Task-Induced Mental Fatigue and Motivation Influence Listening Effort as Measured by the Pupil Dilation in a Speech-in-Noise Task

Objectives: Listening effort and fatigue are common complaints among individuals with hearing impairment (HI); however, the underlying mechanisms, and relationships between listening effort and fatigue are not well understood. Recent quantitative research suggests that the peak pupil dilation (PPD), which is commonly measured concurrent to the performance of a speech-in-noise task as an index of listening effort, may be informative of daily-life fatigue, but it remains unknown whether the same is true for task-induce fatigue. As fatigue effects are known to manifest differently depending on motivation, the main aim of the present study was to experimentally investigate the interactive effects of task-induced fatigue and motivation on the PPD. Design: In a pre-/post- fatigue within-subject design, 18 participants with normal hearing (NH) engaged in a 98-trial-long speech-in-noise task (i.e., a load sequence, approximately 40 min. long), which either excluded or included additional memory demands (light vs. heavy load sequence). Before and after the load sequence, baseline pupil diameter (BPD) and PPD were measured during shorter probe blocks of speech-in-noise tasks. In these probe blocks, if participants correctly repeated more than 60% of the keywords, they could win vouchers of either 20 or 160 Danish krones worth (low incentive vs. high incentive). After each probe block, participants reported their invested effort, tendency for quitting, and perceived performance. Results: The BPD in anticipation of listening declined from pre- to post-load sequence, suggesting an overall decrease in arousal, but the decline did not scale with the magnitude of the load sequence, nor with the amount of monetary incentive. Overall, there was larger pre- to post-load sequence decline in PPD when the load sequence was heavy and when the monetary incentives were low. Post-hoc analyses showed that the decline in PPD was only significant in the heavy-load sequence-low reward condition. The speech-in-noise task performance, self-reported effort, and self-reported tendency to quit listening did not change with the experimental conditions. Conclusions This is the first study to investigate the influence of task-induced fatigue on BPD and PPD. Whereas BPD was not sensitive to the magnitude of previous load sequence and monetary incentives, the decline in PPD from pre- to post- load sequence was significant after the heavy load sequence when the offered monetary incentives were low. This result supports the understanding that fatigue and motivation interactively influence listening effort.

Speech-Driven Facial Animations Improve Speech-in-Noise Comprehension of Humans

Understanding speech becomes a demanding task when the environment is noisy. Comprehension of speech in noise can be substantially improved by looking at the speaker’s face, and this audiovisual benefit is even more pronounced in people with hearing impairment. Recent advances in AI have allowed to synthesize photorealistic talking faces from a speech recording and a still image of a person’s face in an end-to-end manner. However, it has remained unknown whether such facial animations improve speech-in-noise comprehension. Here we consider facial animations produced by a recently introduced generative adversarial network (GAN), and show that humans cannot distinguish between the synthesized and the natural videos. Importantly, we then show that the end-to-end synthesized videos significantly aid humans in understanding speech in noise, although the natural facial motions yield a yet higher audiovisual benefit. We further find that an audiovisual speech recognizer (AVSR) benefits from the synthesized facial animations as well. Our results suggest that synthesizing facial motions from speech can be used to aid speech comprehension in difficult listening environments.

Development of a Training Software to Improve Speech- in-Noise Perception in the Elderly with Noise-Induced Hearing Loss

Background and Aim: The incidence of noise-induced hearing loss (NIHL) is increasing rapidly worldwide. It has been shown that the long-term exposure to noise leads to permanent hearing loss. There is currently no treatment for NIHL and it is mainly managed by preventive measures. This study aimed to develop a training software to improve speech-in-noise (SIN) perception in the elderly suffering from mild-to-moderate NIHL due to temporal fine structure (TFS) damage. Methods: This is a non-randomized clinical trial conducted on 8 older men aged 60-75 years (Mean age = 68 ± 4.5 years) with mild-to-moderate NIHL at high frequencies (43.75±6.0 dBHL) with at least for five years of work in noisy environment. They received rehabilitation using a researcher-developed training software targeted TFS for five weeks. To assess the efficiency of the training software, the signal-to-noise ratio for 50% correct scores (SNR-50%) was calculated using the word-in-noise test before and after the trial. Results: The SNR-50% improved significantly in all participants from 13±2.63 dB to 6.10±2.85 dB (p<0.001). In the multiple linear regression model, the rehabilitation by the training software predicted 68% of improvement in SNR-50% (coefficient of determination=0.676). Conclusion: Rehabilitation by the training software developed based on TFS can improve SIN. perception in the elderly with NIHL.

