Trends in Hearing最新文献

Cochlear-implant listeners show impaired pitch perception compared to normal-hearing listeners. One of the factors limiting pitch sensitivity in multi-electrode as compared to single-electrode stimulation can be intracochlear interactions of electrode signals (i.e., channels). We measured temporal-pitch discrimination sensitivity for loudness-balanced dual-electrode stimuli with various spatio-temporal configurations in listeners with MED-EL implants. We hypothesized a link between pitch sensitivity and tonotopic separation as well as (monaural) temporal electrode asynchrony, the latter resulting in various combinations of inter-pulse intervals in the compound stimuli received by the auditory nerve. Per-electrode stimulus types were high-rate (i.e., 1,000-pps) pulse trains with a 100-Hz amplitude modulation and both with and without additional pulses inserted with short inter-pulse intervals at modulation peaks. The temporal asynchrony had a detrimental effect for tonotopic separations below 2.2 mm but not for separations of 7.1 mm and more. This pattern was largely consistent across stimulus types and can be attributed to spectro-temporal channel interactions. When compared with sensitivity to unmodulated 100-pps pulse trains [Lindenbeck et al., Trends in Hearing, 28, Article 23312165241271340 (2024)], stimuli without short inter-pulse interval pulses yielded lower sensitivity while stimuli with short inter-pulse interval pulses approached low-rate sensitivity for some tonotopic separations. Despite lower sensitivity overall, high-rate pitch cues seemed to be integrated (i.e., improved) more across the two electrodes than low-rate pitch cues when compared to single-electrode stimulation. These results suggest that short inter-pulse interval pulses are beneficial for temporal-pitch sensitivity in dual-electrode configurations.

We describe the development and validation of a self-administered online hearing test, which screens for hearing loss and provides an estimated audiogram. The hearing test computes test results from age, self-reported hearing abilities, and self-assessed pure-tone thresholds. It relies on regression, Bayesian and binary classification, leveraging probabilistic effects of age as well as interfrequency and interaural relationships in audiograms. The test was devised based on development data, collected prospectively in an online experiment from a purposive convenience sample of 251 adult American, Australian, Canadian, and Swiss participants, 58% of whom had hearing loss. Later, we externally validated the hearing test. Validation data were collected prospectively from a representative sample of 156 adult Belgian participants, 15% of whom had hearing loss. Participants completed the hearing test and audiometric assessments at home. The results for the primary screening outcome showed that the hearing test screened for mild hearing losses with a sensitivity of 0.83 [95%-confidence interval (CI): 0.65, 0.96], specificity of 0.94 [CI: 0.89, 0.98], positive predictive value of 0.70 [CI: 0.57, 0.87], and negative predictive value of 0.97 [CI: 0.94, 0.99]. Results for the secondary audiogram estimation outcome showed mean differences between estimated and gold standard hearing thresholds ranging from 2.1 to 12.4 dB, with an average standard deviation of the differences of 14.8 dB. In conclusion, the hearing test performed comparably to state-of-the-art hearing screeners. This test, therefore, is a validated alternative to existing screening tools, and, additionally, it provides an estimated audiogram.

While hearing aids are beneficial in compensating for hearing loss and suppressing ambient noise, they may also introduce an unwanted processing burden to the listener's sensory and cognitive system. To investigate such adverse side effects, hearing aids may be set to a 'transparent mode', aiming to replicate natural hearing through the open ear as best as possible. Such transparent hearing aids have previously been demonstrated to exhibit a small but significant disadvantage in speech intelligibility, with less conclusive effects on self-rated listening effort. Here we aimed to reproduce these findings and expand them with neurophysiological measures of invested listening effort, including parietal alpha power and pupil size. Invested listening effort was measured across five task difficulties, ranging from nearly impossible to easy, with normal-hearing participants in both aided and unaided conditions. Results well reproduced a hearing aid disadvantage for speech intelligibility and subjective listening effort ratings. As to be expected, pupil size and parietal alpha power followed an inverted u-shape, peaking at moderate task difficulties (around SRT50). However, the transparent hearing aid increased pupil size and parietal alpha power at medium task demand (between SRT20 and SRT80). These neurophysiological effects were larger than those observed in speech intelligibility and subjective listening effort, respectively. The results gain plausibility by yielding a substantial association of individual pupil size and individual parietal alpha power. In sum, our findings suggest that key neurophysiological measures of invested listening effort are sensitive to the individual additional burden on speech intelligibility that hearing aid processing can introduce.

