The effect of hearing loss and hearing-aid use on cue-weighting strategies for phoneme categorisation in older adult listeners

Background Hearing loss (HL) changes the way people hear sounds, such as speech, and this may change how listeners use different aspect of the acoustic information conveyed by these sounds. Previous studies have suggested that people with HL use different cue-weighting strategies for categorising phonemes. Some of these strategies may be altered by the use and/experience with hearing aids (HAs). Using the specific consonant-vowel continuum (/da/ to /ga/), the current study aims to investigate how cue-weighting strategies are influenced by HA use, and how the strategies change in the presence of noise. In addition, we will also explore how individual cue-weighting strategies are related to individual differences in self-reported hearing outcomes. Methods Three groups of native English-speaking adults – older adults with bilateral HL that do or do not use HAs, as well as normal-hearing (NH) adults – will complete a phoneme categorisation task of a two-dimensional consonant-vowel continuum between /da/ and /ga/, varying either the consonant burst or the vowel formant transition, or both. Stimuli will be created from natural speech recordings by systematically varying the energy and duration of the consonant burst, and/or the frequency trajectories of the second and the third vowel formants. The task will be performed either in quiet or in speech-shaped noise. Participants with HL will be tested both with and without amplification prescribed according to the NAL-NL2 formula. Unaided self-reported hearing outcomes will be assessed by the short form of the Speech, Spatial and Qualities of Hearing scale (SSQ12). Results Phoneme categorization judgements will be analysed in terms of the relative decisional weights of burst and formant cues, respectively. We expect that, without amplification, both HL groups will show a greater weighting of burst cues than the NH group, as HL has been shown to reduce sensitivity to formant cues. Moreover, as the burst should be more easily masked by noise, weighting of formant cues is expected to increase in noise for all listeners. Finally, amplification is expected to increase audibility, and thus weighting, of formant cues in the HL groups. This effect may be greater in the HA users than non-users, as reliance on burst cues should be less entrenched in HA users. Acknowledgement The study is funded by the Launchpad grant from the Medical Research Foundation. We would also like to thank Sven Mattys and the University of York for their support in stimuli recording.