RP304 - Bridging the Sound Booth and the Real World: Laying Groundwork to Improve Predictions of Listening Performance in Real-World Noisy Environments

Abstract:
A goal for many and scientists and clinicians is to improve the ecological validity of audiologic measurements. Rapid improvements in technology have enabled listening performance and intervention outcome to be measured in ecologically relevant conditions using realistic virtual acoustic scenes. We need to evaluate the reliability and validity of using these acoustic scenes, investigate whether these tests actually improve predictions of listening performance in the real world, and determine whether the types of acoustic scenes used represent the noisy environments people with hearing loss encounter in daily life. We present on two recent studies aiming to address these gaps.

Summary:
Hearing technologies that show benefit in controlled laboratory tests do not provide the same benefit in real-world noisy environments. Researchers are increasingly using virtual acoustic scenes to assess listening performance in realistic conditions and collecting data about listening performance in daily life. We need to evaluate the reliability and validity of using realistic acoustic scenes to measure listening performance, investigate whether these tests improve predictions of listening performance in the real world, and determine whether the acoustic scenes used represent the complex environments listeners with hearing loss actually encounter in daily life.

Methods:
We investigated the reliability and validity of using virtual acoustic scenes to evaluate listening performance. Using available, open-source materials, realistic acoustic scenes were rendered in the lab. Speech perception (SNR50s) and listening effort (vocal reaction times) were measured twice in cafe, church, dinner party, and food court environments, as well as in traditional speech-shaped noise, among 53 young adults with typical hearing. Participants then completed a week of ecological momentary assessment, where they reported on the listening environments they encountered in daily life and rated their speech perception and listening effort. We measured the test-retest reliability of speech perception and listening effort in real-world noise, compared listening performance across environments, and compared real-world noise to traditional speech-shaped noise. We then compared performance measured in the lab environments to ratings in matched real-world environments. For example, speech perception scores measured in the cafe environment in the lab were compared to speech perception ratings in similar cafe environments in the real world. Then in a small pilot study, we asked older adults with hearing loss to complete ecological momentary assessments and audio recordings in the complex listening environments they encountered in daily life. Using these data, we determined the types of noisy environments commonly encountered by older hearing aid users.

Results &

Conclusions:
Differences across tests within the same environment were small, though the intraclass correlation coefficients were larger for the environments with greater variance in their complexity, highlighting but also potential utility in measuring performance in real-world noise. Performance differed across all environments, with scores varying with sound level. Ratings of speech perception and listening effort on ecological momentary assessment in the matched environments in the real world followed the same overall pattern of performance measured in the lab, but individual scores in the lab did not predict real-world ratings, suggesting that the acoustic environment alone cannot account for real-world listening performance. The pilot study of older adults suggests that while research often focuses on cocktail party environments, most complex listening environments were encountered at home, where older adults frequently listen to speech in a background noise of machines, media, or another talker.

This presentation is Intermediate instructional level. Familiarity with measures of listening performance in noise, some knowledge of noise acoustics, and an awareness of in-situ data collection methods are assumed. However, basic concepts will be reviewed, and the focus of the results is on clinical application and future directions of translational research.

Brief Summary of Clinical Takeaways: The clinical takeaway for this session is that the ecological validity and clinical utility speech-in-noise testing can be improved using realistic virtual noisy environments, but the interpretation and use of results need to be considered carefully. Further, the type of noise is only one contributor to real-world speech perception.