Benefit of a Commercially Available Cochlear ImplantProcessor With Dual-Microphone Beamforming:

A Multi-Center Study

*Jace Wolfe, †Aaron Parkinson, ‡Erin C. Schafer, §Jan Gilden, kKathy Rehwinkel,¶Jennifer Mansanares, ¶Elizabeth Coughlan, ¶Jennifer Wright,

¶Jennifer Torres, and #Stephanie Gannaway

*Hearts for Hearing, Oklahoma City, Oklahoma; ÞCochlear Americas; þUniversity of North Texas;§Houston Ear Research Foundation, Houston, Texas; kCenter for Hearing and Balance, St. Louis, Missouri;

¶Denver Ear Associates, Denver, Colorado; and #University of Oklahoma Health Sciences Center,Oklahoma City, Oklahoma, U.S.A.

Objective: Previous research shows that cochlear implant usersexperience significant difficulty with speech perception in noisylistening situations. There is a paucity of research evaluating thepotential improvement in speech recognition in noise providedby a dual-microphone directional system in a commercial im-plant sound processor. The primary objective of this study wasto compare speech recognition in quiet and in noise for theNucleus Freedom and Nucleus 5 CP810 sound processors setto the manufacturer’s default user programs for quiet and noisyenvironments.Research Design: Crossover with repeated-measures design.Setting: This multi-center study was conducted across fourcochlear implant clinics in the United States.Patients: Thirty-five adults with unilateral Nucleus Freedomcochlear implants. All subjects had used their cochlear implantfor at least 6 months and had substantial open-set word rec-ognition as evidenced by a score of at least 40% correct onthe Consonant-Nucleus-Consonant (CNC) monosyllabic wordrecognition test in quiet.Intervention: All subjects (previous users of the NucleusFreedom sound processor) were fitted with the Nucleus 5 soundprocessor. Performance was assessed while these subjects used

each sound processor in the default user program the manu-facturer recommends for quiet and noisy conditions.Main Outcome Measures: Speech recognition was assessedwith CNC monosyllabic words in quiet and sentences in noisefrom the BKB-SIN (Bamford-Kowal-Bench Sentences in Noise)test. The data were analyzed with descriptive statistics and per-formance with each processor in each listening condition wascompared using a repeated-measures analysis of variance.Results:Word recognition in quiet was significantly better withthe Nucleus 5 sound processor when compared to performancewith the Nucleus Freedom processor. In noise, the Nucleus5 sound processor also provided a significant improvement inspeech recognition relative to the performance with the NucleusFreedom.Conclusion: The results of the study suggest that the Nucleus5 sound processor provides significantly better speech recog-nition in quiet and in noise when compared with performancewith the Nucleus Freedom processor. Key Words: CochlearimplantsVDirectional microphone systemsVNoiseVSpeechrecognition.

Otol Neurotol 33:553Y560, 2012.

Cochlear implants are highly successful sensory pros-thetic devices for many adults who have postlingualdeafness. For example, many adult recipients perform atceiling (i.e., 100%) on tests of sentence recognition inquiet (1Y3), and most adult users successfully converseover the telephone (4,5). However, most recipients con-tinue to experience considerable difficulty understand-ing speech in the presence of competing noise (6Y8). Forinstance, for persons with normal hearing, previous re-search using spatially coincident signals (i.e., speech and

Address correspondence and reprint requests to Jace Wolfe, Ph.D.,3525 NW 56th Street, Suite A-150, Oklahoma City, OK 73112; E-mail:jace.wolfe@heartsforhearing.orgThese data have not been published in another journal.This study was funded by a grant from Cochlear Americas. Aaron

Parkinson, Ph.D., is an employee of Cochlear Americas, the manu-facturer of the cochlear implant technology evaluated in this study.However, the first author determined the study design and analyzed thedata, and Dr. Parkinson also was not involved with data collection.The authors disclose no conflicts of interest.Supplemental digital content is available in the text.

Otology & Neurotology33:553Y560 � 2012, Otology & Neurotology, Inc.

553

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

noise at zero degrees azimuth) has shown the speech-to-noise ratio required for 50% correct recognition of sen-tences (Speech Reception Thresholds [SRT]) is between+1 to j5 dB (9Y11). In contrast, cochlear implant re-cipients typically require an SRT of at least +5 dB andoften +10 to +15 dB for similar performance (12).

The difficulties cochlear implant users experiencein noise can be attributed to several factors, which arediscussed in the text of Supplemental Digital Con-tent 1: ‘‘Factors contributing to a cochlear implantuser’s difficulty with speech understanding in noise’’(http://links.lww.com/MAO/A108) (7,8,13Y24).

