Polarity effects on place pitch and loudness for three cochlear-implant designs and at different cochlear sites

Transcription

1 Polarity effects on place pitch and loudness for three cochlear-implant designs and at different cochlear sites Robert P. Carlyon, a) John M. Deeks, and Olivier Macherey MRC Cognition and Brain Sciences Unit, 15 Chaucer Road, Cambridge CB2 7EF, England (Received 18 July 2012; revised 4 May 2013; accepted 15 May 2013) Users of Advanced Bionics, MedEl, and Cochlear Corp. implants balanced the loudness of trains of asymmetric pulses of opposite polarities presented in monopolar mode. For the Advanced Bionics and MedEl users the pulses were triphasic and consisted of a 32-ls central phase flanked by two 32-ls phases of opposite polarity and half the amplitude. The central phase was either anodic (TP-A) or cathodic (TP-C). For the Cochlear Corp. users, pulses consisted of two 32-ls phases of the same polarity separated by an 8-ls gap, flanked by two 32-ls phases of the opposite polarity, each of which was separated from the central portion by a 58-ls gap. The central portion of these quadraphasic pulses was either anodic (QP-A) or cathodic (QP-C), and all phases had the same amplitude. The current needed to achieve matched loudness was lower for the anodic than for the cathodic stimuli. This polarity effect was similar across all electrode locations studied, including the most apical electrode of the MedEl device which stimulates the very apex of the cochlea. In addition, when quadraphasic pulses were presented in bipolar mode, listeners reported hearing a lower pitch when the central portion was anodic at the more apical, than at the more basal, electrode. The results replicate previous reports that, unlike the results of most animal studies, human cochlear implant listeners are more sensitive to anodic than to cathodic currents, and extend those findings to a wider range of cochlear sites, implant types, and pulse shapes. VC 2013 Acoustical Society of America. [ PACS number(s): Ts, Me, Pg [ELP] Pages: I. INTRODUCTION Contemporary cochlear implants (CIs) stimulate the auditory nerve (AN) with a series of symmetric, biphasic pulses. Such stimuli are inefficient, because charge injected into the neural membrane by one phase is immediately recovered by the following phase, but are used in order to prevent the damaging electrochemical reactions that can arise as a result of charge imbalance. Another consequence of these standard pulse shapes is that they make it hard to determine which aspects of the electrical stimulus most effectively excite the AN. Single-cell and compound action potential recordings from animals, using monophasic pulses, generally show that the cathodic ( negative ) phase is the more effective (Miller et al., 1999; Miller et al., 2004). However, recent experiments with human CI users indicated greater activation for anodic ( positive ) current (Macherey et al., 2006; Macherey et al., 2008; van Wieringen et al., 2008; Undurraga et al., 2010; Macherey et al., 2011; Undurraga et al., 2013). One line of evidence comes from stimuli with long inter-phase gaps (IPGs), which permit separate measurement of the effectiveness of each phase, using both the electrically evoked compound action potential (ECAP) and psychophysical masking methods (Macherey et al., 2008; Undurraga et al., 2010). Another finding comes from the use of pseudomonophasic (PS) pulses, in which a short high-amplitude phase is immediately followed by a long low-amplitude phase. It has been shown that when the a) Author to whom correspondence should be addressed. Electronic mail: bob.carlyon@mrc-cbu.cam.ac.uk high-amplitude phase is anodic [ PS-A ; Fig. 1(a)], less current is needed to produce a given loudness than when it is cathodic ( PS-C : Macherey et al., 2006). This finding is consistent with greater sensitivity to anodic stimulation if it is assumed that, for a given charge, the nerve is more efficiently activated by a short high than by a long low phase (e.g., van den Honert and Stypulkowski, 1984). There are a number of possible explanations for the discrepancy between the human and animal results. These include the use of supra-threshold vs threshold measures, differences in survival and degree of myelination of peripheral neural processes, and differences in the anatomical relationship between the implanted electrode and the stimulated nerve fibers (Ranck, 1975; Rubinstein, 1993; McIntyre and Grill, 