The recognition of isolated words in noise improves as words are delayed from the noise onset. This phenomenon, known as adaptation to noise, has been mostly investigated using synthetic noises. The aim here was to investigate whether adaptation occurs for realistic noises and to what extent it depends on the spectrum and level fluctuations of the noise. Forty-nine different realistic and synthetic noises were analyzed and classified according to how much they fluctuated in level over time and how much their spectra differed from the speech spectrum. Six representative noises were chosen that covered the observed range of level fluctuations and spectral differences but could still mask speech. For the six noises, speech reception thresholds (SRTs) were measured for natural and tone-vocoded words delayed 50 (early condition) and 800 ms (late condition) from the noise onset. Adaptation was calculated as the SRT improvement in the late relative to the early condition. Twenty-two adults with normal hearing participated in the experiments. For natural words, adaptation was small overall (mean = 0.5 dB) and similar across the six noises. For vocoded words, significant adaptation occurred for all six noises (mean = 1.3 dB) and was not statistically different across noises. For the tested noises, the amount of adaptation was independent of the spectrum and level fluctuations of the noise. The results suggest that adaptation in speech recognition can occur in realistic noisy environments.

Auditory brainstem response (ABR) interpretation in clinical practice often relies on visual inspection by audiologists, which is prone to inter-practitioner variability. While deep learning (DL) algorithms have shown promise in objectifying ABR detection in controlled settings, their applicability to real-world clinical data is hindered by small datasets and insufficient heterogeneity. This study evaluates the generalizability of nine DL models for ABR detection using large, multicenter datasets. The primary dataset analyzed, Clinical Dataset I, comprises 128,123 labeled ABRs from 13,813 participants across a wide range of ages and hearing levels, and was divided into a training set (90%) and a held-out test set (10%). The models included convolutional neural networks (CNNs; AlexNet, VGG, ResNet), transformer-based architectures (Transformer, Patch Time Series Transformer [PatchTST], Differential Transformer, and Differential PatchTST), and hybrid CNN-transformer models (ResTransformer, ResPatchTST). Performance was assessed on the held-out test set and four external datasets (Clinical II, Southampton, PhysioNet, Mendeley) using accuracy and area under the receiver operating characteristic curve (AUC). ResPatchTST achieved the highest performance on the held-out test set (accuracy: 91.90%, AUC: 0.976). Transformer-based models, particularly PatchTST, showed superior generalization to external datasets, maintaining robust accuracy across diverse clinical settings. Additional experiments highlighted the critical role of dataset size and diversity in enhancing model robustness. We also observed that incorporating acquisition parameters and demographic features as auxiliary inputs yielded performance gains in cross-center generalization. These findings underscore the potential of DL models-especially transformer-based architectures-for accurate and generalizable ABR detection, and highlight the necessity of large, diverse datasets in developing clinically reliable systems.

Previous studies have demonstrated the feasibility of estimating the speech reception threshold (SRT) based on electroencephalography (EEG), termed SRT_neuro, in younger normal-hearing (YNH) participants. This method may support speech perception in hearing-aid users through continuous adaptation of noise-reduction algorithms. The prevalence of hearing impairment and thereby hearing-aid use increases with age. The SRT_neuro estimation is based on envelope reconstruction accuracy, which has also been shown to increase with age, possibly due to excitatory/inhibitory imbalance or recruitment of additional cortical regions. This could affect the estimated SRT_neuro. This study investigated the age-related changes in the temporal response function (TRF) and the feasibility of SRT_neuro estimation across age. Twenty YNH and 22 older normal-hearing (ONH) participants listened to audiobook excerpts at various signal-to-noise ratios (SNRs) while EEG was recorded using 66 scalp electrodes and 12 in-ear-EEG electrodes. A linear decoder reconstructed the speech envelope, and the Pearson's correlation was calculated between the reconstructed and speech-stimulus envelopes. A sigmoid function was fitted to the reconstruction-accuracy-versus-SNR data points, and the midpoint was used as the estimated SRT_neuro. The results show that the SRT_neuro can be estimated with similar precision in both age groups, whether using all scalp electrodes or only those in and around the ear. This consistency across age groups was observed despite physiological differences, with the ONH participants showing higher reconstruction accuracies and greater TRF amplitudes. Overall, these findings demonstrate the robustness of the SRT_neuro method in older individuals and highlight its potential for applications in age-related hearing loss and hearing-aid technology.

Understanding speech in noise is a common challenge for older adults, often requiring increased listening effort that can deplete cognitive resources and impair higher-order functions. Hearing aids are the gold standard intervention for hearing loss, but cost and accessibility barriers have driven interest in alternatives such as Personal Sound Amplification Products (PSAPs). While PSAPs are not medical devices, they may help reduce listening effort in certain contexts, though supporting evidence remains limited. This study examined the short-term effects of bilateral PSAP use on listening effort using self-report measures and electroencephalography (EEG) recordings of alpha-band activity (8-12 Hz) in older adults with and without hearing loss. Twenty-five participants aged 60 to 87 years completed a hearing assessment and a phonological discrimination task under three signal-to-noise ratio (SNR) conditions during two counterbalanced sessions (unaided and aided). Results showed that PSAPs significantly reduced self-reported effort. Alpha activity in the left parietotemporal regions showed event-related desynchronization (ERD) during the task, reflecting brain engagement in response to speech in noise. In the unaided condition, alpha ERD weakened as SNR decreased, with activity approaching baseline. PSAP use moderated this effect, maintaining stronger ERD under the most challenging SNR condition. Reduced alpha ERD was associated with greater self-reported effort, suggesting neural and subjective measures reflect related dimensions of listening demand. These results suggest that even brief PSAP use can reduce perceived and neural markers of listening effort. While not a replacement for hearing aids, PSAPs may offer a means for easing cognitive load during effortful listening. ClinicalTrials.gov, NCT05076045, https://clinicaltrials.gov/study/NCT05076045.

