Effects of mild age-related hearing loss and background noise on speech communication

Our ability to communicate successfully with others can be affected by the presence of noise in the environment and becomes increasingly difficult with advancing age. Much research considers the effect of aging on the ability to understand speech but there is less attention on speech production and the adaptations talkers make to overcome adverse listening conditions, especially in more ecologically valid situations such as during interactive speech. Our aim was to investigate what speech modifications talkers make in background noise and which parameters predict communication efficiency, and if these are influenced by the hearing status of older talkers.

Our study included 83 talkers of Southern British English: 57 older adults aged 65-84, of which 30 had normal hearing (OANH) and 27 (OAHL) had mild age-related hearing loss but did not wear hearing aids (mean PTA .250-4kHz: 27.7 dB, better ear), and 26 younger adults (YA) aged 18-26. Talkers were recorded while they completed an interactive ‘spot the difference’ picture task (diapix) with a younger conversational partner when i) they could each other normally (NORM); ii) they were both in multi-talker babble noise (BAB2). We collected background sensory and cognitive measures (speech-in-noise thresholds, working memory, verbal fluency). From the diapix recordings, we measured acoustic-phonetic features (articulation rate, relative energy in the 1-3 kHz region [ME13], f0, vowel hyperarticulation) and for a subset of talkers (YA=24, OANH=21, OAHL=21) dysfluency rates (filled pauses, false starts, repetitions). Communication efficiency was measured as time to find eight differences between the pictures.

The results showed that OAHL talkers had poorer speech-in-noise thresholds, lower verbal fluency scores and it took them longer to find the eight differences in the diapix tasks (NORM, BAB2) than for YA or OANH talkers. OAHL talkers were also more dysfluent than YA and OANH talkers. Background noise induced slower speech and modifications that resemble Lombard speech (i.e., increase in f0 and ME13) that were greater for OA than for YA talkers. However, when speaking in noise, OAHL talkers made adaptations more consistent with an increase in vocal effort. Only speaking rate, ME13 and frequency of dysfluent repetitions predicted how effectively the two talkers found the differences between the pictures.

Our results suggest that even mild levels of age-related hearing loss can affect communication efficiency and fluency regardless of listening condition. However, background noise can influence speaking effort in OAHL talkers.