COCHLEAR IMPLANT NOISE PROCESSINGAND SPEECH RECOGNITION IN NOISE

Implementation of signal processing advances anddirectional microphones with beam forming possessthe greatest potential to consistently improve perfor-mance in noise for all users of cochlear implants. In-deed, a number of recent studies have examined therelationship between these advances and performancein competing in noise. A thorough review of thesestudies can be found in Supplemental Digital Content2, entitled ‘‘Advances in Cochlear Implant Technol-ogy’’ (http://links.lww.com/MAO/A109) (6Y8,25Y32).

THE COCHLEAR NUCLEUS 5 CP810SOUND PROCESSOR

The Nucleus 5 sound processor received U.S. Food andDrug Administration approval for commercial use inSeptember of 2009. Numerous differences exist betweenthe Nucleus Freedom and Nucleus 5 sound processors,and some of these differences may impact speech rec-ognition in realistic listening situations. First, the Nucleus5 sound processor possesses 2 omni-directional micro-phones, which are calibrated by Cochlear Limited so thatthey are matched in phase and sensitivity within +/j 1 dB.Edwards et al. (33) reported that a sensitivity differencebetween microphones of greater than 2 dB can cause adirectional system to operate with an omni-directional re-sponse, so the process of precise microphone matchingand calibration is imperative to promote optimal direc-tional function. The Nucleus 5 processor also allows foruse of the beam strategy, but the reported polar plotpattern of the Nucleus 5 provides greater attenuationof sounds in the rear hemisphere when compared withthe reported polar plot pattern of the Nucleus Freedomprocessor (34).

Second, the Nucleus 5 sound processor offers an al-ternative directional mode referred to as ‘‘Zoom.’’ Thepreprocessing scheme uses the dual-microphone direc-tional system to provide a fixed, super-cardioid direc-tional response, which provides significant attenuationfor sounds in the rear hemisphere. The polar plot pat-terns for the Zoom scheme and the conventional direc-

tional response in the Everyday program are provided inFigure 1.

Furthermore, the default user programs are different forthe Nucleus 5 and Freedom sound processors. Consistentwith the results of Wolfe et al. (8) and Gifford and Revit(32), which both showed that autosensitivity control(ASC) + adaptive dynamic range optimization (ADRO)provided equivalent speech recognition in quiet and betterspeech recognition in noise when compared with ADROalone, the ‘‘Everyday’’ program of the Nucleus 5 soundprocessor uses ASC+ADRO preprocessing. The ‘‘Noise’’programs in the Nucleus 5 and Freedom sound proces-sors also differ. The Nucleus 5 ‘‘Noise’’ program usesZoom+ASC+ADRO and is recommended for listeningsituations in which the implant user is facing the signal ofinterest, and the competing noise is predominantly behindthe user.

Finally, systematic improvements in the architecture ofthe signal processing chip within the Nucleus 5 soundprocessor were implemented with the intention of en-hancing speech recognition in quiet and in noise. Theintended consequence is that the ASC and directionalpreprocessing systems will be more effective at enhanc-ing the SNR in noisy listening conditions.

Primary Objectives of This StudyGiven the aforementioned differences between the

Nucleus 5 and Freedom processors, the primary objec-tives of this study were as follows: 1) to compare speechrecognition performance in quiet and in noise between theFreedom and the Nucleus 5 processors, and 2) to comparespeech recognition performance in quiet and in noisewhen using the ‘‘Everyday’’ and ‘‘Noise’’ programs in the

FIG. 1. Directional sensitivity of the Nucleus 5 microphonesystem for a sound processor worn on the right ear. The con-ventional microphone response is indicated in blue and the Zoomresponse in red.

554 J. WOLFE ET AL.

Otology & Neurotology, Vol. 33, No. 4, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

Freedom and Nucleus 5 sound processors to validate thenew recommended settings.

MATERIALS AND METHODS

This study was conducted as a single subject, repeated-measures experiment. This design was selected to accommo-date the variability in demographics and outcomes inherent inthe population of persons with hearing loss and cochlear im-plants. A total of 35 subjects were tested at 4 cochlear implantcenters (Table 1). The examiners and subjects were not blindedto the treatment variable as it is not possible to conceal thepresence or absence of the 2 different sound processors used

in the study. This research study was approved by the WesternInstitutional Review Board. The study protocol, along withpotential risks and benefits, were thoroughly explained to eachsubject, and the examiners obtained informed consent beforethe completion of the fitting of the Nucleus 5 sound processorand study assessment.

