Abstract
Objective: The main objective was to investigate the effect of linguistic abilities (lexical-access ability and vocabulary size) on different measures of speech-in-noise recognition in normal-hearing listeners with various levels of language proficiency. Design: Speech reception thresholds (SRTs) were measured for sentences in steady-state (SRTstat) and fluctuating noise (SRTfluc), and for digit-triplets in steady-state noise (DIN). Lexical-access ability was measured with a lexical-decision test and a word-naming test. Vocabulary size was also measured. For the SRT, keyword scoring and sentence scoring were compared. Study sample: To introduce variation in linguistic abilities, three groups of 24 young normal-hearing listeners were included: higher-educated native, lower-educated native, and higher-educated non-native listeners. Results: Lexical-access ability was most accurately measured with combined results of lexical decision and word naming. Lexical-access ability explained 60% of the variance in SRT. The effect of linguistic abilities on SRTs was up to 5.6 dB for SRTstat and 8 dB for SRTfluc. Using keyword scoring reduced this effect by approximately 1.5 dB. For DIN the effect of linguistic ability was less than 1 dB. Conclusions: Lexical-access ability is an important predictor of SRTs in normal-hearing listeners. These results are important to consider in the interpretation of speech-in-noise scores of hearing-impaired listeners.
Acknowledgements
The authors thank all participants who participated in the study, Hans van Beek for technical support, Jaap Peter Kronemeijer, Katrien Devroye, and Sietske Temminghoff for data collection. We also thank Elske Bolk and Andriana Zekveld for their help in defining the keywords, and Birgit Witte for her statistical guidance.
Declaration of interests
The authors report no conflicts of interest.
Notes
1 Notes
Two participants from the non-native group had higher thresholds at one or two frequencies for only one ear. For one participant it was 40 dB HL at 4000 Hz, and for the other participant it was 25 and 30 dB HL at 2000 and 4000 Hz. As all stimuli were diotic and speech-in-noise recognition results of those two participants were not extremes, we did not exclude them.
2 In the main study (Part 1) with keyword scoring, the difference in SRTstat was significantly different between HE-N and LE-N. In this part we measured SRTs in a fixed order, reducing variance due to differences in list content and fatigue.