Hearing Research xxx (2014) 1e8

Contents lists available at ScienceDirect

Hearing Research journal homepage: www.elsevier.com/locate/heares

Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users Reinhold Schatzer a, b, Inna Koroleva c, *, Andreas Griessner a, Sergey Levin c, Vladislav Kusovkov c, Yuri Yanov c, Clemens Zierhofer a a b c

Institute of Mechatronics, University of Innsbruck, Technikerstrabe 13, 6020 Innsbruck, Austria MED-EL GmbH, Fürstenweg 77a, 6020 Innsbruck, Austria St. Petersburg ENT and Speech Research Institute, Bronnitskaja 9, 198013 St. Petersburg, Russia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 May 2014 Received in revised form 24 October 2014 Accepted 5 November 2014 Available online xxx

Early multi-channel designs in the history of cochlear implant development were based on a vocodertype processing of frequency channels and presented bands of compressed analog stimulus waveforms simultaneously on multiple tonotopically arranged electrodes. The realization that the direct summation of electrical fields as a result of simultaneous electrode stimulation exacerbates interactions among the stimulation channels and limits cochlear implant outcome led to the breakthrough in the development of cochlear implants, the continuous interleaved (CIS) sampling coding strategy. By interleaving stimulation pulses across electrodes, CIS activates only a single electrode at each point in time, preventing a direct summation of electrical fields and hence the primary component of channel interactions. In this paper we show that a previously presented approach of simultaneous stimulation with channel interaction compensation (CIC) may also ameliorate the deleterious effects of simultaneous channel interaction on speech perception. In an acute study conducted in eleven experienced MED-EL implant users, configurations involving simultaneous stimulation with CIC and doubled pulse phase durations have been investigated. As pairs of electrodes were activated simultaneously and pulse durations were doubled, carrier rates remained the same. Comparison conditions involved both CIS and fine structure (FS) strategies, either with strictly sequential or paired-simultaneous stimulation. Results showed no statistical difference in the perception of sentences in noise and monosyllables for sequential and paired-simultaneous stimulation with doubled phase durations. This suggests that CIC can largely compensate for the effects of simultaneous channel interaction, for both CIS and FS coding strategies. A simultaneous stimulation paradigm has a number of potential advantages over a traditional sequential interleaved design. The flexibility gained when dropping the requirement of interleaving pulses across electrodes may be instrumental in designing coding strategies for a more accurate transmission of stimulus features such as temporal fine structure or interaural time delays to the auditory nerve. Also, longer pulse phase durations may be implemented while maintaining relatively high stimulation pulse rates. Utilizing longer pulse durations may relax requirements on implant compliance and facilitate the design of more energy-efficient implant receivers for a longer battery lifetime or a reduction in implant size. This article is part of a Special Issue entitled . © 2014 Elsevier B.V. All rights reserved.

Abbreviations: BYU, Brigham Young University; CI, cochlear implant; CIC, channel interaction compensation; CIS, continuous interleaved sampling; CSSS, channel-specific sampling sequences; ENT, ear nose throat; FS, fine structure; FSP, fine structure processing; MCL, most comfortable level; pps, pulses per second; RM ANOVA, repeated measures analysis of variance; SNR, signal-to-noise ratio; UCSF, University of California San Francisco; USB, universal serial bus * Corresponding author. Tel.: þ7 812 316 2256. E-mail address: [email protected] (I. Koroleva).

1. Introduction Cochlear implants (CIs) represent an established treatment for the restoration of functional hearing in adults and children with severe to profound sensorineural hearing loss via direct electrical stimulation of the auditory nerve. With an exponentially growing

http://dx.doi.org/10.1016/j.heares.2014.11.002 0378-5955/© 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002

2

R. Schatzer et al. / Hearing Research xxx (2014) 1e8

number of implant recipients reaching an estimated 324,200 people worldwide as of December 2012 (NIDCD, 2014), CIs are among the most successful neural prostheses. Following early observations on electrical hearing in one patient s in 1957 in Paris (Djourno and Eyries, 1957), CI by Djourno and Eyrie research started on a larger scale in the early 1960s in the United States. In 1961, William House and John Doyle implanted a rudimentary single-electrode device in two patients in Los Angeles. House and his engineer Jack Urban later developed a complete single-channel implant system (House, 1976), implanted from the early 1970s on and later commercialized by 3M to become the very first CI device to receive clinical approval by the U.S. Food and Drug Administration in 1984. While some basic levels of speech recognition were possible with this and later single-channel analog devices developed by other groups, patients enjoying open speech perception without lip reading remained the exception. Limiting factors are a coarse spectral signal representation via a single broad-band stimulation channel in conjunction with limitations in the perception of temporal cues for stimulus frequencies exceeding approximately 300 Hz in most CI recipients (Simmons et al., 1965; Zeng, 2002). From early on, CI researchers believed that implant systems would require multiple channels to exploit the tonotopic organization of the auditory nerve for representing both spectral and temporal information in a meaningful way (Clark, 2013; Hochmair, 2013; Wilson, 2013). In 1961, House and Doyle re-implanted their first single-channel patient with a five-wire scala-tympani electrode system, but the device was soon rejected. In 1969, House and Urban implanted three more subjects with a percutaneous-plug version of the five-channel device, demonstrating place pitch discrimination in one patient who became available as their longterm research subject (House and Urban, 1973). Apparently because channel interaction effects could not be controlled, that patient developed a preference for a stimulation scheme that presented the same broad-band information to each electrode. This prompted the further development of the single-channel instead of the multi-channel system. Research with multi-channel devices continued in 1964 with Blair Simmons at Stanford University, who implanted a cluster of six electrodes in the modiolus portion of the auditory nerve in one subject (Simmons et al., 1965). The team around Robin Michelson at the University of California, San Francisco (UCSF) implanted a single-channel device in 1971 and later developed a four-channel analog implant system. While the Stanford experience did not spin off into a commercial device, the UCSF electrode design was taken up and developed into a commercial CI system by Advanced Bionics Corp in 1993. Donald Eddington and his team at the University of Utah developed a percutaneous (through-the-skin) analog, six-channel device that was later marketed as Ineraid device by Symbion, Inc. (Eddington et al., 1978). In the second half of the 1970s, research groups in Australia and Europe started to develop transcutaneous multi-channel implant systems. Anticipating modern CI designs, those systems comprised an external processor (transmitter) and an implanted receiver that was connected to an array of intracochlear electrodes inserted into the scala tympani. An inductive radio-frequency link supplied power and stimulation data to the implanted receiver. In September of 1976, Claude Chouard and his team at the Saint-Antoine Hospital in Paris implanted the first multi-channel transcutaneous device with an implant receiver built out of discrete electronic components embedded in a Silastic cylinder (Chouard, 1978). In December of 1977, Ingeborg Hochmair-Desoyer, Erwin Hochmair and Kurt Burian at the Technical and Medical Universities in Vienna implanted the first transcutaneous microelectronic multi-channel device built on state-of-the-art hybrid integrated-circuit technology, an eight-