SubjectsThirty-five unilateral cochlear implant recipients of the Nu-

cleus Freedom cochlear implant system were included in thisstudy. Subject demographics are provided in Table 2. Eachsubject met the following inclusion/exclusion criteria:Inclusion criteria: 1) at least 18 years of age at the com-

mencement of the study; 2) unilaterally implanted with a Nu-cleus Freedom cochlear implant system; 3) previous users ofthe Freedom sound processor and the advanced combinationencoder signal coding strategy; 4) open-set word recognition ofat least 40% correct on the CNC word recognition test. This wasnecessary to reduce the likelihood of floor effects on speechrecognition in noise and to subsequently allow for a comparisonof performance between the Freedom and Nucleus 5 processors;5) use of spoken English as primary mode of communication;and 6) after completion of subject consent and presentation of

TABLE 2. Demographic information about participants

Subject AgeAge onset ofloss, right/left

Age onset ofsevere-profound loss, right/left

Etiologyof loss, right/left

1 35.2 8.0/8.0 27.0/27.0 Familial/Familial2 52.2 22.0/22.0 * Familial/Familial3 52.1 27.0/27.0 28.0/28.0 Unk/Unk4 47.1 * 18.0/18.0 Unk/Unk5 37.0 0.8/0.8 4.0/4.0 Immunization Reaction/Immunization Reaction6 21.2 8.0/8.0 8.0/8.0 Unk/Unk7 30.3 24.9/24.9 24.9/24.9 Trauma/Trauma8 55.8 45.0/46.0 45.0/46.0 Unk/Unk9 58.0 12.0/12.0 47.0/47.0 Unk/Unk10 53.4 25.0/25.0 25.0/25.0 Possible Virus/Possible Virus11 40.6 5.0/5.0 24.0/24.0 Trauma/Trauma12 74.4 44.0/44.0 64.0/64.0 Unk/Unk13 84.9 55.0/55.0 80.0/80.0 Familial/Familial14 36.1 2.5/2.5 * Unk/Unk15 65.9 45.8/45.8 * Menieres/Meniere’s16 75.8 36.7/36.7 53.8/53.8 Noise/Noise17 70.2 65.0/65.0 67.0/67.0 Unk/Unk18 68.1 50.0/45.0 64.0/55.0 Menieres/Meniere’s19 62.2 5.0/5.0 56.0/56.0 Unk/Unk20 79.1 60.0/60.0 73.0/73.0 Unk/Unk21 51.2 31.0/31.0 39.0/39.0 Unk/Unk22 72.6 50.0/50.0 65.0/65.0 Noise/Noise23 53.8 34.0/34.0 36.0/36.0 Familial/Familial24 51.6 26.0/26.0 40.0/40.0 Unk/Unk25 66.2 61.0/61.0 61.0/61.0 Ototoxic/Ototoxic26 31.3 23.0/23.0 28.0/23.0 Menieres/Meniere’s27 61.4 53.0/56.0 60.0/56.0 Menieres/Meniere’s28 39.2 24.0/24.0 29.0/29.0 Familial/Familial29 64.2 58.7/58.7 58.7/58.7 Unk/Unk30 68.0 40.0/40.0 61.0/61.0 Noise/Noise31 59.2 50.0/45.0 55.0/55.0 Menieres/Meniere’s32 51.8 40.0/40.0 40.0/40.0 Viral/Viral33 59.9 30.0/30.0 53.0/53.0 Measles/Measles34 72.8 20.0/19.0 70.0/70.0 Noise/Noise35 74.2 48.0/48.0 70.0/70.0 Noise/NoiseMEAN 56.5 33.2/33.0 46.1/45.5Standard deviation 15.5 18.8/18.7 20.1/19.9

* Indicates did not obtain; Unk, unknown.

TABLE 1. Cochlear implant centers involved in study

Center name Location

Hearts for Hearing Oklahoma City, OKCenter for Hearing and Balance St. Louis, MODenver Ear Associates Denver, COHouston Ear Research Foundation Houston, TX

555COCHLEAR IMPLANT PROCESSOR WITH BEAMFORMING

Otology & Neurotology, Vol. 33, No. 4, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

study protocol, subjects were required to exhibit a willingness toparticipate and to comply with all requirements of the protocol.Exclusion criteria: 1) unrealistic expectations regarding the

potential benefits, risks, and limitations associated with thestudy procedures and/or equipment; and 2) an inability tocomplete all test sessions.

Procedures

Session 1: Processor OptimizationAt the initial study visit, the study protocol was discussed

with each subject, informed consent was completed, and ahearing history examination was completed. The subjects’programs for their Freedom processors were optimized if theclinician or recipient felt performance was less than satisfactoryand reprogramming of the sound processor (i.e., adjustments toT-levels [electrical thresholds] and C-levels [levels of electri-cal stimulation considered to be loud but comfortable]) wasnecessary to improve performance.

