15th Speech in Noise Workshop, 11-12 January 2024, Potsdam, Germany 15th Speech in Noise Workshop, 11-12 January 2024, Potsdam, Germany

P07Session 1 (Thursday 11 January 2024, 15:35-18:00)
Examining the effect of voice training on speech-on-speech intelligibility and listening effort in cochlear implant users

Ada Bicer, Thomas Koelewijn, Deniz Başkent
Department of Otorhinolaryngology/Head and Neck Surgery, University Medical Center Groningen, University of Groningen, The Netherlands

Understanding speech in multiple-talker situations can be challenging for people from all hearing backgrounds. However, achieving good speech-on-speech intelligibility is especially challenging when speech is degraded, such as in communication via a cochlear implant (CI). Additionally, understanding speech in multi-talker listening situations can be more effortful for CI users, compared to normal hearing listeners. For normal hearing listeners, listening to familiar or lab-trained voices can be more intelligible in speech-on-speech situations compared to listening to unfamiliar voices. In addition to these intelligibility benefits, voice training might also lead to a reduction in listening effort for normal hearing listeners in speech-on-speech situations. However, sometimes voice training might lead to a listening effort benefit that is not accompanied by an intelligibility benefit, as shown by previous studies with normal hearing listeners. Therefore, the question of whether voice training can lead to a benefit in improving speech intelligibility and/or reducing listening effort for CI users, remains open. The aim of this study is to examine if an explicit voice training can provide a benefit in either speech intelligibility or listening effort, or both, which might improve the quality of life of CI users.

In this study, an explicit voice training was implemented, which involved a talker identification task. During voice training, half of the participants listened to 3 female voices and the other half listened to 3 male voices. As a task, participants selected who was speaking by clicking on the talker’s name, for each sentence that was presented. In total, 600 sentences were used during voice training, while one voice was presented more frequently (400 sentences) that was later used as the trained voice. Following the voice training, speech-on-speech intelligibility was measured by means of a Dutch version of the Coordinate Response Measure (CRM) test. Listening effort was measured during the CRM test using pupillometry. The CRM test involved a target sentence that consists of a call sign, a color, and a number (e.g. “Show the dog where the blue five is”), uttered by the trained voice and an untrained voice, while an unintelligible speech masker was presented in the background, at 0 dB and +6 dB target-to-masker ratios. Results from this ongoing study will be presented.

Last modified 2023-11-29 15:13:43