channel device with eight independent current sources (Desoyer and Hochmair, 1977; Hochmair, 2013). In August of 1978, the Australian group around Graeme Clark at the University of Melbourne implanted their first patient with a transcutaneous microelectronic 10-channel device (Clark et al., 1977). The Melbourne device was later developed further to include 22 stimulation channels. The CI systems developed in Melbourne, Vienna, and Paris later received clinical approval and were commercialized by what became Cochlear Ltd., MED-EL, and MXM-Neurelec, together with Advanced Bionics-Phonak the leading CI companies today. Another transcutaneous multi-channel implant system developed by a research team at the University of Antwerp (Peeters et al., 1989) was marketed for a short period of time as Laura device by Philips Hearing Instruments. In a traditional CI system, an external speech processor picks up sound signals through a microphone and converts them into electrical stimulation parameters according to an algorithm designated as speech or sound coding strategy (Wilson, 2013). In the early days of CIs, multi-channel coding strategies were either presenting speech features extracted from the acoustic signal via pulsatile stimulation that was interleaved across electrodes (Clark et al., 1977; Seligman et al., 1984), or compressed channel output waveforms via simultaneous analog stimulation on multiple electrodes (Eddington, 1980; Merzenich et al., 1984). While feature extraction schemes were limited by an underrepresentation of complex stimulus features, the inherent problem of channel interaction due to electric field summation was e in retrospect e the limiting factor of coding strategies based on a simultaneous stimulation paradigm. The breakthrough in the development of CI systems was the introduction of the Continuous Interleaved Sampling (CIS) strategy by Wilson et al. (1991). As a defining feature of CIS, which can still be considered the standard of CI coding strategies today, stimulation pulses are interleaved across electrodes, but in contrast to feature extraction strategies CIS electrode stimuli sample and present compressed channel envelopes at a relatively high and constant pulse rate. The continuous interleaved sampling of channel outputs prevents a direct electrical summation of currents from multiple electrodes, which is the primary source of detrimental channel interactions. In addition to CIS, researchers and implant manufacturers have developed a variety of other sophisticated coding strategies. These include n-of-m-type strategies dynamically picking spectral peaks for stimulation (Kiefer et al., 2001; McDermott et al., 1992; Skinner et al., 1994; Wilson et al., 1988) or concepts representing spectral fine structure via current steering or virtual channels (Koch et al., 2004; Townshend et al., 1987; Wilson et al., 1994). MED-EL implant systems support different variations of a temporal fine structure (FS) coding strategy, denoted as FSP and FS4 (Lorens et al., 2010; Riss et al., 2009; Schatzer et al., 2010; Zierhofer, 2003a). These FS strategies present both temporal FS and envelope information on multiple apical electrodes, while CIS envelope-based stimulation only is applied on middle and basal electrodes. For instance, the recently released FS4 coding strategy has up to 4 apical FS channels, representing frequencies of up to approximately 800 Hz with both temporal FS and envelope stimulation. While sequential interleaved coding strategies have proven very successful in the last decades, a simultaneous stimulation paradigm has several potential advantages. First, activating multiple electrodes in parallel allows increasing the overall stimulation rate. This facilitates the design of coding strategies such as fine structure concepts by providing a higher amount of flexibility in the spectrotemporal distribution of individual stimulation pulses, for instance to represent fine frequency information or encode interaural time delays in bilateral implant systems. Alternatively, the implementation of a simultaneous stimulation concept makes it