Session 2: Fitting of the Nucleus 5 SoundProcessor and Initial Testing

Approximately 1 to 2 weeks after the initial appointment,each subject was fitted with the Nucleus 5 sound processor. Atthe fitting appointment, the Cochlear Custom Sound 3.0 pro-gramming software was used to convert the Freedom MAP to aMAP for the Nucleus 5 processor. All of the fundamental fittingparameters (e.g., stimulation levels, stimulation rate, maxima,frequency allocation) of the Freedom MAP were preserved inthe Nucleus 5 MAP. The power level required to optimallytransmit the signal between the Nucleus 5 sound processor andthe subject’s cochlear implant was determined, and the situation-specific programs of the Nucleus 5 processor were altered toreflect the manufacturer’s transition to new input process-ing schemes for various listening situations. Specifically, the‘‘Everyday’’ program of the Nucleus 5 processor was convertedto include ASC+ADRO input processing rather than the ADRO-only mode used in the ‘‘Everyday’’ program of the Freedomsound processor. Furthermore, the ‘‘Noise’’ program of the Nu-cleus 5 sound processor was altered to include Zoom+ASC+ADRO rather ASC+ADRO alone as used in the Freedom. In-structions pertaining to care, use, and maintenance of the newprocessor also were provided for each subject.Immediately after the fitting of the new processor, speech

recognition in quiet and in the presence of noise at multiplepresentation levels was assessed with the Freedom and Nucleus5 sound processors. The order in which performance with eachsound processor was assessed along with the speech recognitioncondition (e.g., speech recognition in quiet vs. speech recogni-tion in noise) was counter-balanced across subjects.Open-set speech recognition in quiet was evaluated with 2

full lists (50 words each) of CNC monosyllabic words presentedat 60 dBA from a loudspeaker located directly in front of thesubject. The order of the word lists was randomized acrosssubjects. Using data from Skinner et al. (35), who examinedinterlist difficulty of the CNC monosyllabic word recognitiontest, steps were taken to ensure that lists of equivalent difficultywere administered for the 2 different processors. The subjectswere asked to repeat the words that were heard, and a percent-correct score was determined for each condition. Speech rec-ognition in quiet was assessed, whereas the subjects wore eachof the 2 processors with the manufacturer’s recommended‘‘Everyday’’ program, which again, was ADRO alone and

ASC+ADRO for the Freedom and Nucleus 5 sound processors,respectively.Sentence recognition in continuous multi-talker babble was

assessed with the Bamford-Kowal-Bench Speech in Noise(BKB-SIN) test (9). The BKB-SIN utilizes prerecorded SNRsfrom a +21 dB to a -6 dB in order to determine a listener’s SRT,which was calculated using procedures recommended in theBKB-SIN user manual (9). When using the BKB-SIN, the SNRof the first sentence is presented at a +21 dB SNR, and the SNRof each successive sentence deteriorates so that final sentencewithin a list is presented at a -6 dB SNR. Two full list-pairs ofsentences (16Y20 sentences per list-pair) were presented in eachcondition at 2 sentence presentation levels (65 and 75 dBA), andthe number of key words repeated correctly was determined.The order of the BKB-SIN word lists were randomized acrosssubjects. The sentences were presented from a loudspeaker lo-cated directly in front of the subject (0 degrees in the horizontal),whereas the competing noise was delivered from a loudspeakerlocated directly from the side of the implanted ear.The SRT was determined in the conventional ‘‘Noise’’ pro-

gram (e.g., ASC+ADRO) for both the Freedom and Nucleus 5processor. In addition, performance was evaluated with the new‘‘Noise’’ program (e.g., Zoom+ASC+ADRO) of the Nucleus 5sound processor. Following the testing procedures, the subjectsused the Nucleus 5 sound processor with the ‘‘Everyday’’ and‘‘Noise’’ programs for a 2-week period.

Session 3: Follow-up TestingAfter the 2-week period, the aforementioned battery of speech

recognition tests was readministered. Again, the order of testconditions was counterbalanced across subjects. Within eachtest session as well as across test sessions, the examiners ensuredthat the stimulation parameters (i.e., underlying MAP, T- andC-levels, volume control and microphone sensitivity controlsettings, etc.) were identical across the 2 different soundprocessors.

RESULTS

Speech Recognition in QuietAverage speech recognition in quiet on the CNC word

lists is shown in Figure 2. The data from the 2 listening

FIG. 2. Mean CNC word recognition scores obtained at Visits 2and 3 for the Freedom with ADRO and the Nucleus 5 with ASC+ADRO. Note: vertical lines represent 1 standard deviation.

