Aniket Saoji

Relationship between perception of spectral ripple and speech recognition in cochlear implant and vocoder listeners

Spectral resolution has been reported to be closely related to vowel and consonant recognition in cochlear implant (CI) listeners. One measure of spectral resolution is spectral modulation threshold (SMT), which is defined as the smallest detectable spectral contrast in the spectral ripple stimulus. SMT may be determined by the activation pattern associated with electrical stimulation. In the present study, broad activation patterns were simulated using a multi-band vocoder to determine if similar impairments in speech understanding scores could be produced in normal-hearing listeners. Tokens were first decomposed into 15 logarithmically spaced bands and then re-synthesized by multiplying the envelope of each band by matched filtered noise. Various amounts of current spread were simulated by adjusting the drop-off of the noise spectrum away from the peak (40-5 dBoctave). The average SMT (0.25 and 0.5 cyclesoctave) increased from 6.3 to 22.5 dB, while average vowel identification scores dropped from 86% to 19% and consonant identification scores dropped from 93% to 59%. In each condition, the impairments in speech understanding were generally similar to those found in CI listeners with similar SMTs, suggesting that variability in spread of neural activation largely accounts for the variability in speech perception of CI listeners.

Spectral modulation detection and vowel and consonant identifications in cochlear implant listenersa)

Speech understanding by cochlear implant listeners may be limited by their ability to perceive complex spectral envelopes. Here, spectral envelope perception was characterized by spectral modulation transfer functions in which modulation detection thresholds became poorer with increasing spectral modulation frequency (SMF). Thresholds at low SMFs, less likely to be influenced by spectral resolution, were correlated with vowel and consonant identifications [Litvak, L. M. et al. (2008). J. Acoust. Soc. Am. 122, 982-991] for the same listeners; while thresholds at higher SMFs, more likely to be affected by spectral resolution, were not. Results indicate that the perception of broadly spaced spectral features is important for speech perception.

Relationship between auditory function of nonimplanted ears and bimodal benefit

Objectives: Patients with a cochlear implant (CI) in one ear and a hearing aid in the other ear commonly achieve the highest speech-understanding scores when they have access to both electrically and acoustically stimulated information. At issue in this study was whether a measure of auditory function in the hearing aided ear would predict the benefit to speech understanding when the information from the aided ear was added to the information from the CI. Design: The subjects were 22 bimodal listeners with a CI in one ear and low-frequency acoustic hearing in the nonimplanted ear. The subjects were divided into two groups—one with mild-to-moderate low-frequency loss and one with severe-to-profound loss. Measures of auditory function included (1) audiometric thresholds at 750 Hz or lower, (2) speech-understanding scores (words in quiet and sentences in noise), and (3) spectral-modulation detection (SMD) thresholds. In the SMD task, one stimulus was a flat spectrum noise and the other was a noise with sinusoidal modulations at 1.0 peak/octave. Results: Significant correlations were found among all three measures of auditory function and the benefit to speech understanding when the acoustic and electric stimulation were combined. Benefit was significantly correlated with audiometric thresholds (r = −0.814), acoustic speech understanding (r = 0.635), and SMD thresholds (r = −0.895) in the hearing aided ear. However, only the SMD threshold was significantly correlated with benefit within the group with mild-to-moderate loss (r = −0.828) and within the group with severe-to-profound loss (r = −0.896). Conclusions: The SMD threshold at 1 cycle/octave has the potential to provide clinicians with information relevant to the question of whether an ear with low-frequency hearing is likely to add to the intelligibility of speech provided by a CI.

Use of Phantom Electrode Technique to Extend the Range of Pitches Available Through a Cochlear Implant

Objective: The range of pitch sensations available in cochlear implants (CIs) is conventionally thought to be limited by the location of the most apical and basal electrodes. However, partial bipolar stimulation, in which current is distributed to two intracochlear electrodes and one extracochlear electrode, can produce “phantom electrode” (PE) pitch percepts that extend beyond the pitch range available with physical electrodes. The goals of this study were (1) to determine the PE configuration that generated the lowest pitch relative to monopolar (MP) stimulation of the most apical electrode and (2) to determine the amount of pitch shift produced by different PE configurations. Design: Ten Advanced Bionics CI users (9 unilateral and 1 bilateral), implanted with the CII or HiRes 90k implant and the HiFocus 1, HiFocus 1j, or Helix electrode arrays participated in this study. PEs were created by simultaneously stimulating the primary and compensating electrodes in opposite phase. To test different PE configurations, the proportion of current delivered to the compensating electrode (σ) and the electrode separation between the primary and compensatory electrode (D) were varied. To estimate the relative pitch of PEs, the lowest pitched PEs with primary electrodes 4 and 8 were compared with subsets of MP electrodes (1, 2, 3, 4, 5 and 5, 6, 7, 8, 9, respectively). Results: In all subjects, it was possible to identify σ and D values that produced a PE that was lower in pitch than the MP stimulation of the primary electrode. In some subjects, increasing σ and/or D produced progressively lower pitch percepts, whereas in others, PE pitch changed nonmonotonically with σ and/or D. The amount of PE pitch shift could be estimated only for 14 cases; in seven cases, the pitch shift was <1 MP electrode, and in seven other cases, the pitch shift was between 1 and 2 MP electrodes. Conclusions: PE stimulation can elicit pitch percepts lower than that of the most apical MP electrode; the PE pitch is lower by the equivalent of 0.5 to 2 MP electrodes.

