Reproducibility in Cognitive Hearing Research: Theoretical Considerations and Their Practical Application in Multi-Lab Studies

Identifying Links Between Latent Memory and Speech Recognition Factors

OBJECTIVES

The link between memory ability and speech recognition accuracy is often examined by correlating summary measures of performance across various tasks, but interpretation of such correlations critically depends on assumptions about how these measures map onto underlying factors of interest. The present work presents an alternative approach, wherein latent factor models are fit to trial-level data from multiple tasks to directly test hypotheses about the underlying structure of memory and the extent to which latent memory factors are associated with individual differences in speech recognition accuracy. Latent factor models with different numbers of factors were fit to the data and compared to one another to select the structures which best explained vocoded sentence recognition in a two-talker masker across a range of target-to-masker ratios, performance on three memory tasks, and the link between sentence recognition and memory.

DESIGN

Young adults with normal hearing (N = 52 for the memory tasks, of which 21 participants also completed the sentence recognition task) completed three memory tasks and one sentence recognition task: reading span, auditory digit span, visual free recall of words, and recognition of 16-channel vocoded Perceptually Robust English Sentence Test Open-set sentences in the presence of a two-talker masker at target-to-masker ratios between +10 and 0 dB. Correlations between summary measures of memory task performance and sentence recognition accuracy were calculated for comparison to prior work, and latent factor models were fit to trial-level data and compared against one another to identify the number of latent factors which best explains the data. Models with one or two latent factors were fit to the sentence recognition data and models with one, two, or three latent factors were fit to the memory task data. Based on findings with these models, full models that linked one speech factor to one, two, or three memory factors were fit to the full data set. Models were compared via Expected Log pointwise Predictive Density and post hoc inspection of model parameters.

RESULTS

Summary measures were positively correlated across memory tasks and sentence recognition. Latent factor models revealed that sentence recognition accuracy was best explained by a single factor that varied across participants. Memory task performance was best explained by two latent factors, of which one was generally associated with performance on all three tasks and the other was specific to digit span recall accuracy at lists of six digits or more. When these models were combined, the general memory factor was closely related to the sentence recognition factor, whereas the factor specific to digit span had no apparent association with sentence recognition.

CONCLUSIONS

Comparison of latent factor models enables testing hypotheses about the underlying structure linking cognition and speech recognition. This approach showed that multiple memory tasks assess a common latent factor that is related to individual differences in sentence recognition, although performance on some tasks was associated with multiple factors. Thus, while these tasks provide some convergent assessment of common latent factors, caution is needed when interpreting what they tell us about speech recognition.

Streamlining experiment design in cognitive hearing science using OpenSesame

Auditory science increasingly builds on concepts and testing paradigms originated in behavioral psychology and cognitive neuroscience - an evolution of which the resulting discipline is now known as cognitive hearing science. Experimental cognitive hearing science paradigms call for hybrid cognitive and psychobehavioral tests such as those relating the attentional system, working memory, and executive functioning to low-level auditory acuity or speech intelligibility. Building complex multi-stimuli experiments can rapidly become time-consuming and error-prone. Platform-based experiment design can help streamline the implementation of cognitive hearing science experimental paradigms, promote the standardization of experiment design practices, and ensure reliability and control. Here, we introduce a set of features for the open-source python-based OpenSesame platform that allows the rapid implementation of custom behavioral and cognitive hearing science tests, including complex multichannel audio stimuli while interfacing with various synchronous inputs/outputs. Our integration includes advanced audio playback capabilities with multiple loudspeakers, an adaptive procedure, compatibility with standard I/Os and their synchronization through implementation of the Lab Streaming Layer protocol. We exemplify the capabilities of this extended OpenSesame platform with an implementation of the three-alternative forced choice amplitude modulation detection test and discuss reliability and performance. The new features are available free of charge from GitHub: https://github.com/elus-om/BRM_OMEXP .

Clinical Trials and Outcome Measures in Adults With Hearing Loss

Clinical trials are designed to evaluate interventions that prevent, diagnose or treat a health condition and provide the evidence base for improving practice in health care. Many health professionals, including those working within or allied to hearing health, are expected to conduct or contribute to clinical trials. Recent systematic reviews of clinical trials reveal a dearth of high quality evidence in almost all areas of hearing health practice. By providing an overview of important steps and considerations concerning the design, analysis and conduct of trials, this article aims to give guidance to hearing health professionals about the key elements that define the quality of a trial. The article starts out by situating clinical trials within the greater scope of clinical evidence, then discusses the elements of a PICO-style research question. Subsequently, various methodological considerations are discussed including design, randomization, blinding, and outcome measures. Because the literature on outcome measures within hearing health is as confusing as it is voluminous, particular focus is given to discussing how hearing-related outcome measures affect clinical trials. This focus encompasses how the choice of measurement instrument(s) affects interpretation, how the accuracy of a measure can be estimated, how this affects the interpretation of results, and if differences are statistically, perceptually and/or clinically meaningful to the target population, people with hearing loss.