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002

R. Schatzer et al. / Hearing Research xxx (2014) 1e8

possible to increase pulse durations while keeping stimulation rates relatively high. This represents a further step towards reaching a present engineering goal in the development of CIs, which is to minimize system size and reduce power consumption in order to improve wearability and battery life or size, or both. Because loudness is roughly proportional to stimulus pulse charge, i.e. the product of current amplitude and duration, longer pulse durations relax the need for high pulse amplitudes to achieve sufficient stimulus loudness. Lower current amplitudes put less stringent requirements on implant compliance voltages, which in turn allows for a more energy-efficient design of the implant electronics. The possibility of implementing longer phase duration pulses at relatively high pulse rates with simultaneous stimulation may also be useful for patients with higher charge requirements. Simultaneous stimulation paradigms have been previously investigated in CI users, for both analog (Zwolan et al., 2005) and paired simultaneous pulsatile configurations (Bonnet et al., 2012; Buechner et al., 2005). However, speech test performance, in particular for speech presented in competition with noise, was significantly lower with simultaneous stimulation as compared to interleaved sequential stimulation. A simultaneous activation of electrodes in a monopolar configuration results in a direct electrical field summation and exacerbates channel interactions. It is however conceivable to extend the previous simultaneous pulsatile approaches by adding a mechanism to compensate, at least in part, for the electrical field summation through a controlled reduction of simultaneous current pulse amplitudes. Presumably, this also ameliorates the inherent problem of increased channel interactions with simultaneous stimulation. The mechanism, designated as channel interaction compensation (CIC), has been previously described and investigated in CI users (Bader et al., 2013; Zierhofer and Schatzer, 2008). A goal of the present simultaneous stimulation concept with CIC and doubled phase durations is to reduce the stimulation pulse amplitudes, thereby allowing for a reduction of the implant compliance voltage and consequently of the implant power consumption. In contrast to virtual channel or current steering concepts such as Advanced Bionics’ HiRes Fidelity 120 strategy (Firszt et al., 2007), which aim at providing a more accurate spectral signal representation through a CI, the CIC algorithm aims at preserving the spectral resolution of a sequential CIS stimulation paradigm while relaxing implant power requirements. The recently introduced HiRes Optima strategy on the other hand is designed to improve battery lifetime without compromising implant performance (Ajimsha and Mathias, 2014). Specifically, the present study investigates the efficacy of the CIC mechanism in combination with a doubling of the pulse phase duration for configurations where pairs of simultaneous biphasic pulses are applied on adjacent electrodes in monopolar mode. While Bader et al. (2013) already compared a similar CIC configuration to an experimental CIS condition, the present study extends the prior findings by investigating both FS and CIS strategies with interleaved and simultaneous stimulation, respectively, and by comparing those settings with the participants' everyday OPUS2 speech processor map. The hypothesis was that for both types of coding strategy speech test performance with paired simultaneous stimulation and CIC would be on a par with sequential stimulation. Effects of simultaneous channel interaction on loudness and speech perception may depend on channel pulse rate (Bonnet et al., 2012; Middlebrooks, 2004; Schatzer et al., 2013). While CIS as investigated here and in Bader et al. uses relatively high carrier pulse rates above the pitch saturation limit on all channels, FS coding presents low-rate temporal information on select apical channels. Hence, simultaneous stimulation configurations with CIC and doubled pulse durations were additionally investigated with FS coding

3

strategies in the present study. Similar levels of test performance were also expected for the experimental settings and the everyday processor map which served as a control. A prospective study design with an acute comparison of processor conditions was adopted. 2. Methods 2.1. Subjects Eleven experienced adult MED-EL implant users participated in this study. Subject demographics are shown in Table 1. The mean age of the participants was 43.8 years (range 26.0e68.1 years), the average implant experience was 1.6 years (range 3 monthse3.8 years). All subjects except S08 are postlingually deaf. Subject S08 is perilingually deaf and her speech production is moderately impaired. All participants are implanted with a 31-mm MED-EL Standard array with a contact spacing of 2.4 mm. Two recipients use a MED-EL SONATA device, all others are PULSAR (Zierhofer, 2003b) users. S01 had the two basal-most electrodes 11 and 12 switched off clinically and during speech testing due to unpleasant sound sensations, but impedances were within normal ranges on those two electrodes so that a complete insertion of the array cannot be excluded. All remaining subjects had fully functional basal electrodes, indicating full insertions of their arrays. The study was conducted at the Research Institute of ENT & Speech (ENT Institute) in St. Petersburg, Russia. The study was approved by the Ethics Review Board of the ENT Institute. 2.2. Channel interaction compensation CIC as applied for all paired-simultaneous configurations in the present study is described in detail in Zierhofer and Schatzer (2008). The basis for the algorithm is a one-dimensional model of the cochlea (Kral et al., 1998) and a model for the electric field potential which is centered on the stimulating electrode and exponentially decreasing towards the apex and base with decay constants a and b, respectively. Constants a and b are exponentials of a decay parameterd l (mm), normalized to the electrode spacing  d  d (mm), i.e. a ¼ e lapex and b ¼ e lbase . This model can be used to calculate the approximate longitudinal intracochlear field potential in a sequential stimulation paradigm. In that paradigm, masking effects will occur if pulses are applied within an absolute refractory period and on different electrodes that excite spatially overlapping populations. The effective resulting potential is the contour of the corresponding electrode potentials (c.f. Fig. 3 in Zierhofer and Schatzer, 2008). Directly applying the sequential pulse sequence in parallel and without compensation would lead to a summation of the overlapping potentials and thus to a significant increase in loudness and a consequent decrease in hearing performance. Consonant and vowel test scores in four subjects studied by Zierhofer and Schatzer were indeed significantly lower with uncompensated paired simultaneous stimulation as compared to CIS, with the only exception of consonant test scores in subject S4. The aim of CIC is to stimulate two or more electrodes simultaneously in a way that the resulting summation potential closely approximates the contour potential resulting from sequential stimulation. CIC is an efficient mathematical algorithm that calculates reduced simultaneous current pulse amplitudes from a set of known sequential amplitudes in a way that the resulting summation potential exactly matches the sequential contour potential at the sites of stimulation. CIC is based on the assumptions of a monopolar electrode configuration and simultaneous pulses of equal polarity. It is important to note that because of not providing for multipolar configurations and alternating pulse polarities across simultaneously activated electrodes, CIC is not producing a spectral