556 J. WOLFE ET AL.

Otology & Neurotology, Vol. 33, No. 4, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

conditions in quiet were analyzed using a 2-way repeated-measures analysis of variance with the independent var-iables of test session (Sessions 2 and 3) and condition(Freedom: ASC+ADRO, Nucleus 5: ASC+ADRO). Ac-cording to the analysis, no significant difference in speechrecognition in quiet was detected across test sessions(F1,34 = 0.05, p = 0.83); however, a significant effect ofcondition was detected (F1,34 = 15.44, p = 0.0004). Therewas no significant interaction effect between test sessionand condition (F1,34 = 0.16, p = 0.69). Post hoc testingusing the Tukey-Kramer multiple comparison test re-vealed significantly better speech recognition in quiet inthe Nucleus 5 with ASC+ADRO condition relative to theFreedom with ADRO condition. In summary, speech rec-ognition in quiet did not improve across the 2 visits, butperformance was significantly better with the ‘‘Everyday’’program of the Nucleus 5 processor when compared withthe ‘‘Everyday’’ program of the Freedom processor. Acrossthe 2 test sessions, the mean CNC word score was 72.7%with the Freedom processor and 75.8% with the Nucleus5 processor.

Because group analyses indicated no effect of testsession, individual CNC scores were averaged across the2 evaluations and plotted in Figure 3. Each point rep-resents an individual subject’s mean score obtained forFreedom, as indicated by reference to the abscissa, as wellthe same individual’s mean score obtained with Nucleus 5,as indicated by reference to the ordinate. Thus, the soliddiagonal indicates identical scores being obtained witheach processor. Scores above the diagonal show improvedperformance with the Nucleus 5 processor, with scoresbelow indicative of better performance with Freedom.Scores outside the curved lines indicate changes that weresignificant based on binomial comparisons (36). Theresults suggest a systematic trend for small gains in CNCword recognition with the Nucleus 5 processor, consistentwith the small gain in mean performance. Six subjects

demonstrated relatively small but significant improve-ment with the Nucleus 5 processor, based on binomialcomparisons, and 1 subject showed a small but significantimprovement with Freedom over Nucleus 5. Twenty-eight of the 35 subjects showed no significant differencein scores with Nucleus 5 and Freedom.

Speech Recognition in NoiseAverage speech recognition in noise on the BKB-SIN

test is shown in Figure 4. These data were analyzed usinga 3-way repeated-measures analysis of variance with theindependent variables of test session (Sessions 2 and 3),signal level (65 and 75 dBA), and condition (Freedom:ASC+ADRO, Nucleus 5: ASC+ADRO, Nucleus 5:Zoom+ASC+ADRO). The analysis revealed no signif-icant effect of test session (F1,34 = 0.65, p = 0.42), nosignificant effect of signal level (F1,34 = 0.40, p = 0.53),and a significant effect of test condition (F2,34 = 1926.2,p G 0.000001). No 2- or 3-way interaction effects weredetected ( p 9 0.05). A post hoc Tukey-Kramer multiplecomparison test showed that each condition was sig-nificantly different from the others (p G 0.05) with theNucleus 5 with Zoom+ASC+ADRO providing the bestperformance followed by the Nucleus 5 with ASC+ADRO and Freedom with ASC+ADRO conditions. Over-all, the analysis suggested that regardless of test sessionor signal level, users received the greatest benefit inspeech recognition in noise from the new ‘‘Noise’’ pro-gram on the Nucleus 5 processor. Collapsing scoresobtained across test session and presentation level, themean speech reception thresholds (SRTs) in noise were8.6 and 7.8 dB when converting a user from Freedomwith ASC+ADRO to Nucleus 5 with ASC+ADRO, re-spectively. However, when using the recommended pro-gram on the Nucleus 5 processor (i.e., Zoom+ASC+ADRO), the subjects scored 1.8 dB. This represented animprovement of 6.8 and 6.0 dB in SRT in noise relative tothe Freedom ASC+ADRO and Nucleus 5 ASC+ADRO,respectively.

FIG. 3. Individual CNC word recognition scores for Freedom(abscissa) and Nucleus 5 (ordinate) processors. Curved lines indi-cated 95% confidence limits for a 200-item test (two 100-item pre-sentations across 2 evaluations) based on binomial comparisons.

FIG. 4. Mean BKB-SIN scores obtained at Visits 2 and 3 for theFreedom with ASC+ADRO and the Nucleus 5 with ASC+ADROand ASC+ADRO+Zoom. Note: vertical lines represent 1 standarddeviation.

557COCHLEAR IMPLANT PROCESSOR WITH BEAMFORMING

Otology & Neurotology, Vol. 33, No. 4, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

Because group analyses indicated no effect of test ses-sion or presentation level, individual data were collapsedacross these factors. Figure 5 shows individual scores,averaged across the 2 test evaluations as well as test level,for the BKB-SIN test for the subjects using the Free-dom and Nucleus 5 sound processors with ASC+ADRO(top panel). The abscissa in each graph indicates the scoreobtained for each subject with Freedom and the ordi-nate the score obtained with Nucleus 5. As in Figure 2,the diagonal indicates comparable performance with the2 processors. The upper and lower diagonal indicate the95% critical difference (1.6 dB) for 8 list-pairs (2 list pairs� 2 presentation levels � 2 test evaluations), per theBKB-SIN test manual. In other words, any scores that fallon or within these 2 lines indicate no significant differ-ences between Freedom and Nucleus 5. Of the 35 subjectstested, 7 subjects showed significantly better performancewith the Nucleus 5 processor, and 28 subjects showedcomparable performance. However, with Zoom activated(Fig. 5, bottom panel), all 35 subjects showed signifi-cantly better performance with the Nucleus 5 processorwhen compared with Freedom using ASC+ADRO, thesubjects’ recommended noise program before conversionto the Nucleus 5 processor.