Excitation Patterns of Simultaneous and Sequential Dual-Electrode Stimulation in Cochlear Implant Recipients

Objective: Both simultaneous (SI) and sequential stimulation of intracochlear electrodes can be used to generate pitches that are intermediate to the physical electrodes (PEs). The goal of this study was to compare the spread of neural excitation for SI and sequential dual-electrode stimulation with the spread of neural excitation for the intermediate electrode using electrically evoked compound action potentials. Design: Seven Advanced Bionics cochlear implant users with either CII or HiRes 90k implant and HiFocus 1 or HiFocus 1j electrode array participated in this study. A masker-probe subtraction method was used to derive neural excitation patterns for SI nonadjacent dual-electrode stimulation, apical and basal-first sequential nonadjacent dual-electrode stimulation, and the intermediate PE. For apical-first sequential (SEa) stimulation, the masker pulse on the apical electrode immediately preceded the masker pulse on the basal electrode, and vice versa for basal-first sequential stimulation (SEb). The electrodes used for dual-electrode stimulation were separated by an intermediate PE, which represents a spatial distance of approximately 2 mm. Current levels necessary to achieve comfortable loudness were determined for each masker and probe stimulus. During the evoked compound action potential measurements, the masker was fixed in location, whereas the probe was varied across a subset of electrodes in the array. Neural responses were calculated by subtracting the response to the probe from the masked response. Results: Neural excitation patterns were normalized to their peak and analyzed in terms of their area and center of gravity. The area and center of gravity for SI nonadjacent dual-electrode stimulation were similar to those of the intermediate PE. In contrast, the area for the two modes of sequential nonadjacent dual-electrode (SEa and SEb) stimulation differed significantly from the intermediate PE. The center of gravity for SEa stimulation also differed significantly from the intermediate PE, whereas there was no significant difference in the center of gravity between SEb stimulation and the intermediate PE. Conclusions: Peripheral neural activation patterns suggest a similar spread of excitation for SI dual-electrode stimulation and the intermediate PE. The spread of excitation associated with sequential dual-electrode stimulation is generally different from the intermediate PE, and it varies depending on the order of the sequential pulses.

Electrocochleography in Cochlear Implant Recipients With Residual Hearing: Comparison With Audiometric Thresholds.

Objectives: To determine whether electrocochleography (ECoG) thresholds, especially cochlear microphonic and auditory nerve neurophonic thresholds, measured using an intracochlear electrode, can be used to predict pure-tone audiometric thresholds following cochlear implantation in ears with residual hearing. Design: Pure-tone audiometric thresholds and ECoG waveforms were measured at test frequencies from 125 to 4000 Hz in 21 Advanced Bionics cochlear implant recipients with residual hearing in the implanted ear. The “difference” and “summation” responses were computed from the ECoG waveforms measured from two alternating phases of stimulation. The interpretation is that difference responses are largely from the cochlear microphonic while summating responses are largely from the auditory nerve neurophonic. The pure-tone audiometric thresholds were also measured with same equipment used for ECoG measurements. Results: Difference responses were observed in all 21 implanted ears, whereas summation response waveforms were observed in only 18 ears. The ECoG thresholds strongly correlated (r2 = 0.87, n = 150 for difference response; r2 = 0.82, n = 72 for summation response) with audiometric thresholds. The mean difference between the difference response and audiometric thresholds was −3.2 (±9.0) dB, while the mean difference between summation response and audiometric thresholds was −14 (±11) dB. In four out of 37 measurements, difference responses were measured to frequencies where no behavioral thresholds were present. Conclusions: ECoG thresholds may provide a useful metric for the assessment of residual hearing in cochlear implant subjects for whom it is not possible to perform behavioral audiometric testing.

Design and Evaluation of a Cochlear Implant Strategy Based on a “Phantom” Channel

Unbalanced bipolar stimulation, delivered using charge balanced pulses, was used to produce “Phantom stimulation”, stimulation beyond the most apical contact of a cochlear implant’s electrode array. The Phantom channel was allocated audio frequencies below 300Hz in a speech coding strategy, conveying energy some two octaves lower than the clinical strategy and hence delivering the fundamental frequency of speech and of many musical tones. A group of 12 Advanced Bionics cochlear implant recipients took part in a chronic study investigating the fitting of the Phantom strategy and speech and music perception when using Phantom. The evaluation of speech in noise was performed immediately after fitting Phantom for the first time (Session 1) and after one month of take-home experience (Session 2). A repeated measures of analysis of variance (ANOVA) within factors strategy (Clinical, Phantom) and interaction time (Session 1, Session 2) revealed a significant effect for the interaction time and strategy. Phantom obtained a significant improvement in speech intelligibility after one month of use. Furthermore, a trend towards a better performance with Phantom (48%) with respect to F120 (37%) after 1 month of use failed to reach significance after type 1 error correction. Questionnaire results show a preference for Phantom when listening to music, likely driven by an improved balance between high and low frequencies.

