
Purpose Ocular vestibular evoked myogenic potentials (oVEMPs) are widely used to assess utriculo-ocular function. However, oVEMPs have traditionally been elicited using transient stimuli, providing limited insight into temporal and nonlinear processing of the vestibular input. This study used amplitude-modulated stimuli to characterize temporal modulation transfer function (TMTF) and quantify nonlinear harmonic distortions in oVEMP responses. Methods Fourteen young adult females with normal hearing and no vestibular history were tested. Sinusoidally amplitude-modulated tones (500 Hz carrier, 100% modulation depth, 123 dB FL) were delivered via bone conduction across ten modulation frequencies (11 to 397 Hz). Responses were analyzed using fast Fourier transformation and quantified via amplitude, signal-to-noise ratio (SNR), phase coherence (PC), and harmonic components. Results Robust oVEMP responses were elicited across a broad range of modulation frequencies. Peak amplitude occurred at 79 Hz, while SNR and PC peaked between 79 and 113 Hz. Using a statistical SNR detection criterion of > 6.13 dB, the effective response range extended from 37 to 113 Hz. Harmonic distortion products were consistently detected up to the third harmonic, with peak response amplitudes and response rates converging near 110 Hz across harmonic components. Conclusions These findings provide the first characterization of the amplitude-modulated oVEMP TMTF and demonstrate temporal processing and nonlinear encoding within the utricular-ocular pathway. Extending this approach to vestibular disorders could help identify deficits not captured by transient measures, advance our understanding of vestibular pathophysiology, and improve clinical diagnosis. Optimal modulation frequencies for clinical and research applications of amplitude-modulated oVEMPs appear to lie between 37 and 113 Hz.
Recent advances in precision hearing diagnostics and machine learning are enabling a new era of personalized audio processing. The identification of cochlear synaptopathy (CS) as an early contributor to sensorineural hearing loss (SNHL) has motivated the development of novel, non-invasive diagnostic tools to complement standard audiometry. At the same time, novel machine-learning approaches have lifted the performance of noise reduction in hearing aids and introduce new avenues for audio processing that can compensate for different aspects of SNHL. However, integrating these insights into effective, individualized strategies for hearing aids and hearables remains challenging. Numerical approaches based on auditory modeling offer promising solutions. Computational auditory models can simulate the effects of SNHL on speech processing, providing insight into how audio processing may compensate for specific pathologies. When combined with modern machine-learning techniques, these models can be optimized so that simulations of normal and impaired auditory processing can guide the training of audio processing algorithms. These advanced nonlinear approaches may provide audio processing that partially restores impaired auditory function without relying on the prior assumptions underlying conventional hearing-aid processing, such as prescribed gain rules, wide dynamic-range compression, or fixed attack and release times. This paper provides an overview of existing and novel hearing loss compensation methods and an example of how precision hearing diagnostics, auditory modeling, and machine learning can be integrated to create personalized audio processing methods. DNN approaches to hearing diagnostics and treatment have the potential to drive the next generation of hearing aids and hearables, enabling targeted intervention even at the earliest stages of hearing damage.
The relative contributions of spectral and temporal cues to voice pitch perception in cochlear implant (CI) users remain unclear. This study aimed to directly quantify these contributions in dynamic speech stimuli with clinical n-of-m processors. Fundamental frequency (F0) discrimination was measured using consonant–vowel triplets in two experiments. Sixteen (eight female) postlingually deaf adult CI users with n-of-m strategies participated in experiment 1 (F0 range 76–242 Hz). Five (three female) also participated in experiment 2 (F0 range 242–768 Hz). Stimuli were presented acoustically or via direct stimulation. For the latter, pulse patterns were recorded in advance using the participants’ own settings and modified to suppress temporal or spectral cues. In experiment 1, temporal suppression drastically decreased F0 discrimination. Spectral suppression decreased the F0 discrimination of only some participants, while spectral and temporal suppression combined decreased performance further. Conversely, in experiment 2, spectral suppression caused a drop in F0 discrimination while temporal suppression had no effect. Analysis of the unaltered stimuli showed that with increasing F0, temporal cues decreased and spectral cues increased in availability. The novel prerecording and streaming method effectively combines realistic stimulation patterns with precise control at the pulse level. Our results indicate that in voice pitch discrimination, postlingually deafened adult CI users rely predominantly on the most available cue, which depends on the F0. Temporal cues dominate when discriminating F0s within the typical male to female voice range, while spectral cues tend to dominate within the typical female to child voice range.
Cochlear implant (CI) users struggle to understand speech under noisy conditions. This has long been attributed to channel interactions: a given neuron not only responds to electrical stimulation from the closest electrode but also is sensitive to stimulation from electrodes located further away. Shaping the stimulation voltage into a very sharp profile (e.g. using tripolar instead of the more common monopolar stimulation) has been proposed as a solution, only with limited success. We argue this is due to several factors: the strong correlation between envelopes of adjacent CI channels with speech sounds, the necessity to recruit enough neurons for sufficient loudness, and, at loud levels, the similarity between monopolar and sharp stimulation output voltages at locations that are at least 3–4 electrodes away from a given stimulation electrode. We describe a “Hilltop” method to effectively reduce the voltage more than 3–4 electrodes away from a given electrode, while keeping the same percept of place pitch as with monopolar stimulation. We assess this method psychophysically in human CI users (N = 7 males, 9 females) and with computational modelling. Based on computational modelling and on psychophysical measures of charge summation, we show that the voltage difference between monopolar and Hilltop stimulation persists at the level of the spiral ganglion neurons. The results warrant investigations with more complex, multichannel stimuli. We also describe possible limitations of this strategy, including the presence of sidelobes of opposite effective polarity in some participants, and suggest ways to mitigate their effects.
The amplitudes of the human compound action potential (CAP) and auditory brainstem response (ABR) are influenced by the cochlear traveling wave. To synchronize auditory nerve (AN) firing, short, rising-frequency chirps have been used to compensate for traveling wave delays; however, for high presentation levels, basal spread of excitation may disrupt AN synchrony. Further, unintended spread of excitation over the time course of the chirp may adapt basal AN fibers, thereby reducing CAP and ABR amplitudes—opposite of the objective of using chirps. Previous work has altered the rate of frequency change as a function of chirp level to maximize AN synchrony. Here, we evaluated, in normal-hearing adults (N = 12, six males), an alternative approach of attenuating low-frequency chirp energy and hypothesized this manipulation would increase CAP and ABR amplitudes relative to the original, flat-spectrum chirp for high stimulus levels. Low-frequency-attenuated chirps produced larger CAP and ABR wave amplitudes than clicks and the original chirp for nearly all stimulus levels. To explore mechanisms underlying these results, we derived AN firing distributions for chirps and clicks. The distributions for low-frequency-attenuated chirps were consistent with greater AN synchrony compared to those for clicks and the original chirp. ABR waveforms were reasonably simulated by convolving the derived AN firing distributions with an impulse response waveform associated with the ABR electrode montage. These findings suggest that low-frequency-attenuated chirps and predicted AN firing distributions may be useful in evaluating human AN function, including the effects of AN fiber loss, and other disease states.