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002

4

R. Schatzer et al. / Hearing Research xxx (2014) 1e8

Table 1 Subject demographics. All participants are MED-EL recipients with an implant experience ranging from 3 months to 3.8 years (average 1.6 years). The table includes each subject's preferred OPUS2 setting which was also evaluated during speech testing. In patients using an FSP fine structure map, the number of apical FS channels is indicated in parentheses. Subject

Gender

Age (yrs)

Etiology

Duration deafness (yrs)

Implant use (yrs)

Implant type

Electrode type

OPUS2 setting and frequency range (Hz)

S01 S02 S03 S04 S05 S06 S07 S08 S09 S10 S11 Mean

m m m m f m m f m f m

41.6 67.1 36.2 38.0 55.3 68.1 55.3 31.2 26.6 26.0 36.2 43.8

Progressive Unknown Progressive Trauma Ototoxic drugs Progressive Trauma Progressive Progressive Ototoxic drugs Progressive

10 2 10 10 25 4 10 11 1 5 4 8.4

0.6 0.8 0.4 2.9 1.2 3.8 3.8 0.5 0.2 2.0 1.1 1.6

PULSAR SONATA SONATA PULSAR PULSAR PULSAR PULSAR PULSAR PULSAR PULSAR PULSAR

Standard Standard Standard Standard Standard Standard Standard Standard Standard Standard Standard

FSP(1), 160-8k5 HDCIS, 200-7k FSP(2), 150-8k5 HDCIS, 250-6k5 FSP(3), 100-8k5 FSP(1), 200-6k5 HDCIS, 150-7k5 FSP(2), 130-8k5 HDCIS, 200-7k FSP(1), 150-6k5 FSP(1), 150-7k5

sharpening of intracochlear field potentials as other approaches based on similar field summation models (Townshend and White, 1987; van den Honert and Kelsall, 2007). Instead, contour potentials resulting from CIC typically have a shallower profile in between electrode positions than the original sequential contours. This deviation of the CIC from the sequential profile is smallest for adjacent and largest for maximally separated simultaneous electrodes. Hence, as illustrated in Fig. 1, the tested simultaneous configurations had pairs of adjacent electrodes activated in parallel. 2.3. Stimulation strategy settings The tested stimulation strategy configurations are summarized in Table 2. The six configurations included five experimental settings, in addition to each participant's preferred clinical OPUS2 processor map used on a daily basis. The five experimental settings included a sequential and paired-simultaneous CIS configuration, as well as one sequential and two paired-simultaneous FS configurations with four temporal FS channels each. Coding strategy type

and parameters for the clinical maps varied across subjects, as indicated in Table 2. For each subject, the experimental test settings utilized the same number of channels (channels switched off in the clinical map were also deactivated) and the same frequency range from 100 to 8500 Hz. FS settings CSSS, CSSSp2, and FS4p2 represented temporal fine structure information on the four apical-most channels via “channel specific sampling sequences” (CSSS) (Zierhofer, 2003a). CSSS channels 1 and 2 represented within-channel fine structure by generating a brief double-pulse sequence each time a positive zero crossing was detected on the corresponding band filter output. CSSS channels 3 and 4 presented single-pulse sequences at each positive zero crossing. All experimental settings were fitted separately on an OPUS1 processor by first adjusting maximum comfortable levels (MCLs) individually for each electrode and setting, and then balancing the obtained stimulation levels at MCL in loudness. Finally, the processor volume was adjusted separately for each setting in order for the speech stimuli to be comfortably loud. This procedure was

Fig. 1. Illustration of sequential versus paired-simultaneous stimulation configurations and respective phase durations for the CIS test setting.

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002

R. Schatzer et al. / Hearing Research xxx (2014) 1e8 Table 2 Stimulation strategy settings. Name

Strategy

FS channels

FS channel picking

Frequency range (Hz)

Stimulation

Clinical

0e3

e

e e

e e

Patient specific 100e8k5 100e8k5

Sequential

CIS CISp2

HDCIS or FSP CIS CIS

CSSS CSSSp2

FS FS

[1 2][3 4] [1 2]p[3 4]p

1 of 2 2 simulta-neous

100e8k5 100e8k5

FS4p2

FS

[1 2 3 4]

1 of 4

100e8k5

Sequential Paired simultaneous, doubled phase duration Sequential Paired simultaneous, doubled phase duration Paired simultaneous, doubled phase duration