DISCUSSION

The results of this study with experienced adult recip-ients serve to validate the change in input processingschemes used within the ‘‘Everyday’’ and ‘‘Noise’’ pro-grams for the Nucleus 5 sound processor. Analyses ofspeech recognition performance in quiet and in noiseyielded several noteworthy findings.

A clinically small, but significant improvement, formonosyllabic word recognition in quiet was observedfor subjects using the Nucleus 5: ASC+ADRO whencompared with the Freedom: ADRO alone. In addition,a small but significant improvement was noted in meanSRTs for same device comparison, with both processorsusing ASC+ADRO. These small but statistically signif-icant findings may be related to a series of incrementalchanges in the digital signal-processing pathway of theNucleus 5 sound processor and/or to the improvementsin the microphone technology of the Nucleus 5.

Second, speech recognition in noise across test ses-sions was significantly better by 6 to 6.8 dB when subjectsused the new ‘‘Noise’’ program (Zoom+ASC+ADRO) ofthe Nucleus 5 sound processor relative to performancewith the previous ‘‘Noise’’ program (ASC+ADRO) withboth the Nucleus 5 and Freedom sound processors, re-spectively. Considering the relatively moderate improve-ment in sentence recognition in noise of 0.8 dB that wasobserved with use of the Nucleus 5 sound processor andASC+ADRO versus the Freedom with ASC+ADRO, thesubstantial improvement in performance in noise shouldbe attributed to the directional properties of the Zoomfeature. The improvements with the new ‘‘Noise’’ pro-gram are large given the estimation that a 1 dB improve-ment in SRT may correspond to as much as a 7% to 9%improvement speech recognition in the presence of com-peting noise (22). Furthermore, the mean absolute SRTof 1.8 dB observed with use of the Nucleus 5 ‘‘Noise’’program is quite impressive, given the favorable SNRrequired without Zoom. Given the aforementioned diffi-culties cochlear implant users have experienced in noise,this finding indicates that the sophisticated directionaltechnology should be considered for all adult cochlear im-plant recipients. This finding also demonstrates substantial

FIG. 5. Individual BKB-SIN scores for Freedom versus Nucleus 5 both programmed with ASC+ADRO, top panel. Bottom panel showsindividual scores for Freedom (ASC+ADRO) versus Nucleus 5 with Zoom activated (Zoom+ASC+ADRO). Upper and lower diagonal linesindicate a T1 critical difference based on the presentation of 8 list-pairs (2 list-pairs per condition � 2 presentation levels � 2 test sessions)per individual (see BKB-SIN test manual).

558 J. WOLFE ET AL.

Otology & Neurotology, Vol. 33, No. 4, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

progress in cochlear implant technology over the past2 decades, especially considering the fact that early co-chlear implant systems did not typically allow for open-set speech recognition in quiet.

It should be noted that the improvement observed withZoom in this study may overestimate the benefit ob-tained in the real world. Speech recognition in noise wasmeasured in an audiometric sound booth, and the noisesignal was presented from one loudspeaker located di-rectly to the side of the implanted ear (i.e., relatively closeto the location of the fixed null point in the algorithm).The improvement in SRT achieved with ASC+ADROand the Zoom directional feature relative to the Freedomwith ASC+ADRO is consistent with directional benefitobserved in hearing aid studies, in which directionalbenefit was evaluated in a low-reverberant environmentwith the competing noise signal arising from one location(37Y39). Ricketts (37), however, reported that directionalbenefit in hearing aids decreases by approximately 4 dBwhen measured in a typically reverberant environment(e.g., living room) in the presence of a diffuse competingnoise.

Given the large directional benefit observed in thisstudy, the authors contend that the test condition used inthis study should be considered for clinical use. Perfor-mance with the ‘‘Noise’’ program (Zoom+ASC+ADRO)of the Nucleus 5 relative to the ‘‘Everyday’’ program(ASC+ADRO) of the Nucleus 5 would serve to verify theappropriate function of the directional system. This is animportant step and should be completed longitudinally.As previously noted, precise matching of the omni-directional microphones is required to achieve the desireddirectional response (33). Additionally, accumulation ofdebris over time in the microphones’ protective coversmay differentially influence the input to each microphone,which would negatively influence any attempts at match-ing the responses of the 2 microphones. However, demon-strating a significant improvement in speech recognitionin noise with the Zoom feature verifies the appropriatefunction of the dual-microphone system and allows theclinician to replace microphone covers or replace theprocessor if directional benefit is not observed.