1999; Rattay, 1999; Rattay et al., 2001). This last factor might be responsible for another species difference, in which ECAP thresholds in response to monophasic pulses were lower for the cathodic polarity in the cat, but for anodic pulses in the guinea pig (Miller et al., 1998). If the effect of polarity on auditory sensitivity does indeed depend on this anatomical relationship, then it might be influenced by the site of stimulation; this might particularly be the case for electrodes near the apex of the cochlea, because Rosenthal s canal (which contains the spiral ganglion) follows the cochlear duct only for the most basal turns (Kawano et al., 1996). Furthermore, there is post-mortem evidence that the degeneration of peripheral processes in human CI users is less marked at the apex of the cochlea, providing another reason why the site of activation (and hence, possibly, the polarity effect) might differ between apical and more-basal electrodes (Fayad and Linthicum, 2006). J. Acoust. Soc. Am. 134 (1), July /2013/134(1)/503/7/$30.00 VC 2013 Acoustical Society of America 503

2 FIG. 1. Schematic representation of the pseudomonophasic-anodic (PS-A), triphasic-anodic (TP-A), and quadraphasic-anodic (QP-A) waveforms. Recently, Undurraga et al. (2010) described some ECAP measures showing that polarity sensitivity was largely invariant across basal, middle, and apical electrodes of the Advanced Bionics HiRes90k and CII implants. In common with all of the recent experiments demonstrating higher sensitivity to anodic current in CI users, those findings were obtained using Advanced Bionics implants, which permit a wider choice of pulse shapes and IPGs than are available with other devices. A limitation, however, is that the electrode array is only approximately 16 mm long, precluding stimulation of sites close to the cochlear apex. The present study reports loudness balancing experiments between asymmetric pulse trains of opposite polarity in users of the Advanced Bionics, Cochlear Corp., and MedEl implants. The use of the latter implant is important because it has a much longer (30 mm) electrode array than the others, allowing us to measure polarity sensitivity at the very apex of the cochlea. Unfortunately, only the Advanced Bionics device permits the use of PS pulses, so for the Advanced Bionics and MedEl device we used triphasic (TP) pulses, in which a central phase has twice the amplitude of the first and third phases [Fig. 1(b)]. The research software available for the Cochlear device permits only symmetric pulses, so we used another asymmetric stimulus using two symmetric pulses of opposite leading polarity, in which the IPG was set to as long as possible (58 ls) and with the two pulses separated by the minimum possible delay [8 ls; Fig. 1(c), quadraphasic (QP) pulses]. A second aim of the study was to use QP pulses presented in bipolar mode to investigate the effects of polarity on placeof-excitation. We have previously shown, using PS pulses, that place pitch is lower when the short, high-amplitude phase is anodic at the more apical electrode of the bipolar pair than when it is anodic at the more basal electrode (Macherey et al., 2011; Macherey and Carlyon, 2012). This finding has recently been bolstered by ECAP evidence that AN activation is greater near the electrode of a bipolar pair that conveys the high-amplitude anodic phase of a PS pulse (Undurraga et al., 2012). Furthermore, we showed that when these stimuli are presented via two electrodes at the apex of the Advanced Bionics array, place pitch is lower than obtained with the clinically standard symmetric pulses in monopolar mode. A potential practical application of this finding might be when a short electrode array is inserted in order to preserve residual low-frequency hearing; it may be possible to extend the site of excitation more apically, without inserting the electrode further and possibly damaging the remaining hair cells. Unfortunately PS pulses are only possible with the Advanced Bionics device. We therefore wished to determine whether a similar finding could be obtained with QP pulses. The structure of this article is as follows. We first describe the loudness results, starting with TP pulse trains in users of the Advanced Bionics device, in order to provide a direct link to the previous body of work obtained with that make of implant. Next, we describe results from eight users of the MedEl implant, for TP pulses presented to the apex, middle, and base of the array, in order to compare polarity sensitivity across as wide a range of cochlear locations as possible. Third, we report the loudness results obtained using the QP pulses in users of the Cochlear device. Finally, we describe our place-pitch experiments obtained with QP pulses. II. EXPERIMENT 1: LOUDNESS JUDGMENTS A. Method 1. Subjects A total of 18 subjects took part, including 5, 8, and 5 users of the Advanced Bionics, MedEl, and Cochlear Corp. devices, respectively. Details of the devices and electrodes used for each subject are given in Table I. All listeners were implanted with standard-length arrays. 