the moreoccluded ear. However, there is still a largedegree of variation not explained by the

FFiigguurree 77.. Rated occlusion averaged over all subjects (taken from Figure 4) as a function of the acoustic massof the air column in the vent. Perceived occlusion can be predicted from the acoustic mass of the vent accord-ing to the equation given in the lower right corner.


247

measured occlusion effect. This would appearto be accounted for by the lower reliability ofthe objective measurements compared to thesubjective judgments.

A final question to be addressed in thisstudy was whether or not perceived occlusioncould be predicted for individuals with goodlow-frequency hearing. In light of thediscussion on reliability of objective occlusionmeasurements and subjective judgments, itis clear that the measured occlusion effect isnot a strong predictor of perceived occlusionin individual cases. However, as there are nodata points in the lower right corner of thescattergram in Figure 6, a single-edgedestimation is possible. For example, if themeasured occlusion effect is 15 dB or higher,perceived occlusion is expected to be threescaling units or above, indicating thatperception of own voice as “natural” soundingis highly unlikely to occur.

We propose that an alternative way topredict perceived occlusion is to base it ongeometric vent dimensions. As mentionedpreviously in this paper, the effectiveness ofa vent in allowing free passage of low-frequency sound is determined by the acousticmass of the air column in the vent. Theincrease in ear canal sound pressure level forself-generated sound is also mainlydetermined by this acoustic mass (Ma), whichis given by:

Ma = Constant • Vent length(Vent diameter)2

where the constant contains the density of theair. To prove this assumption, we have plottedthe rated occlusion taken as the averagesfrom Figure 4 as a function of the acousticmass of the vent types used in this study. Theacoustic mass is calculated from the equationabove utilizing the average vent dimensionsand taking end corrections for the vent lengthinto account (Dillon 2001). For a semi-logarithmic plot (Figure 7), we find a closelinear relation between both variables: theacoustic mass explains about 97% of thevariance of the perceived occlusion. Otherfactors, such as middle ear compliance asstudied by Carle et al (2002), seem to play aminor role for this group of subjects withnormal middle ear function.

Given this relation, the perceived occlusioncan be predicted by the geometric dimensionsof the vent. This relation is illustrated inFigure 8. Plotting the calculated occlusion asfunction of vent length and vent diameterresults in a three-dimensional surface. Thisdiagram can be used to assess the expectedocclusion from the vent dimensions when theearmold is manufactured; for example, witha long narrow vent (rear corner of Figure 8),one will never get an acceptable result interms of experienced occlusion. We havemarked the position of the types of earmoldsused in this study on this surface as blacksquares to illustrate how the vent dimensionsaffect the occlusion.

FFiigguurree 88.. Three-dimensional plot of rated occlusion as a function of vent length and diameter calculated accord-ing to the equation given in Figure 7. The positions of the types of earmolds used in this study are marked onthis surface as black squares.


248

CCOONNCCLLUUSSIIOONNSS

• The FlexVent earmolds are significantly(p < 0.01) less occluding than traditionalearmolds with equivalent cross-sectionalvent area.

• The soft silicone eartip provides anonoccluding hearing aid fitting.

• For individuals, subjective evaluationappears to be a more reliable indicator ofocclusion than objective measurement ofthe occlusion effect.

• Perception of occlusion appears to be wellpredicted by the acoustic mass associatedwith the vent in the otoplastic.

AAcckknnoowwlleeddggmmeenntt.. The authors thank two anony-mous reviewers for helpful comments on an earlierversion of this manuscript. This study was supportedby GN Resound.

RREEFFEERREENNCCEESS

Biering-Sørensen M, Pedersen F, Parving A. (1994)Is there a relationship between the acoustic occlusioneffect and the sensation of occlusion? Scand Audiol23:111–116.

Brooks DN. (1994) Some factors influencing choice oftype of hearing aid in the UK: behind-the-ear or in-the-ear. Br J Audiol 28:91–98.

Brügel FJ, Schorn K, Hofer R. (1992) Whatimprovements can be gained by additional boring inhearing aid fitting? Laryngorhinootologie 71:79–82.

Carle R, Laugesen S, Nielsen C. (2002) Observationson the relations among occlusion effect, compliance,and vent size. J Am Acad Audiol 13:25–37.

Cienkowski KM, Pimentel V. (2001) The hearing aid'effect' revisited in young adults. Br J Audiol35:289–295.

Davies JE, John DG, Stephens SD. (1991)Intermediate hearing tests as predictors of hearingaid acceptance. Clin Otolaryngol 16:76–83.

Davis A. (2003) Population study of the ability tobenefit from amplification and the provision of ahearing aid in 55–74-year-old first-time hearing aidusers. Int J Audiol 42(Suppl. 2):2S39–52.

Dempsey JJ. (1990) The occlusion effect created bycustom canal hearing aids. Am J Otol 11:44–46.

Dillon H. (1991) Allowing for real ear venting effectswhen selecting the coupler gain of hearing aids. EarHear 12:406–416.

Dillon H. (2001) Hearing Aids. New York: Thieme.