Development and Psychometric Evaluation of Persian Dynamic Spatial Quick Speech-in-Noise Test in Adults with Normal Hearing

Background and Aim: Spatial hearing is a prerequisite for the proper function of the listener in complex auditory environments. In the present study, a Persian version of the dynamic spatial-quick speech in noise (DS-QSIN) has been developed with respect to all possible factors affecting the test and to run five lists for normal hearing subjects and assessment of reliability. Methods: To construct five new lists according to the original quick speech in noise (QSIN) test, we used frequent, familiar, and difficult words to construct unpredictable sentences. After determining the content and face validity of the sentences, 30 selected sentences were played using a DS-QSIN software for 35 subjects aged 18–25 years. The reliability of the test was assessed after repeating the test after two weeks. Results: According to expert judges, these 30 sentences showed acceptable content and face validity with the changes. The average signal-to-noise ratio (SNR) loss of five lists was –5.2 dB. No significant difference was seen between men and women in all lists. The results indicate no difference in the average SNR loss between the five lists. Regarding the reliability assessment, the test-retest correlation coefficient was 0.5 to 0.7 (p<0.05). The intra-class correlation coefficient between test-retest was statistically significant (p>0.001) and confirmed that the lists have high reliability and repeatability. Conclusion: DS-QSIN test showed good validity and reliability and can be helpful in diagnosis and selecting the best method for rehabilitation of people with a spatial hearing disorder.

Speech-in-Noise Audiometry in Adults: A Review of the Available Tests for French Speakers

<b><i>Background:</i></b> Difficulty understanding speech in background noise is the reason of consultation for most people who seek help for their hearing. With the increased use of speech-in-noise (SpIN) testing, audiologists and otologists are expected to evidence disabilities in a greater number of patients with sensorineural hearing loss. The purpose of this study is to list validated available SpIN tests for the French-speaking population. <b><i>Summary:</i></b> A review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. PubMed and Scopus databases were searched. Search strategies used a combination of 4 keywords: speech, audiometry, noise, and French. There were 10 validated SpIN tests dedicated to the Francophone adult population at the time of the review. Some tests use digits triplets as speech stimuli and were originally designed for hearing screening. The others were given a broader range of indications covering diagnostic or research purposes, determination of functional capacities and fitness for duty, as well as assessment of hearing amplification benefit. <b><i>Key Messages:</i></b> As there is a SpIN test for almost any type of clinical or rehabilitation needs, both the accuracy and duration should be considered for choosing one or the other. In an effort to meet the needs of a rapidly aging population, fast adaptive procedures can be favored to screen large groups in order to limit the risk of ignoring the early signs of forthcoming presbycusis and to provide appropriate audiological counseling.

Speech-driven Facial Animations Improve Speech-in-Noise Comprehension of Humans

Understanding speech becomes a demanding task when the environment is noisy. Comprehension of speech in noise can be substantially improved by looking at the speake's face, and this audiovisual benefit is even more pronounced in people with hearing impairment. Recent advances in AI have allowed to synthesize photorealistic talking faces from a speech recording and a still image of a person's face in an end-to-end manner. However, it has remained unknown whether such facial animations improve speech-in-noise comprehension. Here we consider facial animations produced by a recently introduced generative adversarial network (GAN), and show that humans cannot distinguish between the synthesized and the natural videos. Importantly, we then show that the end-to-end synthesized videos significantly aid humans in understanding speech in noise, although the natural facial motions yield a yet higher audiovisual benefit. We further find that an audiovisual speech recognizer benefits from the synthesized facial animations as well. Our results suggest that synthesizing facial motions from speech can be used to aid speech comprehension in difficult listening environments.