Perception and prediction of loudness in sound coding strategies using simultaneous electric stimulation

A computational model to simulate spectral modulation and speech perception experiments of cochlear implant users

Speech understanding in cochlear implant (CI) users presents large intersubject variability that may be related to different aspects of the peripheral auditory system, such as the electrode-nerve interface and neural health conditions. This variability makes it more challenging to proof differences in performance between different CI sound coding strategies in regular clinical studies, nevertheless, computational models can be helpful to assess the speech performance of CI users in an environment where all these physiological aspects can be controlled. In this study, differences in performance between three variants of the HiRes Fidelity 120 (F120) sound coding strategy are studied with a computational model. The computational model consists of (i) a processing stage with the sound coding strategy, (ii) a three-dimensional electrode-nerve interface that accounts for auditory nerve fiber (ANF) degeneration, (iii) a population of phenomenological ANF models, and (iv) a feature extractor algorithm to obtain the internal representation (IR) of the neural activity. As the back-end, the simulation framework for auditory discrimination experiments (FADE) was chosen. Two experiments relevant to speech understanding were performed: one related to spectral modulation threshold (SMT), and the other one related to speech reception threshold (SRT). These experiments included three different neural health conditions (healthy ANFs, and moderate and severe ANF degeneration). The F120 was configured to use sequential stimulation (F120-S), and simultaneous stimulation with two (F120-P) and three (F120-T) simultaneously active channels. Simultaneous stimulation causes electric interaction that smears the spectrotemporal information transmitted to the ANFs, and it has been hypothesized to lead to even worse information transmission in poor neural health conditions. In general, worse neural health conditions led to worse predicted performance; nevertheless, the detriment was small compared to clinical data. Results in SRT experiments indicated that performance with simultaneous stimulation, especially F120-T, were more affected by neural degeneration than with sequential stimulation. Results in SMT experiments showed no significant difference in performance. Although the proposed model in its current state is able to perform SMT and SRT experiments, it is not reliable to predict real CI users' performance yet. Nevertheless, improvements related to the ANF model, feature extraction, and predictor algorithm are discussed.

Assessing the relationship between neural health measures and speech performance with simultaneous electric stimulation in cochlear implant listeners

OBJECTIVES

The relationship between electrode-nerve interface (ENI) estimates and inter-subject differences in speech performance with sequential and simultaneous channel stimulation in adult cochlear implant listeners were explored. We investigated the hypothesis that individuals with good ENIs would perform better with simultaneous compared to sequential channel stimulation speech processing strategies than those estimated to have poor ENIs.

METHODS

Fourteen postlingually deaf implanted cochlear implant users participated in the study. Speech understanding was assessed with a sentence test at signal-to-noise ratios that resulted in 50% performance for each user with the baseline strategy F120 Sequential. Two simultaneous stimulation strategies with either two (Paired) or three sets of virtual channels (Triplet) were tested at the same signal-to-noise ratio. ENI measures were estimated through: (I) voltage spread with electrical field imaging, (II) behavioral detection thresholds with focused stimulation, and (III) slope (IPG slope effect) and 50%-point differences (dB offset effect) of amplitude growth functions from electrically evoked compound action potentials with two interphase gaps.

RESULTS

A significant effect of strategy on speech understanding performance was found, with Triplets showing a trend towards worse speech understanding performance than sequential stimulation. Focused thresholds correlated positively with the difference required to reach most comfortable level (MCL) between Sequential and Triplet strategies, an indirect measure of channel interaction. A significant offset effect (difference in dB between 50%-point for higher eCAP growth function slopes with two IPGs) was observed. No significant correlation was observed between the slopes for the two IPGs tested. None of the measures used in this study correlated with the differences in speech understanding scores between strategies.

CONCLUSIONS

The ENI measure based on behavioral focused thresholds could explain some of the difference in MCLs, but none of the ENI measures could explain the decrease in speech understanding with increasing pairs of simultaneously stimulated electrodes in processing strategies.

Applications of Phenomenological Loudness Models to Cochlear Implants

Cochlear implants electrically stimulate surviving auditory neurons in the cochlea to provide severely or profoundly deaf people with access to hearing. Signal processing strategies derive frequency-specific information from the acoustic signal and code amplitude changes in frequency bands onto amplitude changes of current pulses emitted by the tonotopically arranged intracochlear electrodes. This article first describes how parameters of the electrical stimulation influence the loudness evoked and then summarizes two different phenomenological models developed by McKay and colleagues that have been used to explain psychophysical effects of stimulus parameters on loudness, detection, and modulation detection. The Temporal Model is applied to single-electrode stimuli and integrates cochlear neural excitation using a central temporal integration window analogous to that used in models of normal hearing. Perceptual decisions are made using decision criteria applied to the output of the integrator. By fitting the model parameters to a variety of psychophysical data, inferences can be made about how electrical stimulus parameters influence neural excitation in the cochlea. The Detailed Model is applied to multi-electrode stimuli, and includes effects of electrode interaction at a cochlear level and a transform between integrated excitation and specific loudness. The Practical Method of loudness estimation is a simplification of the Detailed Model and can be used to estimate the relative loudness of any multi-electrode pulsatile stimuli without the need to model excitation at the cochlear level. Clinical applications of these models to novel sound processing strategies are described.

Effect of Channel Interaction on Vocal Cue Perception in Cochlear Implant Users

Speech intelligibility in multitalker settings is challenging for most cochlear implant (CI) users. One possibility for this limitation is the suboptimal representation of vocal cues in implant processing, such as the fundamental frequency (F0), and the vocal tract length (VTL). Previous studies suggested that while F0 perception depends on spectrotemporal cues, VTL perception relies largely on spectral cues. To investigate how spectral smearing in CIs affects vocal cue perception in speech-on-speech (SoS) settings, adjacent electrodes were simultaneously stimulated using current steering in 12 Advanced Bionics users to simulate channel interaction. In current steering, two adjacent electrodes are simultaneously stimulated forming a channel of parallel stimulation. Three such stimulation patterns were used: Sequential (one current steering channel), Paired (two channels), and Triplet stimulation (three channels). F0 and VTL just-noticeable differences (JNDs; Task 1), in addition to SoS intelligibility (Task 2) and comprehension (Task 3), were measured for each stimulation strategy. In Tasks 2 and 3, four maskers were used: the same female talker, a male voice obtained by manipulating both F0 and VTL (F0+VTL) of the original female speaker, a voice where only F0 was manipulated, and a voice where only VTL was manipulated. JNDs were measured relative to the original voice for the F0, VTL, and F0+VTL manipulations. When spectral smearing was increased from Sequential to Triplet, a significant deterioration in performance was observed for Tasks 1 and 2, with no differences between Sequential and Paired stimulation. Data from Task 3 were inconclusive. These results imply that CI users may tolerate certain amounts of channel interaction without significant reduction in performance on tasks relying on voice perception. This points to possibilities for using parallel stimulation in CIs for reducing power consumption.