The long-term stability of neural responses to cochlear implant (CI) stimulation and programmed stimulation levels remains unclear. Although smaller cohort studies suggest stabilization within months postimplant, reprogramming still consumes significant clinical time. The aim of this study was to investigate the resilience of the auditory nerve to prolonged stimulation from CIs and identify changes in the clinically provided stimulation levels over time. Stimulation parameters (n = 14,072 MAPs), electrophysiological auditory nerve thresholds (n = 23,215), and slopes of amplitude growth functions (n = 17,849) were obtained from 664 bilaterally implanted children (n = 1,291 devices) followed between September 2003 and July 2022. Stimulation parameters stabilized within 12 months following implantation for most, but not all, devices (75.3% and 75.4% of devices for C-levels and T-levels, respectively). Electrophysiological measures demonstrated very minor changes per year postimplant (slopes: mean [SE] = 0.03 [0.002] μV/CU/year [95% CI: 0.02-0.03]; thresholds: mean [SE] = 0.35 [0.06] CU/year [95% CI: 0.24-0.47]). While age at implantation did not relate to clinically meaningful changes in electrophysiological measures (slopes: mean [SE] = 0.02 [0.002] μV/CU/year [95% CI: 0.01-0.02]; thresholds: mean [SE] = 0.07 [0.08] CU/year [95% CI: -0.08 to 0.23]), stimulation levels decreased for children implanted at older ages (T-levels before plateau: mean [SE] = -0.47 [0.03] CU/year [95% CI: -0.53 to -0.42]; C-levels before plateau: mean [SE] = -0.78 [0.03] CU/year [95% CI: -0.85 to -0.72]). These findings indicate long-term neural and CI programming stability, suggesting potential for directing clinical time to care in areas other than reprogramming after the first year of implant use.

Speech-in-noise testing is a valuable component of audiological examination that can provide estimates of a listener's ability to communicate in their everyday life. It has long been recognized, however, that the acoustics of real-world environments are complex and variable and not well represented by a typical clinical test setup. The first aim of this study was to quantify real-world environments in terms of several acoustic parameters that may be relevant for speech understanding (namely speech-likeness, interaural coherence, and interaural time and level differences). Earlier acoustic analyses of binaural recordings in natural environments were extended to binaural re-creations of natural environments that included conversational speech embedded in recorded backgrounds and allowed a systematic manipulation of signal-to-noise ratio. The second aim of the study was to examine these same parameters in typical clinical speech-in-noise tests and consider the "acoustic realism" of such tests. We confirmed that the parameter spaces of natural environments are poorly covered by those of the most commonly used clinical test with one frontal loudspeaker. We also demonstrated that a simple variation of the clinical test, which uses two spatially separated loudspeakers to present speech and noise, leads to better coverage of the parameter spaces of natural environments. Overall, the results provide a framework for characterizing different listening environments that may guide future efforts to increase the real-world relevance of clinical speech-in-noise testing.

The combination of directional microphones and noise reduction (DIR + NR) in hearing aids offers substantial improvement in speech intelligibility and reduction in listening effort in spatial acoustic scenarios. Pupil dilation can be used to infer ocular markers of listening effort. However, pupillometry is also known to crucially depend on luminance. The present study investigates the effects of a state-of-the-art DIR + NR algorithm (implemented in commercial hearing aids) on pupil dilation of hearing aid users both in darkness and ambient light conditions. Speech intelligibility and peak pupil dilations (PPDs) of 29 experienced hearing aid users were measured during a spatial speech-in-noise-task at a signal-to-noise ratio (SNR) matching the individual's speech reception threshold. While speech intelligibility improvements due to DIR + NR were substantial (about 35 percentage points) and independent of luminance, PPDs were only significantly reduced due to DIR + NR in ambient light, but not in darkness. This finding suggests that the reduction in PPD due to DIR + NR (most likely through improvement in SNR) is dependent on luminance and should be interpreted with caution as a marker for listening effort. Relations of reduction in PPD due to DIR + NR in ambient light to subjectively reported long-term fatigue, age, and pure-tone average were not statistically significant, which indicates that all patients benefitted similarly in listening effort from DIR + NR, irrespective of these patient-specific factors. In conclusion, careful control of luminance needs to be taken in hearing aid studies inferring listening effort from pupillometry data.