2. Stimuli All stimuli consisted of 400-ms pulse trains presented at a rate of 99 pulses per second (pps), similar to that used in a previous study of the polarity of PS pulses on loudness (Macherey et al., 2006). Presentation was always in monopolar mode, which, for the Cochlear device, used both the case and an extra-cochlear ball electrode as the return ( MP1 þ 2 mode). Symmetric pulses consisted of two 32-ls oppositepolarity phases separated by an IPG of 0 ls (Advanced Bionics), 2.1 ls (MedEl), or 8 ls (Cochlear). The polarity of the first phase was anodic (SYM-A) or cathodic (SYM-C). TP pulses [Fig. 1(b)] consisted of two same-polarity 32-ls phases separated, with zero (Advanced Bionics) or 2.1-ls (MedEl) IPG, by a 32-ls phase of the opposite polarity and twice the amplitude. An exception was made for MedEl listeners M5 and M6, for whom the phase duration was increased to 90 ls; this reduced the current levels used and avoided distortion associated with compliance limits of the implant. Compliance limits were not reached for any other participant. The polarity of the high-amplitude central phase was either anodic (TP-A) or cathodic (TP-C). QP-A pulses [Fig. 1(c)] consisted of a SYM-C pulse with an IPG of 58 ls 504 J. Acoust. Soc. Am., Vol. 134, No. 1, July 2013 Carlyon et al.: Polarity effects on loudness and pitch in cochlear implants

3 TABLE I. The first three columns show the subject number, device type, and electrodes tested for each subject. Note that electrodes are numbered from apex to base for the MedEl and Advanced Bionics devices, but from base to Apex for the Cochlear device. The remaining columns show the CLLs for each subject and electrode for the SYM-A stimulus and, depending on implant type, the TP-A or QP-A stimulus used in Experiment 1. CLL (SYM-A): db re 1 ma CLL (TP/QP-A): db re 1mA Subject Device Electrodes Apical Mid Basal Apical Mid Basal M1 Pulsar_ci100 1,7, M2 Pulsar_ci100 1,7, M3 Pulsar_ci100 2,7, M4 Sonata_ti100 1,7, M5 Sonata_ti100 1,7, M6 Sonata_ti100 1,7, M7 Sonata_ti100 1, M8 Sonata_ti100 1,7, A1 HiRes90k 1, A2 HiRes90k 1, A3 HiRes90k 1, A4 HiRes90k 1, A5 HiRes90k 1, C1 CI24M 22, C2 CI24M 22, C3 CI24M 20, C4 CI24CA 19, C5 CI24M 20, followed 8 ls later by a SYM-A pulse having the same IPG. QP-C pulses were identical but of opposite polarity. In the Cochlear device all electrodes are shorted between pulses when the monopolar mode is used, potentially leading to a small amount of current leakage during this gap. Stimulus generation and presentation was achieved using research software and hardware provided by each manufacturer: BEDCS (Advanced Bionics), RIB2 (MedEl), and NIC2 (Cochlear). For the Advanced Bionics and Nucleus patients we used the APEX software interface (Laneau et al., 2005) for stimulus presentation and data collection. In each case this served as a wrapper around the BEDCS and NIC2 routines. The APEX software was modified from the publicly available version to incorporate BEDCS functionality. For the MedEl patients, custom MATLAB programs called the RIB2 routines and collected the data. In all cases stimuli were checked using a test implant and a digital storage oscilloscope. For most subjects, stimuli were presented, in different conditions, to each of two electrodes, at the apex and middle of the electrode array. For seven of the eight MedEl users, data were also obtained from a basal electrode. However, because of time constraints, we sometimes omitted measurements with symmetric pulses on one or more electrodes. Because of this, and because no significant effects of polarity were observed for symmetric pulses, we focus instead on the results obtained for asymmetric pulse shapes. 