performed to compensate for the charge and consequent loudness increase for stimulation pulses presented at the same level but with longer phase durations. For settings CIS, CSSS, and with each subject's clinical OPUS2 map, electrodes were stimulated sequentially. CSSS arranged the four FS channels in two “selected groups” [1 2] and [3 4], where in each of the two groups the channel with the largest amplitude was dynamically picked for stimulation (Kals et al., 2010). With CISp2 and CSSSp2, groups of two adjacent channels each were stimulated simultaneously, as illustrated in Fig. 1. With FS4p2, only CIS channels 5e12 were stimulated simultaneously in groups of two neighboring channels, while the four FS channels were configured as one selected group [1 2 3 4] with the highestamplitude channel among the four being picked for stimulation. CIC parameters a and b were set to a ¼ 0.76 and b ¼ 0.67 for all paired-simultaneous conditions, i.e. a shallower decay in apical direction was assumed. The specific values correspond to the means of the CIC parameters obtained from prior estimates based on objective and psychophysical measures in CI users. For the objective-measure parameter estimate, the decay constants were calculated from exponential functions fitted to current-field telemetry measurements from an anonymized patient database. For the psychophysical parameter estimate, subjects were asked to match both loudness and timbre or pitch between running processors alternating sequential and simultaneous stimulation with steady-state wide-band noise input signals (Bader et al., 2013). Parameters obtained from the two methods were similar but showed a high inter-subject variability. However, speech test performance is relatively insensitive to CIC parameter changes, as long as the decay constants are not too small (Zierhofer and Schatzer, 2008). As results from that study have shown that simultaneous stimulation without CIC significantly deteriorates speech perception in CI users, our set of test conditions did not include a simultaneous setting with decay constants set to zero, i.e. CIC turned off. Across test settings, electrodes were stimulated in monopolar configuration using biphasic pulses. For the sequential stimulation strategies CIS and CSSS, pulse phase durations were 30 ms, except for subjects S04 (apical four electrodes) and S07 (all electrodes), where MCL charge levels required a phase duration of 53 ms. For all paired-simultaneous stimulation strategies, pulse phase durations were twice as long, i.e. 60 ms or 106 ms, respectively. 2.4. Stimuli and test procedure Speech test materials included both the Russian-language version of the Oldenburg matrix sentence test (Zokoll et al., 2013)

5

and the Russian monosyllable test developed by Brigham Young University (BYU). The Russian matrix test comprises 8 lists of 20 sentences each. Each sentence has the same structure “Name Verb Numeral Adjective Object”, drawing from a closed set of words for each category, and is a nonsense sentence. The speaker is female, and the masking noise has the same long-term spectrum as the test material. The BYU test is an open-set monosyllabic word test and comprises 4 lists of 50 words each. For this study, the male-speaker version of this test was utilized. Speech test stimuli were fed to an OPUS1 (for all experimental test settings) or OPUS2 (for each subject's own processor map setting) processor via direct input cable connected to the output of an M-Audio Universal Serial Bus (USB) sound card installed on a Dell laptop computer. The presentation level for the speech stimuli was set below clipping level and adjusted to individual comfortable loudness via processor volume control. A Baaske Medical USB isolator cable was utilized to connect the sound device to the mains-powered laptop computer. The OPUS1 and OPUS2 processors were powered via a MED-EL Diagnostic Interface Box 2. Because CI listeners sometimes may not achieve an intelligibility score above 50 percent correct at high signal-to-noise ratios (SNRs) required for measuring an adaptive speech reception threshold, the sentence test lists were presented at an individually-adjusted constant SNR across conditions. The SNR was determined such as to avoid ceiling or flooring effects by presenting several practice lists to each subject. The first list was presented in quiet, and for the subsequent lists the SNR was adjusted in 5-dB or larger steps, if necessary, until performance in percent correct fell below approximately 85% correct. The so determined individual SNRs were 10 dB for S02, 5 dB for S01, S05, and S06, and 0 dB for all remaining subjects. One list of the Russian matrix test was presented for each of the six test conditions, except for subject S01, where two lists per condition were presented in block-randomized order. Results were scored as percent of words correctly identified. The BYU monosyllable test was presented in quiet, except for subjects S04, S10, and S11, who were tested at 10, 15, and 15 dB SNR in speech-spectrum-shaped CCITT noise, respectively, to avoid ceiling performance levels. Due to the limited testing time available with subject S08, the BYU monosyllable test could not be performed in this one instance. For practice and familiarization purposes to the male speaker's voice, prior to the actual monosyllable testing each subject was presented with a bisyllabic word list of the same speaker. The testing order of conditions and monosyllable list used for each condition were randomized across subjects. As the BYU monosyllable test comprises only 4 lists of 50 items each, lists were re-used. In order to counterbalance potential learning effects, re-used lists varied depending on the randomization, and list items were presented in random order. Test results were scored as percent of words correctly identified. The study was conducted in an acute setting, with all experimental processor maps fitted only minutes before being evaluated in speech testing. Thus, subjects had no listening experience with those settings prior to testing. 3. Results Results for the Russian matrix sentence test are shown in Fig. 2. Mean test performance ranged from 76.2% for the clinical OPUS2 condition to 81.4% correct for the CSSSp2 FS condition with pairedsimultaneous stimulation. A one-way repeated measures (RM) ANOVA shows no statistically significant differences among the six test conditions (p ¼ 0.236, F(5,50) ¼ 1.411). A “rationalized” arcsine transformation (Studebaker, 1985) was applied to the original data in order to meet the parametric test requirements of normal distributions and equal variances.

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002

6

R. Schatzer et al. / Hearing Research xxx (2014) 1e8

Fig. 2. Group results for the recognition of Russian matrix sentences in eleven subjects. Sentence lists were presented at a constant SNR for each subject, with the actual SNR individually determined prior to testing to avoid ceiling effects. Differences among the six test conditions are not statistically significant (p ¼ 0.236).