Cogmed Training Does Not Generalize to Real-World Benefits for Adult Hearing Aid Users: Results of a Blinded, Active-Controlled Randomized Trial

Objectives:

Performance on working memory tasks is positively associated with speech-in-noise perception performance, particularly where auditory inputs are degraded. It is suggested that interventions designed to improve working memory capacity may improve domain-general working memory performance for people with hearing loss, to benefit their real-world listening. We examined whether a 5-week training program that primarily targets the storage component of working memory (Cogmed RM, adaptive) could improve cognition, speech-in-noise perception and self-reported hearing in a randomized controlled trial of adult hearing aid users with mild to moderate hearing loss, compared with an active control (Cogmed RM, nonadaptive) group of adults from the same population.

Design:

A preregistered randomized controlled trial of 57 adult hearing aid users (n = 27 experimental, n = 30 active control), recruited from a dedicated database of research volunteers, examined on-task learning and generalized improvements in measures of trained and untrained cognition, untrained speech-in-noise perception and self-reported hearing abilities, pre- to post-training. Participants and the outcome assessor were both blinded to intervention allocation. Retention of training-related improvements was examined at a 6-month follow-up assessment.

Results:

Per-protocol analyses showed improvements in trained tasks (Cogmed Index Improvement) that transferred to improvements in a trained working memory task tested outside of the training software (Backward Digit Span) and a small improvement in self-reported hearing ability (Glasgow Hearing Aid Benefit Profile, Initial Disability subscale). Both of these improvements were maintained 6-month post-training. There was no transfer of learning shown to untrained measures of cognition (working memory or attention), speech-in-noise perception, or self-reported hearing in everyday life. An assessment of individual differences showed that participants with better baseline working memory performance achieved greater learning on the trained tasks. Post-training performance for untrained outcomes was largely predicted by individuals’ pretraining performance on those measures.

Conclusions:

Despite significant on-task learning, generalized improvements of working memory training in this trial were limited to (a) improvements for a trained working memory task tested outside of the training software and (b) a small improvement in self-reported hearing ability for those in the experimental group, compared with active controls. We found no evidence to suggest that training which primarily targets storage aspects of working memory can result in domain-general improvements that benefit everyday communication for adult hearing aid users. These findings are consistent with a significant body of evidence showing that Cogmed training only improves performance for tasks that resemble Cogmed training. Future research should focus on the benefits of interventions that enhance cognition in the context in which it is employed within everyday communication, such as training that targets dynamic aspects of cognitive control important for successful speech-in-noise perception.

Low-sound-level auditory processing in noise-exposed adults

Early signs of noise-induced hearing damage are difficult to identify, as they are often confounded by factors such as age, audiometric thresholds, or even music experience. Much previous research has focused on deficits observed at high intensity levels. In contrast, the present study was designed to test the hypothesis that noise exposure causes a degradation in low-sound-level auditory processing in humans, as a consequence of dysfunction of the inner hair cell pathway. Frequency difference limens (FDLs) and amplitude modulation depth discrimination (MDD) were measured for five center frequencies (0.75, 1, 3, 4 and 6 kHz) at 15 and 25 dB sensation level (SL), as a function of noise exposure, age, audiometric hearing loss, and music experience. Forty participants, aged 33-75 years, with normal hearing up to 1 kHz and mild-to-moderate hearing loss above 2 kHz, were tested. Participants had varying degrees of self-reported noise exposure, and varied in music experience. FDL worsened as a function of age. Participants with music experience outperformed the non-experienced in both the FDL and MDD tasks. MDD thresholds were significantly better for high-noise-exposed, than for low-noise-exposed, participants at 25 dB SL, particularly at 6 kHz. No effects of age or hearing loss were observed in the MDD. It is possible that the association between MDD thresholds and noise exposure was not causal, but instead was mediated by other factors that were not measured in the study. The association is consistent, qualitatively, with a hypothesized loss of compression due to outer hair cell dysfunction.