3. Procedure For each subject and electrode, the SYM-A pulse train was initially set to a comfortable loudness level ( CLL ), defined as 6 Most Comfortable on the Advanced Bionics Loudness Scale. The TP-A pulse train was then loudnessbalanced to it. Finally, the TP-C stimulus was loudnessbalanced to TP-A and the SYM-C stimulus was loudness balanced to SYM-A. (For the Cochlear device we used QP instead of TP pulses.) Throughout this article we present our results in terms of the polarity effect, defined as the loudness-balanced current for the cathodic stimulus minus that for the anodic stimulus of the same shape. The value of the CLL for each stimulus and electrode for the SYM-A and TP-A stimuli are given in Table I. The loudness balancing task was similar to that described in previous publications from our laboratory (Carlyon et al., 2010; Macherey et al., 2011), and was based on a method described by McKay and McDermott (1998). For each pair of stimuli (call them A and B ) to be compared, two adjustments were initially performed. First, the level of A was fixed and that of B adjusted by the subject. B was fixed at this adjusted level and A was then adjusted to match it. The loudness-balanced level for A was defined as the level at which it was initially fixed, and that for B differed from this value by the average difference between A and B at the end of each adjustment. The two stimuli were presented sequentially separated by 500 ms, with the adjustable stimulus presented last. The subject then adjusted its level for the next presentation by clicking one of six different virtual buttons on a computer screen, corresponding to three different step sizes in each direction. Subjects were encouraged to bracket each match before indicating that the stimuli were of equal loudness by clicking on a seventh button. The different step sizes were 1, 2, and 4 bits (where 1 bit corresponds to approximately 1.18, 2.36, 4.72, or 9.45 la, depending on the amplitude range required by the J. Acoust. Soc. Am., Vol. 134, No. 1, July 2013 Carlyon et al.: Polarity effects on loudness and pitch in cochlear implants 505

4 FIG. 2. Difference in current level needed to obtain equal loudness for cathodic vs anodic TP [(a) Advanced Bionics and (b) MedEl users] or QP [(c) Cochlear users] waveforms. A positive value indicates that the anodic stimulus was adjusted to a lower level. Filled squares show mean data across listeners, 6one standard error. For each device, the (open) symbol used for each subject number (Table I) was determined by the following scheme: 1: Circles, 2: Upward triangle, 3: Downward triangle, 4: Diamond, 5: Hourglass, 6: Asterisk, 7: Plus sign, 8: Square. Only the first five symbols were used for the Advanced Bionics and Cochlear groups. subject) for MedEl, 4, 8, or 12 la for Advanced Bionics, and 1, 2, or 3 current units (CUs, where 1 CU corresponds to 0.17 db) for the Cochlear device. This whole procedure was repeated between six and eight times for each electrode and subject. The mean values for each subject and condition were then entered into repeated-measures analyses of variances (ANOVAs), using the Huynh-Feldt sphericity correction. B. Results The polarity effect for the TP stimuli and for the Advanced Bionics users is plotted in Fig. 2(a) for electrodes at the apex and middle of the array. The mean values were 1.6 and 1.7 db for the apical and middle electrodes, respectively. These values are very similar and show that a lower current was needed for the TP-A than for the TP-C stimulus at both sites. This is consistent with the analogous finding for PS-A and PS-C pulse trains reported by Macherey et al. (2006), and with the conclusion that anodic current most effectively stimulates the AN of human CI users. A twoway (electrode X polarity) repeated-measures ANOVA on the balanced levels revealed a significant effect of polarity [(F1,4) ¼ 8.9; p < 0.05)]. There was also a significant effect of electrode [F(1,4) ¼ 51.1; p < 0.01], indicating that the balanced levels were lower overall for the apical than for the middle electrode ( 7.5 vs 5.3 db re 1 ma). Note that the overall stimulus level would have been primarily determined by the CLL level set for the SYM-A stimulus, and so the between-electrode difference does not reflect anything particular to the asymmetric pulse shapes used here. Indeed, the difference between the CLLs for the SYM-A stimulus at the apical vs middle electrodes was 2.4 db, similar to the 2.1 db difference for the balanced TP-A stimuli. The interaction between polarity and electrode was not significant. The polarity effect for the TP stimuli and for the