Differences between sequential and paired-simultaneous stimulation modalities for strategies CIS and CSSS were further analyzed with a two-way RM ANOVA. There is no significant interaction between the two factors of strategy and stimulation modality (p ¼ 0.349, F(1,10) ¼ 0.964), nor a significant main effect of stimulation mode (p ¼ 0.672, F(1,10) ¼ 0.190) or coding strategy (p ¼ 0.067, F(1,10) ¼ 4.227). Results of the BYU monosyllable test are shown in Fig. 3. Ten subjects have completed the test, with the mean test performance ranging from 45.4% for CISp2 to 51.2% correct for the clinical OPUS2 condition. Again, no statistically significant differences between the tested settings are found in an RM ANOVA (p ¼ 0.547, F(5,45) ¼ 0.813). A two-way RM ANOVA comparing CIS and CSSS for sequential and paired-simultaneous stimulation modalities reveals no significant interaction between the two factors of strategy and stimulation modality (p ¼ 0.628, F(1,9) ¼ 0.252), nor significant main effects of stimulation mode (p ¼ 0.752, F(1,9) ¼ 0.106) or coding strategy (p ¼ 0.658, F(1,9) ¼ 0.210). In order to assess the potential advantages on power consumption with the simultaneous coding strategies and doubled phase durations, the reduction of the stimulation amplitudes has been measured for both the tested CIS and FS settings. The mean reduction of the MCL pulse amplitudes weighted with the respective volume settings for equal loudness was 4.78 ± 0.24 dB for stimulation with CISp2 as compared to CIS, and 4.79 ± 0.23 dB for CSSSp2 as compared to CSSS. 4. Discussion Speech test results for paired-simultaneous stimulation with CIC and doubled phase duration were on a par to results with sequential stimulation, both for sentences presented in noise and monosyllables presented in quiet in most subjects. This was the case for both types of coding strategies investigated, i.e. for an envelope-based CIS as well as for a temporal FS strategy, and with one default set of CIC parameters utilized across subjects and coding strategies respectively channel pulse rate settings. Implant listeners did not volunteer that subjective sound quality or pitch was different between sequential and simultaneous stimulation paradigms within a particular type of coding strategy. This suggests

Fig. 3. Group results for the recognition of Russian BYU monosyllables in ten subjects. Differences among the six test conditions are not statistically significant (p ¼ 0.547).

that CIC with a default set of interaction parameters does not produce excitation profiles that are systematically skewed compared to sequential stimulation profiles, as this would likely result in consistent differences in pitch quality. Finally, speech test performance with all experimental (and unfamiliar) conditions was not statistically different from results obtained with the subjects’ everyday OPUS2 processor map. Simultaneous stimulation paradigms have been previously investigated clinically in CI users, both for analog (Zwolan et al., 2005) and paired simultaneous pulsatile configurations (Bonnet et al., 2012; Buechner et al., 2005). In a semi-acute comparison of CIS and paired simultaneous coding strategies in 27 Advanced Bionics implant users, Bonnet et al. consistently found poorer monosyllabic word scores with paired simultaneous stimulation. When channel pulse rates were kept constant, word recognition was significantly lower with paired simultaneous stimulation in quiet and in a þ10 dB SNR. In the study by Buechner et al., speech perception for both HochmaireSchulzeMoser sentences in noise at 10 dB SNR and Freiburger monosyllabic words in quiet was significantly lower with paired simultaneous stimulation as compared to interleaved pulsatile CIS stimulation, after one month of listening experience with each of the test conditions. Sound quality as assessed by questionnaire was also significantly lower for the paired-simultaneous strategy compared to the CIS strategy. Bonnet et al. and Buechner et al. applied simultaneous stimulation on maximally separated electrodes and reduced electrode MCLs as required for equal loudness to the CIS maps. In contrast, in our study pairs of adjacent electrodes were activated simultaneously, and pulse amplitudes were dynamically reduced according to the CIC algorithm previously described. In all studies electrical stimuli were presented in monopolar electrode configuration. Zwolan et al. compared speech recognition and patient preference with simultaneous analog stimulation (SAS) and CIS in 25 postlingual adult users of the Clarion 1.2 device, using a longitudinal ABAB crossover design with a cumulative listening experience of six weeks per strategy. SAS used a bipolar electrode configuration in order to minimize simultaneous channel interactions, whereas CIS used monopolar stimulation. At the replication interval, scores for Hearing in Noise Test sentences were significantly higher for the CIS strategy than for the SAS strategy, in both quiet and noise. A statistically significant majority of participants also

