Talker variability in Mandarin spoken word recognition: Comparing normal and hearing-impaired listeners

Abstract

During natural speech processing, talker variability is inherently embedded in the acoustic signal and may affect the speed and accuracy of processing. Prior studies have demonstrated the effect of talker variability in normal listeners. What is the effect of inter-talker variability in people with hearing impairment? This study aims to explore the question in native speakers of Mandarin Chinese, trying to compare listeners with and without hearing impairment. Participants are instructed to listen to pre-recorded pairs of Mandarin Chinese words (e.g., in English translation, happy – sad) and make responses only to the second item of each pair (e.g., sad) in a short-term priming paradigm. The speed and accuracy of the responses are measured. The fact that the pairs were spoken by either the same or different native speakers of Mandarin Chinese allows us to examine the effect of speaker voice change on the reaction time and accuracy. The hearing-impaired listeners appear to be facilitated more by the same speaker voice condition in terms of response accuracy, compared with the listeners without hearing impairment. There also seems to be more semantic facilitation to the hearing-impaired in terms of reaction time, although this effect does not seem to be modulated by talker variability.