MedEl users is plotted in Fig. 2(b), for electrodes at the apex, middle, and base of the array. It can be seen that the overall size of the effect is very similar across electrodes. One subject (M5) showed an anomalous finding in that more current was required for TP-A than for TP-C at the basal electrode, and another subject (M2) showed the same finding for the apical electrode. Data from one subject at the base of the array are missing, so the results are described and analyzed in two ways. For the seven subjects tested at all electrodes, the polarity effect was 1.14, 1.18, and 0.34 db at the apical, middle, and basal sites, respectively. A two-way ANOVA revealed a main effect of polarity [F(1,6) ¼ 14.69, p < 0.01], an effect of electrode [F(2,12) ¼ 7.89, p < 0.01], but, crucially, no interaction [F(2,12) ¼ 0.96, p ¼ 0.40]. The absence of an interaction means that, although less current was needed for TP-A than for TP-C pulse trains, the size of this polarity effect did not differ across the electrode array. The main effect of the electrode reflected the fact that, as for the Advanced Bionics listeners, the mean CLL was lower (for both polarities) for the apical electrode ( 9.1 db re 1mA) than for the middle ( 6.5 db) or basal ( 6.8 db) electrodes. The pattern of results was very similar when the data from all eight subjects were analyzed, this time including only those electrodes (apical and middle) tested for all subjects. There were once again significant effects of polarity [F(1,7) ¼ 8.12, p ¼ 0.025] and of electrode [F(1,7) ¼ 10.79, p < 0.02], but no interaction. The polarity effect for the QP stimuli and for the Cochlear device users is plotted in Fig. 2(c), for electrodes at the apex and middle of the array. The mean values were 1.2 and 1.4 db for the apical and middle electrodes, respectively. A repeated measures ANOVA revealed, once again, significant effects of polarity [F(1,4) ¼ 17.8, p < 0.02] and place [F(1,4) ¼ 17.0, p < 0.02], with the direction of the effects being the same as for the other two devices (lower current levels needed for anodic stimuli and for apical stimulation). Averaged across polarity, the mean current levels for the QP stimuli were 10.4 and 7.9 db re 1mAforthe apical and mid-array electrodes, respectively. These values are somewhat lower than for the other devices, probably due to the fact that the total duration of all the phases in a QP pulse is twice that for a TP pulse (Fig. 1). There was no interaction between electrode and polarity. Note that the size of the polarity effect was at least as large for the QP stimulus as for the TP stimuli presented to the Advanced Bionics and MedEl listeners. 506 J. Acoust. Soc. Am., Vol. 134, No. 1, July 2013 Carlyon et al.: Polarity effects on loudness and pitch in cochlear implants

5 III. EXPERIMENT 2. PITCH JUDGMENTS A. Methods 1. Subjects and stimuli Four of the Cochlear subjects from experiment 1 (all except C4) took part. For each subject, stimuli were presented at either the apex or the middle of the electrode array, on electrode pairs (22,19) and (13,10), respectively. An exception was made for subject C5 for whom the apical pair was electrodes (20,17) because electrodes 19 and 22 were not enabled in the clinical map. The duration of all pulse trains was 400 ms. QP pulses were used, as in Experiment 1. As the stimuli were presented in bipolar mode the waveforms at the two electrodes were polarity-inverted versions of each other. Here we use the term QP-A to refer to the situation where the central portion of the pulse is anodic at the more apical electrode of the bipolar pair, and QP-C to refer to the opposite case. Prior to the main part of the experiment QP-A and QP-C were loudness balanced to each other using the same procedure as in Experiment 1, with the exception that levels were based on an average of four matches per condition. Pitch judgments were obtained at two pulse rates, 25 and 1031 pps. As in a previous study (Macherey and Carlyon, 2012), the lower rate was chosen so as to avoid any possible influence of temporal pitch cues, whereas the higher rate was more representative of the rates used in speech processing strategies. In each trial the subject was presented, sequentially, with two QP pulse trains of opposite polarity, in random order and separated by a 300-ms gap. They were required to indicate which stimulus had the higher pitch. No feedback was given after each response. Each combination of electrode and pulse