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002

R. Schatzer et al. / Hearing Research xxx (2014) 1e8

indicated their final subjective preference for the CIS than for the SAS strategy. In contrast to all three clinical studies in Advanced Bionics recipients mentioned above, we found equal levels of speech test performance between simultaneous and sequential stimulation strategies. This suggests that CIC, at least for only two adjacent electrodes stimulated simultaneously as in the present study, can effectively compensate for the deleterious effects of simultaneous channel interaction. Conversely, neither a maximum separation of paired simultaneous electrodes accompanied by an adjustment of electrode MCLs as in Buechner et al. and Bonnet et al., nor a bipolar configuration of simultaneously stimulated electrodes as in Zwolan et al. could sufficiently limit the negative effects of simultaneous electrode interactions on speech perception. A comparison of the present concept based on paired simultaneous stimulation with extended phase durations and CIC to more recent coding strategies such as HiRes Fidelity 120 from Advanced Bionics (Firszt et al., 2007) seems limited. HiRes 120 applies simultaneous pulsatile stimulation on adjacent electrodes to “steer” the centroid of neural excitation between those sites of stimulation, in an attempt to present more information by increasing spectral resolution. In contrast, the concept presented here aims at reproducing the spectral and temporal resolution achieved with interleaved sequential stimulation while allowing for an increase of pulse phase durations and improving power efficiency. Although the present results seem promising, the small sample size and consequently limited statistical power of the analysis demand some caution in generalizing the findings. Nevertheless, the flexibility gained when dropping the requirement of interleaving pulses across electrodes may be instrumental in designing coding strategies for a more accurate transmission of stimulus features such as temporal fine structure or interaural time delays to the auditory nerve. Acknowledgments We thank our subjects for their time and efforts. This research has been supported by MED-EL. References Ajimsha, K.M., Mathias, N., 2014. Evaluation of the battery lifetime improvement with the HiRes Optima™ strategy in Harmony™ cochlear implant users. In: 13th International Conference on Cochlear Implants and Other Implantable Auditory Technologies, Munich, Germany. Bader, P., Kals, M., Schatzer, R., Griessner, A., Zierhofer, C., 2013. Compensation for channel interaction in a simultaneous cochlear implant coding strategy. J. Acoust. Soc. Am. 133, 4124e4132. Bonnet, R.M., Boermans, P.P., Avenarius, O.F., Briaire, J.J., Frijns, J.H., 2012. Effects of pulse width, pulse rate and paired electrode stimulation on psychophysical measures of dynamic range and speech recognition in cochlear implants. Ear. Hear. 33, 489e496. Buechner, A., Frohne-Buechner, C., Stoever, T., Gaertner, L., Battmer, R.D., Lenarz, T., 2005. Comparison of a paired or sequential stimulation paradigm with advanced bionics' high-resolution mode. Otol. Neurotol. 26, 941e947. Chouard, C.H., 1978. Multiple intracochlear electrodes for rehabilitation in total deafness. Otolaryngol. Clin. North Am. 11, 217e233. Clark, G.M., 2013. The multichannel cochlear implant for severe-to-profound hearing loss. Nat. Med 19, 1236e1239. Clark, G.M., Tong, Y.C., Black, R., Forster, I.C., Patrick, J.F., Dewhurst, D.J., 1977. A multiple electrode cochlear implant. J. Laryngol. Otol. 91, 935e945. Desoyer, I., Hochmair, E., 1977. Implantable eight-channel stimulator for the deaf. Proc. Eur. Solid State Circuits Conference 87e88. Djourno, A., Eyries, C., 1957. Prothese auditive par excitation electrique a distance du nerf sensoriel a l'aide d'un bodinage inclus a demeure. La Presse medicale 65, 1417. Eddington, D.K., 1980. Speech discrimination in deaf subjects with cochlear implants. J. Acoust. Soc. Am. 68, 885e891. Eddington, D.K., Dobelle, W.H., Brackmann, D.E., Mladejovsky, M.G., Parkin, J.L., 1978. Auditory prostheses research with multiple channel intracochlear stimulation in man. Ann. Otol. Rhinol. Laryngol 87, 1e39.