Furthermore, evaluating performance between the Nu-cleus 5 ‘‘Noise’’ and ‘‘Everyday’’ programs can serve asan instructional exercise to educate recipients of the ben-efits associated with and proper use of the Zoom feature.Of course, along with the demonstration of improve-ment in performance in the audiometric booth, it is stillimperative that the clinician explain the rationale of di-rectional systems (i.e., provision of maximal sensitivityfor sounds arriving from the front with attenuation ofsounds from behind, on the basis that the user will facethe signal of interest in a noisy listening situation) aswell as the most appropriate use of a program employingdirectional microphones, such as Zoom. Finally, the cli-nician must keep in mind that the relatively large im-provement in the SRT provided by Zoom (e.g., 7 dB) maystill be insufficient for recipients requiring 1 to 20 dBSNR for adequate understanding in noise. Remote micro-

phone technology, such as FM systems, should be con-sidered for these recipients.

CONCLUSION

Word recognition testing in quiet and sentence recog-nition testing in noise verified that the new recommendeddefault ‘‘Everyday’’ program will deliver, at a minimum,comparable hearing performance to both the ‘‘Everyday’’and ‘‘Noise’’ programs in the Freedom sound processor.

Speech recognition in noise was significantly betterwith the ‘‘Noise’’ program (Zoom+ASC+ADRO) of theNucleus 5 sound processor when compared with speechrecognition in noise obtained with the ‘‘Noise’’ program(ASC+ADRO) of the Freedom processor.

The new ‘‘Everyday’’ and ‘‘Noise’’ programs are appro-priate and beneficial for adult recipients of the Nucleus 5sound processor.

REFERENCES

1. Firszt JB, Holden LK, Skinner MW, et al. Recognition of SpeechPresented at Soft to Loud Levels by Adult Cochlear Implant Reci-pients of Three Cochlear Implant Systems. Ear and Hear 2004;25:375Y87.

2. Gifford R, Dorman M, Shallop J, Sydlowski S. Evidence for theexpansion of adult cochlear implant candidacy. Ear Hear 2010;31:186Y94.

3. Gifford RH, Shallop JK, Peterson AM. Speech Recognition Mate-rials and Ceiling Effects: Considerations for Cochlear ImplantPrograms. Audiol Neurotol 2008;13:193Y205.

4. Uziel A, Sillon M, Vieu A, et al. Ten-year follow-up of a con-secutive series of children with multi-channel cochlear implants.Otol Neurotol 2007;28:615Y28.

5. Zeng FG. Trends in cochlear implants. Trends in Amplification2004;8:1Y34.

6. Spahr A, Dorman MF, Loiselle LH. Performance of patients usingdifferent cochlear implant systems: effects of input dynamic range.Ear Hear 2007;28:260Y75.

7. Schafer EC, Thibodeau LM. Speech recognition abilities of adultsusing cochlear implants with FM systems. J Am Acad Audiol 2004;15:678Y91.

8. Wolfe J, Schafer EC, Heldner B, Mulder H, Vincent B. Evaluationof speech recognition in noise with cochlear implants and dynamicFM. J Am Acad Audiol 2009;20:409Y21.

9. Etymotic Research. BKB-SIN Speech-in-Noise Test. Elk GroveVillage, IL: Etymotic Research, 2005.

10. Plomp R, Mimpen AM. Speech-reception threshold for sentences as afunction of age and noise level. J Acoust Soc Am 1979;66:1333Y42.

11. Wilson RH, McArdle RA, Smith SL. An evaluation of the BKB-SIN, HINT, QuickSIN, and Win materials on listeners with normalhearing and listeners with hearing loss. J Sp Lang Hear Res 2007;50:844Y56.

12. Parkinson AJ, Parkinson W, Tyler RS, et al. Speech perceptionperformance in experienced cochlear-implant patients receiving theSPEAK processing strategy in the Nucleus Spectra-22 cochlear im-plant. J Sp Lang Hear Res 1998;41:1073Y87.

13. Lorenzi C, Gilbert G, Carn H, et al. Speech perception problems ofthe hearing impaired reflect inability to use fine temporal structure.Proc Nat Acad Sci Unit St Amer 2006;103:18866Y9.

14. Fu QJ, Nogaki G. Noise susceptibility of cochlear implant users: therole of spectral resolution and smearing. J Assoc Res Otolaryngol2005;6:19Y27.