Spectral cues for understanding speech in quiet and in noise

thresholds observed at 0.25 and 0.50 peaks/octave (Saoji et al., 2009). The bandwidth and audio-frequency range of the noise band carrier has little effect on SMD thresholds of NH listeners. That is, sensitivity to 1cycle/octave modulation obtained to a four-octave noiseband is not likely different than that obtained to a one-octave noiseband tested at the upper or lower range of the original broadband stimulus. In the present study, SMD thresholds were obtained to noiseband stimuli of differentbandwidthsand audio-frequency regionsat different spectral modulation frequencies. The SMD thresholds were compared with the listener’s speech understanding scores to determine the relationship between SMD thresholds and performance. The goal was to determine the spectral cues (i.e. spectral modulation frequencies and audio-frequency ranges) required to model speech understanding in quiet and in noise.

Enhanced hearing in noise for cochlear implant recipients: clinical trial results for a commercially available speech-enhancement strategy.

Objective To demonstrate benefits for speech perception and everyday listening in quiet and in noise with a speech-enhancement strategy called ClearVoice, which was designed to improve listening in complex acoustic environments without compromising hearing in quiet. Study Design A 2-week randomized crossover design was used to evaluate ClearVoice in 46 adults unilaterally implanted with a CII/HiRes 90K cochlear implant who had at least 6 months experience with HiRes Fidelity 120 sound processing. Speech perception was assessed using the AzBio sentences presented in quiet, in speech-spectrum noise and in multitalker babble. Subjective listening benefit and strategy preference were assessed with a questionnaire. ClearVoice has 3 gain settings (low, medium, and high), each intended as a full-time listening option according to individual preference. Speech understanding after acute use of ClearVoice-low was compared with HiRes Fidelity 120 during an initial test session. Speech perception abilities were compared with HiRes Fidelity 120 after 2 weeks of exclusive use of ClearVoice-medium, and after 2 weeks of exclusive use of ClearVoice-high. During a fifth week, participants were fit with 3 programs for comparison (HiRes Fidelity 120, ClearVoice-medium, and ClearVoice-high), after which, they reported preference and everyday listening benefits via a questionnaire. Results ClearVoice significantly improved speech understanding in speech-spectrum noise and multitalker babble, did not compromise listening in quiet, was preferred for everyday listening, and provided improved hearing in real-life situations. Conclusion ClearVoice improves hearing in noise for cochlear implant recipients who use HiRes Fidelity 120 sound processing.

Adding simultaneous stimulating channels to reduce power consumption in cochlear implants

&NA; Sound coding strategies for Cochlear Implant (CI) listeners can be used to control the trade‐off between speech performance and power consumption. Most commercial CI strategies use non‐simultaneous channel stimulation, stimulating only one electrode at a time. One could add parallel simultaneous stimulating channels such that the electrical interaction between channels is increased. This would produce spectral smearing, because the electrical fields of the simultaneous stimulated channels interact, but also power savings. The parallel channels produce a louder sensation than sequential stimulation. To test this hypothesis we implemented different sound coding strategies using a research interface from Advanced Bionics: the commercial F120 strategy using sequential channel stimulation (one channel equals two electrodes with current steering) and the Paired strategy, consisting of simultaneous stimulation with two channels. Here, the electrical field of both channels will interact, requiring less current on each channel to perceive the same loudness as with F120. However, channel interaction between the independent channels may reduce speech recognition or understanding. This can be diminished by adding an inverse‐polarity stimulation channel between both channels. This strategy is termed Paired with Flanks. Additionally, Triplet with three channels and an adjacent Flank style was investigated. For each strategy we measured speech intelligibility with the Hochmair‐Schulz‐Moser sentence test. Spectral resolution was assessed using a spectral modulation depth detection task. Results show that Paired without Flanks obtains similar performance while reducing the current by 20% on average compared to F120. Triplet with and without Flanks shows overall poorer performance when compared to F120. All strategies inhibit the option to increase the pulse width which would result in even further decreased power consumption. HighlightsNew cochlear implant stimulation mode for reduced electrical interaction proposed.Simultaneous stimulation with additional inversely phased electrodes (Flanks).Flanks can reduce electrical interaction to some degree.Polarity effects need to be considered when stimulating simultaneously with both polarities.Paired strategy shows best trade‐off between performance and power consumption.