7

Firszt, J.B., Koch, D.B., Downing, M., Litvak, L., 2007. Current steering creates additional pitch percepts in adult cochlear implant recipients. Otol. Neurotol 28, 629e636. Hochmair, I., 2013. The importance of being flexible. Nat. Med 19, 1240e1244. House, W.F., 1976. Cochlear implants. Ann. Otol. Rhinol. Laryngol. 85 (Suppl. 27), 1e93. House, W.F., Urban, J., 1973. Long term results of electrode implantation and electronic stimulation of the cochlea in man. Ann. Otol. Rhinol. Laryngol 82, 504e517. Kals, M., Schatzer, R., Krenmayr, A., Vermeire, K., Visser, D., Bader, P., Neustetter, C., Zangerl, M., Zierhofer, C., 2010. Results with a cochlear implant channel-picking strategy based on “selected groups”. Hear. Res. 260, 63e69. Kiefer, J., Hohl, S., Sturzebecher, E., Pfennigdorff, T., Gstoettner, W., 2001. Comparison of speech recognition with different speech coding strategies (SPEAK, CIS, and ACE) and their relationship to telemetric measures of compound action potentials in the nucleus CI 24M cochlear implant system. Audiology 40, 32e42. Koch, D.B., Osberger, M.J., Segel, P., Kessler, D., 2004. HiResolution and conventional sound processing in the HiResolution bionic ear: using appropriate outcome measures to assess speech recognition ability. Audiol. Neurootol. 9, 214e223. Kral, A., Hartmann, R., Mortazavi, D., Klinke, R., 1998. Spatial resolution of cochlear implants: the electrical field and excitation of auditory afferents. Hear. Res. 121, 11e28. Lorens, A., Zgoda, M., Obrycka, A., Skarzynski, H., 2010. Fine structure processing improves speech perception as well as objective and subjective benefits in pediatric MED-EL COMBI 40þ users. Int. J. Pediatr. Otorhinolaryngol. 74, 1372e1378. McDermott, H.J., McKay, C.M., Vandali, A.E., 1992. A new portable sound processor for the University of Melbourne/Nucleus limited multielectrode cochlear implant. J. Acoust. Soc. Am. 91, 3367e3371. Merzenich, M.M., Rebscher, S.J., Loeb, G.E., Byers, C.L., Schindler, R.A., 1984. The UCSF Cochlear Implant Project. State of Development. In: Keidel, W.D., Finkenzeller, P. (Eds.), Advances in Audiology, Cochlear Implants in Clinical Use, Vol. 2. Karger, Basel, pp. 119e144. Middlebrooks, J.C., 2004. Effects of cochlear-implant pulse rate and inter-channel timing on channel interactions and thresholds. J. Acoust. Soc. Am. 116, 452e468. NIDCD, 2014. NIDCD Fact Sheet e Cochlear Implants [Online]. Available by National Institute on Deafness and Other Communication Disorders. https://www.nidcd. nih.gov/health/hearing/pages/coch.aspx (posted November 2013; verified December 10, 2014). Peeters, S., Marquet, J., Offeciers, F.E., Bosiers, W., Kinsbergen, J., Van Durme, M., 1989. Cochlear implants: the Laura prosthesis. Journal of medical engineering & technology 13, 76e80. Riss, D., Arnoldner, C., Reiss, S., Baumgartner, W.D., Hamzavi, J.S., 2009. 1-year results using the Opus speech processor with the fine structure speech coding strategy. Acta Otolaryngol. 129, 988e991. Schatzer, R., Bader, P., Griessner, A., Zierhofer, C., 2013. Place and rate dependency of channel interaction with simultaneous stimulation. Conference on Implantable Auditory Prostheses, Lake Tahoe, CA. Schatzer, R., Krenmayr, A., Au, D.K., Kals, M., Zierhofer, C., 2010. Temporal fine structure in cochlear implants: preliminary speech perception results in Cantonese-speaking implant users. Acta Otolaryngol. 130, 1031e1039. Seligman, P.M., Patrick, J.F., Tong, Y.C., Clark, G.M., Dowell, R.C., Crosby, P.A., 1984. A signal processor for a multiple-electrode hearing prosthesis. Acta Otolaryngol. Suppl. 411, 135e139. Simmons, F.B., Epley, J.M., Lummis, R.C., Guttman, N., Frishkopf, L.S., Harmon, L.D., Zwicker, E., 1965. Auditory Nerve: Electrical Stimulation in Man. Science 148, 104e106. Skinner, M.W., Clark, G.M., Whitford, L.A., Seligman, P.M., Staller, S.J., Shipp, D.B., Shallop, J.K., Everingham, C., Menapace, C.M., Arndt, P.L., et al., 1994. Evaluation of a new spectral peak coding strategy for the nucleus 22 channel cochlear implant system. Am. J. Otol. 15 (Suppl. 2), 15e27. Studebaker, G.A., 1985. A “rationalized” arcsine transform. J. Speech Hear. Res. 28, 455e462. Townshend, B., White, R.L., 1987. Reduction of electrical interaction in auditory prostheses. IEEE Trans. Biomed. Eng. BME 34, 891e897. Townshend, B., Cotter, N., Van Compernolle, D., White, R.L., 1987. Pitch perception by cochlear implant subjects. J. Acoust. Soc. Am. 82, 106e115. van den Honert, C., Kelsall, D.C., 2007. Focused intracochlear electric stimulation with phased array channels. J. Acoust. Soc. Am. 121, 3703e3716. Wilson, B.S., 2013. Toward better representations of sound with cochlear implants. Nat. Med 19, 1245e1248. Wilson, B.S., Lawson, D.T., Zerbi, M., Finley, C.C., 1994. Recent developments with the CIS strategies. In: Hochmair-Desoyer, I.J., Hochmair, E.S. (Eds.), Advances in Cochlear Implants. Proceedings of the 3rd International Cochlear Implant Conference, Innsbruck, Austria, April 1993. Manz, Vienna, pp. 103e112. Wilson, B.S., Finley, C.C., Lawson, D.T., Wolford, R.D., Eddington, D.K., Rabinowitz, W.M., 1991. Better speech recognition with cochlear implants. Nature 352, 236e238. Wilson, B.S., Finley, C.C., Farmer Jr., J.C., Lawson, D.T., Weber, B.A., Wolford, R.D., Kenan, P.D., White, M.W., Merzenich, M.M., Schindler, R.A., 1988. Comparative studies of speech processing strategies for cochlear implants. Laryngoscope 98, 1069e1077. Zeng, F.G., 2002. Temporal pitch in electric hearing. Hear. Res. 174, 101e106. Zierhofer, C.M., 2003a. Electrical nerve stimulation based on channel specific sampling sequences. U.S. patent 6,594,525 B1.

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002

8

R. Schatzer et al. / Hearing Research xxx (2014) 1e8

Zierhofer, C.M., 2003b. Multichannel cochlear implant with neural response telemetry. U.S. patent 6,600,955 B1. Zierhofer, C.M., Schatzer, R., 2008. Simultaneous intracochlear stimulation based on channel interaction compensation: analysis and first results. IEEE Trans. Biomed. Eng 55, 1907e1916. Zokoll, M.A., Hochmuth, S., Warzybok, A., Wagener, K.C., Buschermohle, M., Kollmeier, B., 2013. Speech-in-noise tests for multilingual hearing screening and diagnostics1. Am. J. Audiol 22, 175e178.

Zwolan, T.A., Kileny, P.R., Smith, S., Waltzman, S., Chute, P., Domico, E., Firszt, J., Hodges, A., Mills, D., Whearty, M., Osberger, M.J., Fisher, L., 2005. Comparison of continuous interleaved sampling and simultaneous analog stimulation speech processing strategies in newly implanted adults with a Clarion 1.2 cochlear implant. Otol. Neurotol. 26, 455e465.

Please cite this article in press as: Schatzer, R., et al., Speech perception with interaction-compensated simultaneous stimulation and long pulse durations in cochlear implant users, Hearing Research (2014), http://dx.doi.org/10.1016/j.heares.2014.11.002