15. Nelson PB, Su-Hyun J. Understanding speech in modulated inter-ference: Cochlear implant users and normal-hearing listeners. JAcoust Soc Am 2003;113:961Y8.

559COCHLEAR IMPLANT PROCESSOR WITH BEAMFORMING

Otology & Neurotology, Vol. 33, No. 4, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.

16. Stickney GS, Zeng FG, Litovsky R, Assman P. Cochlear implantspeech recognition with speech maskers. J Acoust Soc Am 2004;116:1081Y91.

17. Zeng FG, Galvin JJ. Amplitude Mapping and Phoneme Recognitionin Cochlear Implant Listeners. Ear Hear 1999;20:60Y74.

18. Zeng FG, Shannon RV, Hellman WS. Physiological processes under-lying psychophysical laws. In: Palmer AR, Rees A, Summerfield AQ,Meddis R, eds. Psychophysical and Physiological Advances in Hear-ing. London: Whurr Publishers Ltd, 1998:473Y81.

19. Buchman CA, Roush PA, Teagle HFB. Auditory neuropathy char-acteristics in children with cochlear nerve deficiency. Ear and Hear2006;27:399Y408.

20. Slattery WH, Luxford WM. Cochlear implantation in the congenitalmalformed cochlear. Laryngoscope 1995;105:1184Y7.

21. Kelly AS, Purdy SC, Thorne PR. Electrophysiological and speechperception measures of auditory processing is experienced adultcochlear implant users. Clin Neurophys 2005;116:1235Y46.

22. Litovsky RY, Parkinson A, Arcaroli J, Sammeth C. Simultaneousbilateral cochlear implantation in adults: a multicenter study. EarHear 2006;27:714Y31.

23. Schafer EC, Thibodeau LM. Speech recognition performance ofchildren using cochlear implants and FM systems. J Educ Audiol2003;11:15Y26.

24. Wolfe J, Schafer ES. Optimizing the benefit of sound processorcoupled to personal FM systems. J Am Acad Audiol 2008;19:585Y94.

25. Spahr A, Dorman M. Performance of subjects fit with the AdvancedBionics CII and Nucleus ESPrit 3G cochlear implant devices. ArchOtolaryngol Head Neck Surg 2004;130:624Y8.

26. Wolfe J, Schafer ES, John A, et al. The effect of front-end pro-cessing on cochlear implant performance of children. Otol Neurotol2011;32:533Y8.

27. Dawson PW, Decker JA, Psarros CE. Optimizing dynamic range inchildren using the Nucleus cochlear implant. Ear Hear 2004;25:230Y41.

28. James CJ, Blamey PJ, Martin L, Swanson B, Just Y, Macfarlane D.

Adaptive dynamic range optimization for cochlear implants: a pre-liminary study. Ear Hear 2002;23:49SY58S.

29. Amlani A. Efficacy of directional microphone hearing aids: a meta-analytic perspective. J Am Acad Audiol 2001;12:202Y14.

30. Bentler RA. The effectiveness of microphones and noise reductionschemes in digital hearing aids: a systematic review of the evidence.J Am Acad of Audiol 2005;16:473Y84.

31. Spriet A, Van Deun L, Eftaxiadis K, et al. Speech understanding inbackground noise with the two-microphone adaptive beamformerBEAM in the Nucleus Freedom cochlear implant system. Ear Hear2007;28:62Y72.

32. Gifford RH, Revit LJ. Speech perception for cochlear implantrecipients in a realistic background noise: effectiveness of pre-processing strategies and external options for improving sentencerecognition in noise. J Am Acad Audiol 2010;21:441Y51.

33. Edwards B, Hou Z, Struck C, Dharan P. Signal-processing algo-rithms for a new software-based, digital hearing device. Hear J1998;51:44Y52.

34. Cochlear Technology Spotlight: Informational Series for Profes-sionals. Dual omni-directional microphone technology. CochlearAmericas 2010.

35. Skinner MW, Holden LK, Fourakis MS, et al. Evaluation of equi-valency in two recordings of monosyllabic words. J Am Acad Audiol2006;17:350Y66.

36. Thornton AR, Raffin MJM. Speech-discrimination scores modeledas a binomial variable. J Sp Lang Hear Res 1978;21:507Y18.

37. Ricketts TA. Impact of noise source configuration in directionalhearing aid benefit and performance. Ear Hear 2000;21:194Y205.

38. Ricketts TA, Dahr S. Aided benefit across directional and omni-directional hearing aid microphones for behind-the-ear hearing aids.J Am Acad Audiol 1999;10:180Y9.

39. Valente M, Fabry DA, Potts LG. Recognition of speech in noisewith hearing aids using dual microphones. J Am Acad Audiol 1995;6:440Y9.

560 J. WOLFE ET AL.

Otology & Neurotology, Vol. 33, No. 4, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.