rate was run in two blocks of 50 trials, leading to a total of 100 trials per data point. An exception was subject C2 who only did 50 trials/point for the 25-pps pulse trains. B. Results Figure 3 shows the percentage of trials in which the QP-A stimulus was judged higher in pitch than the QP-C FIG. 3. Percentage of trials on which the QP-A stimulus was judged higher in pitch than the QP-C stimulus in Experiment 2. (a) Pulse rate ¼ 25 pps. (b) Pulse rate ¼ 1031 pps. The solid horizontal line shows chance performance, with 95% confidence intervals shown by the dashed lines. Symbols for individual subjects are the same as in Fig. 2. stimulus. The horizontal solid line at 50% represents chance performance, and the dashed lines show 95% confidence limits based on 100 trials/point. For the 25-pps pulse trains [Fig. 3(a)], all subjects consistently reported the QP-A stimulus as having a lower pitch than the QP-C stimulus, for both electrodes. This is consistent with the pattern of results obtained for PS pulses with the Advanced Bionics device (Macherey et al., 2011; Macherey and Carlyon, 2012) and with the idea that anodic current is more effective than cathodic current at exciting the AN of human CI users. A similar pattern was observed for the pps pulse trains [Fig. 3(b)], with two exceptions (C1 on the apical pair and C2 on the middle pair of electrodes). In addition, the size of the effect was somewhat reduced for subject C5. This mirrors our observation in a recent article (Macherey and Carlyon, 2012) that polarity effects on place pitch in bipolar mode are less consistent at high rates. In that article we noted that one subject reversed his pitch judgments from session to session. Although we cannot be sure of the reason for this, we argued that it may be due to temporal pitch cues being stronger for stimuli presented to more apical sites (Macherey et al., 2011), and with this counteracting the effects of place-of-excitation. IV. DISCUSSION A. Comparison with loudness effects observed with different pulse shapes Averaged across electrodes, the polarity effect observed with TP pulses was 1.7 db for the Advanced Bionics device and 0.7 db for the MedEl implant. These values are somewhat smaller than the 2.5 db effect observed by Macherey et al. (2006) for 99-pps PS pulse trains with a 97-ls phase duration. This difference is unlikely to be due to the longer overall phase duration used by them, because they performed a supplementary experiment that showed no effect of phase duration on the polarity effect. The smaller effect observed here with TP pulses is perhaps not surprising, because the ratio between the duration of the central high-amplitude phase and the total duration of the two flanking low-amplitude phases of our TP stimuli was only two, whereas the long low phase of Macherey et al. PS stimuli was eight times longer than that of the short high phase. To provide a more direct comparison of the two pulse shapes, we tested two Advanced Bionics listeners with PS pulses having duration ratios of two and eight, using an electrode in the middle of the array, as in the Macherey et al. study. For these two listeners the average polarity effect was 1.24 db for the TP stimuli, and 1.09 and 1.77 db for the PS stimuli with duration ratios of two and eight, respectively. Hence, for a given duration ratio, the size of the duration effect does not appear to depend markedly between the two waveform shapes. B. Differences in CLLs across electrodes CLLs were significantly lower for stimuli presented to the most apical electrode of each device than they were for the other electrodes. This is surprising because, for stimuli J. Acoust. Soc. Am., Vol. 134, No. 1, July 2013 Carlyon et al.: Polarity effects on loudness and pitch in cochlear implants 507

6 presented in monopolar mode, CLLs are usually largely constant across the electrode array (e.g., Bierer, 2007). One possible reason for this discrepancy is that these equalloudness profiles are usually obtained at the much higher pulse rates currently employed in signal-processing strategies, rather than the 99-pps used here. However, it should be stressed that we did not formally loudness balance the stimuli presented to different electrodes against each other, but rather adjusted the current separately for the SYM-A stimuli on each electrode so that the subject indicated that it was at a comfortable level. Hence some caution should be exercised when interpreting this