PP321 - Enhanced Speech Recognition and Reduced Listening Effort with On-Demand Hearing Aid Processing

Abstract: While MarkeTrak 2022 indicates that more than 80% of hearing aid owners are satisfied with their devices, understanding speech in difficult listening environments remains a challenge. With the goal of helping users in difficult situations, an on-demand processing feature was designed for hearing aids to provide additional signal processing to optimize listening experiences in difficult environments. Data from two lab-based studies demonstrated that on-demand processing can effectively enhance speech recognition and reduce listening effort when listening to speech in background noise, providing an option for users to optimize their hearing aids in situations when default settings may not be sufficient.

Summary:

While the MarkeTrak 2022 survey data indicate that more than 80% of hearing aid owners are satisfied with their devices, understanding speech in difficult listening environments remains a challenge. With the goal of helping users in these difficult situations, an on-demand processing feature was designed for Starkey hearing aids to provide additional signal processing to optimize listening experiences. When activated by the user, this feature prompts the hearing aid to analyze the current listening environment and then apply specialized changes to the gain, microphone, and noise reduction settings to improve listening outcomes.

The purpose of this project was to demonstrate the effects of app-based, on-demand processing on speech recognition and listening effort. Data are from two lab-based studies. In study 1, twelve experienced hearing-aid users using occluded couplings were fitted with Starkey Genesis AI 24 RIC RT devices. They had symmetrical mild to moderately severe hearing loss. The participants completed speech recognition tasks (IEEE sentences presented in a real-life recording of restaurant noise) with and without the on-demand processing. In study 2, the participants were 20 experienced hearing-aid users with symmetrical mild to moderately severe hearing loss. They were fitted with the same hearing aids as used in study 1. They completed a listening effort task using the Adaptive Categorical Listening Effort Scaling (ACALES) in a diffuse modulated masker (English matrix test in the ICRA250 noise) with and without the on-demand processing. For both studies, speech was presented from 0-degree azimuth. In the speech recognition test, the noise level was fixed at 63 dB SPL. The participants were evaluated under their individualized signal-to-noise (SNR) to achieve 70% speech recognition without the on-demand processing. In the listening effort test, the noise level was fixed at 65 dB SPL and the speech level was varied to create the different SNRs for listening effort evaluation. Results from the two studies showed that the on-demand processing provided significant benefits for speech recognition in noise (11% improvement, t(11) = -2.55, p = 0.027) and listening effort (+1.13 dB in SNR benefit, t(19) = 2.08, p =0.05).

The on-demand feature evaluated in the present studies provides an option for hearing aid users to optimize their hearing aids in situations when they have trouble hearing with the default settings. These findings support the conclusion that the on-demand processing can provide additional speech enhancement and noise reduction to help communication in challenging situations.