Dillon H, Birtles G, Lovegrove R. (1999) Measuringthe outcomes of a national rehabilitation program:normative data for the Client Oriented Scale ofImprovement (COSI) and the Hearing Aid User´sQuestionnaire (HAUQ). J Am Acad Audiol 10: 67–79.

Dyrlund O, Bisgaard N. (1991) Acoustic feedbackmargin improvements in hearing instruments usinga prototype DFS (digital feedback suppression) system.Scand Audiol 20:49–53.

Dyrlund O, Henningsen LB, Bisgaard N, Jensen JH.(1994) Digital feedback suppression (DFS):characterization of feedback-margin improvements ina DFS hearing instrument. Scand Audiol 23:135–138.

French-Saint George M, Barr-Hamilton RM. (1978)Relief of the occluded ear sensation to improve earmoldcomfort. J Am Audiol Soc 4:30–35.

Greenberg JE, Zurek PM, Brantley M. (2000)Evaluation of feedback-reduction algorithms forhearing aids. J Acoust Soc Am 108:2366–2376.

Groth J, Søndergaard MB. (2004) Disturbance causedby varying propagation delay in non-occluding hearingaid fittings. Int J Audiol 43:594–599.

Grover BC, Martin MC. (1979) Physical and subjectivecorrelates of earmould occlusion. Audiology18:335–350.

Hansen MØ. (1997) Occlusion effects. Part I. Hearingaid users' experiences of the occlusion effect comparedto the real ear sound level. Ph.D. thesis, TechnicalUniversity of Denmark.

Henningsen LB, Dyrlund O, Bisgaard N, Brink B.(1994) Digital feedback suppression (DFS). Clinicalexperiences when fitting a DFS hearing instrumenton children. Scand Audiol 23:117–122.

Jerram JC, Purdy SC. (2001) Technology, expectations,and adjustment to hearing loss: predictors of hearingaid outcome. J Am Acad Audiol 12:64–79.

Jespersen CT, Asgaard J. (2003) Measuring occlusioneffect with a new vent design including different ventdiameters. Paper presented at 6th EuropeanFederation of Audiology Societies Congress, Crete,Greece.

Kiessling J, Margolf-Hackl S, Geller S, ØstergaardOlsen S. (2003) Field test of a non-occluding hearinginstrument. Hear J 56:36–41.

Killion MC, DeVilbiss EL, Stewart JK. (1988) Anearplug with uniform 15-dB attenuation. Hear J 41:14–17.

Killion MC, Wilber LA, Gudmundsen GI. (1985) Insertearphones for more interaural attenuation. HearInstrum 36:346–347.

Kochkin S. (2000) MarkeTrak V. Why my hearingaids are in the drawer: the consumer´s perspective.Hear J 53:34–42.

Kuk FK. (1991) Perceptual consequence of vents inhearing aids. Br J Audiol 25:163–169.

MacKenzie K, Browning GG, McClymont LG. (1989)Relationship between earmould venting, comfort andfeedback. Br J Audiol 23:335–337.

May A, Dillon H. (1992) Comparison of physicalmeasurements of the occlusion effect with subjectivereports. Paper presented at the Audiological SocietyConference (organized by the National AcousticLaboratories), Barossa Valley, South Australia.


249

Mueller HG. (1992) Assessment of fittingarrangements, special circuitry, and features. In:Mueller HG, Hawkins DB, Northern JL. ProbeMicrophone Measurements: Hearing Aid Selectionand Assessment. San Diego: Singular PublishingGroup, 201–225.

Murray DJ, Hanson JV. (1992) Application of digitalsignal processing to hearing aids: a critical survey. JAm Acad Audiol 3:145–152.

Saunders GH, Cienkowski KM. (1996) Refinement andpsychometric evaluation of the Attitudes Toward Lossof Hearing Questionnaire. Ear Hear 17:505–519.

Staab WJ. (1997) Deep canal hearing aids. In: ChasinM, ed. CIC Handbook. San Diego: Singular PublishingGroup, 1–30.

Staab WJ, Finlay B. (1991) A fitting rationale fordeep canal hearing instruments. Hear Instrum42:6–10, 48.

Stephens SD, Meredith R, Callaghan DE, Hogan S,Rayment A. (1990) Early intervention andrehabilitation: factors influencing outcome. ActaOtolaryngol Suppl 476:221–225.

Studebaker GA, Zachman TA. (1979) Investigation ofthe acoustics of earmold vents. J Acoust Soc Am47:1107–1115.

van den Brink RH, Wit HP, Kempen GI, van HeuvelenMJ. (1996) Attitude and help-seeking for hearingimpairment. Br J Audiol 30:313–324.

Warland A, Tonning F. (1993) Factors to considerwhen in-the-canal hearing instruments are used inaural rehabilitation. Scand Audiol 22:47–55.

Wilson PS, Fleming DM, Donaldson I. (1993)Prevalence of hearing loss among people aged 65years and over: screening and hearing aid provision.Br J Gen Pract 43:406–409.

occlusion effect of earmolds with different venting systems · pdf fileall four types of...

Documents