result. C. Practical implications Although the present results reveal a difference in CLL between TP-A and TP-C stimuli, the difference observed is unlikely to have significant implications for power consumption, which in contemporary devices is likely to be dominated by signal processing rather than by the efficiency of the electrode-neuron interface (Seligman and Shepherd, 2004). Indeed, when we calculated the mean CLLs across all subjects and for the middle electrodes of the MedEl and Advanced Bionics devices, the value of 6.7 db re 1 ma for the TP-A stimulus was significantly greater than that for the standard symmetric pulse shapes used clinically (e.g., 7.7 db for SYM-A; t-test df ¼ 11, p < 0.001). Note, however, that this finding does not apply to all asymmetric pulse shapes; for example, Macherey et al. (2006) asked subjects to loudnessbalance 99-pps pulse trains, and found that the level needed for a PS-A stimulus, with a first phase duration of 97 ls and a second phase duration eight times longer, was 3.4 db lower than for the corresponding SYM-A stimulus. The effect of pulse shape on overall sensitivity is well captured by the model described by Carlyon et al. (2005). The model passes the electrical waveform through a low pass filter and estimates threshold from the root-mean-squared level at the output of a 10-ms integration window. For the stimuli used here, the model predicts that the TP stimulus should have a threshold 0.8 db higher than for the SYM stimulus, whereas that for a PS stimulus with a duration ratio of 8 should by 3.7 db lower than for the SYM stimulus. Note, however, that the model does not attempt to account for the effects of polarity, and models thresholds rather than CLLs. One potential application does, however, arise from the observation that using symmetric stimuli in bipolar mode produces a locus of excitation close to that electrode where the most effective phase is anodic (Macherey et al., 2010; Macherey et al., 2011). Previous evidence was obtained using PS pulses with the Advanced Bionics device, and showed that it was possible to produce a lower place pitch than is achievable with symmetric pulses using either bipolar or monopolar stimulation (Macherey et al., 2011). As noted in Sec. I, this might be useful when trying to preserve residual low-frequency hearing, as it would (albeit slightly) allow one to reduce insertion depth while still exciting a sufficiently apical part of the cochlea. The results of Macherey et al. indicate that the shift in place-of-excitation from using asymmetric pulses in bipolar mode would correspond to between about 0.6 and 1.1 mm. The finding that this neural steering can be achieved using QP pulses means that it could be implemented in the Cochlear device, one version of which has a short electrode array suitable for hearing preservation. In addition, this form of activation is likely to be more unimodal than the bimodal pattern of excitation that occurs with symmetric pulses in bipolar mode, a fact that might lead to better speech perception (Frijns et al., 1996; Undurraga et al., 2012). D. Summary (1) The polarity sensitivity observed in previous psychophysical and electrophysiological studies with human CI users generalizes to a wide range of cochlear sites, including those stimulated by the apical end of the 30- mm MedEl electrode array. (2) Polarity effects in monopolar mode can be observed using QP and TP pulse shapes. The size of the effect falls roughly in between that obtained with PS stimuli having duration ratios of two and eight. (3) Our results suggest that it will be possible to manipulate the place of excitation in the cochlea using TP and QP waveforms in bipolar mode. This may include the excitation of slightly more apical sites than is possible with symmetric waveforms. Bierer, J. A. (2007). Threshold and channel interaction in cochlear implant users: Evaluation of the tripolar electrode configuration, J. Acoust. Soc. Am. 121, Carlyon, R. P., Deeks, J. M., and McKay, C. M. (2010). The upper limit of temporal pitch: Stimulus duration, conditioner pulses, and the number of electrodes stimulated, J. Acoust. Soc. Am. 127, Carlyon, R. P., van Wieringen, A., Deeks, J. M., Long, C. J., Lyzenga, J., and Wouters, J. (2005). Effect of inter-phase gap on the sensitivity of cochlear implant users to electrical stimulation, Hear. Res. 205, Fayad, J. N., and Linthicum, F. H., Jr. (2006). Multichannel cochlear implants: Relation of histopathology to performance, Laryngoscope 116, Frijns, J. H. M., de Snoo, S. L., and ten Kate, J. H. (1996). Spatial selectivity in a rotationally symmetric model of the electrically stimulated cochlea, Hear. Res. 95, Kawano, A., Seldon, H. L., and Clark, G. M. (1996). Computer-aided three-dimensional reconstruction in human cochlear maps: Measurement of the lengths of organ of corti, outer wall, inner wall, and Rosenthal s canal, Ann. Otol. Rhinol. Laryngol. 105, Laneau, J., Boets, B., Moonen, M., van Wieringen, A., and Wouters, J. (2005). A flexible auditory research platform using acoustic or electric stimuli for adults and young children, J. Neurosci. Methods 142, Macherey, O., and Carlyon, R. P. (2012). Place-pitch manipulations with cochlear implants, J. Acoust. Soc. Am. 131, Macherey, O., Carlyon, R. P., Deeks, J. M., van Wieringen, A., and Wouters, J. (2008). Higher sensitivity of human auditory nerve fibers to positive electrical currents, J. Assoc. Res. Otolaryngol. 9, Macherey, O., Deeks, J. M., and Carlyon, R. P. (2011). Extending the limits of place and temporal pitch perception in cochlear implant users, J. Assoc. Res. Otolaryngol. 12, Macherey, O., van Wieringen, A., Carlyon, R. P., Deeks, J. M., and Wouters, J. (2006). Asymmetric pulses in cochlear implants: Effects of pulse shape, polarity and rate, J. Assoc. Res. Otolaryngol. 7, Macherey, O., van Wieringen, A., Carlyon, R. P., Dhooge, I., and Wouters, J. (2010). Forward-masking patterns produced by symmetric and asymmetric pulse shapes in electric hearing, J. Acoust. Soc. Am. 127, McIntyre, C. C., and Grill, W. M. (1999). Excitation of central nervous system neurons by nonuniform electric fields, Biophys. J. 76, J. Acoust. Soc. Am., Vol. 134, No. 1, July 2013 Carlyon et al.: Polarity effects on loudness and pitch in cochlear implants

7 McKay, C. M., and McDermott, H. J. (1998). Loudness perception with pulsatile electrical stimulation: The effect of interpulse intervals, J. Acoust. Soc. Am. 104, Miller, C. A., Abbas, P. J., Hay-McCutcheon, M. J., Robinson, B. K., Nourski, K. V., and Jeng, F.-C. (2004). Intracochlear and extracochlear ECAPs suggest antidromic action potentials, Hear. Res. 198, Miller, C. A., Abbas, P. J., Robinson, B. K., Rubinstein, J. T., and Matsuoka, A. J. (1999). Electrically evoked single-fiber action potentials from cat: Responses to monopolar, monophasic stimulation, Hear. Res. 130, Miller, C. A., Abbas, P. J., Rubinstein, J. T., Robinson, B. K., Matsuoka, A. J., and Woodworth, G. (1998). Electrically evoked compound action potentials of guinea pig and cat: Responses to monopolar, monophasic stimulation, Hear. Res. 119, Ranck, J. B. (1975). Which elements are excited in electrical stimulation of mammalian nervous system: A review, Brain Res. 98, Rattay, F. (1999). The basic mechanism for the electrical stimulation of the nervous system, Neuroscience 89, Rattay, F., Lutter, P., and Felix, H. (2001). A model of the electrically excited human cochlear neuron I. Contribution of neural substructures to the generation and propagation of spikes, Hear. Res. 153, Rubinstein, J. T. (1993). Axon termination conditions for electrical simulation, IEEE Trans. Biomed. Eng. 40, Seligman, P. M., and Shepherd, R. K. (2004). Cochlear implants, in Neuroprosthetics: Theory and Practice, edited by K. W. Horch and G. S. Dhillon (World Scientific Publishing, Singapore). Undurraga, J. A., Carlyon, R. P., Wouters, J., and van Wieringen, A. (2012). Spread of excitation varies for different electrical pulse shapes and stimulation modes in cochlear implants, Hear. Res. 290, Undurraga, J. A., van Wieringen, A., Carlyon, R. P., Macherey, O., and Wouters, J. (2010). Polarity effects on neural responses of the electrically stimulated auditory nerve at different cochlear sites, Hear. Res. 269, Undurraga, J. A., van Wieringen, A., Carlyon, R. P., Macherey, O., and Wouters, J. (2013). The polarity sensitivity of the electrically stimulated human auditory nerve measured at the level of the brainstem, J. Assoc. Res. Otolaryngol. 14(3), van den Honert, C., and Stypulkowski, P. H. (1984). Physiological properties of the electrically stimulated auditory nerve. II. Single fiber recordings, Hear. Res. 14, van Wieringen, A., Macherey, O., Carlyon, R. P., Deeks, J. M., and Wouters, J. (2008). Alternative pulse shapes in electrical hearing, Hear. Res. 242, J. Acoust. Soc. Am., Vol. 134, No. 1, July 2013 Carlyon et al.: Polarity effects on loudness and pitch in cochlear implants 509