Quantifying the contribution of contextual information in speech perception in noise for native and non-native listeners

• Published in: The Journal of the Acoustical Society of America 2008-10, Vol.124 (4_Supplement), p.2593-2593
• Main Authors: Nishi, Kanae, Lewis, Jessica, Kopun, Judy, Stelmachowicz, Patricia G.
• Format: Article
• Language: English
• ISSN: 0001-4966; 1520-8524
• DOI: 10.1121/1.4808988

When speech is degraded, listeners tend to rely on contextual information, if available. The goal of the current study was to evaluate the use of lexical, syntactic, and semantic context for native and non-native listeners. Using an adaptive tracking method, listeners’ reliance on contextual information was assessed as signal-to-noise ratios for 70%-correct speech perception performance (SNR70) for single words and short grammatically and semantically appropriate sentences. Adult and child native speakers of American English and adult Spanish speakers learning English as a non-native language served as listeners. For all groups, SNR70 was significantly higher for words than sentences. No group difference was found for sentence SNR70, but significantly higher word SNR70 was observed for the non-native group than the two native groups, indicating greater benefit of context for non-native listeners. When non-native listeners' SNR70 measures were subjected to discriminant analysis to determine similarity to native groups, three subgroups were identified: (1) nativelike for both words and sentences, (2) nativelike only for sentences, and (3) not nativelike for either words or sentences. Possible application of this method to determine candidacy for auditory training to improve speech perception in noise will be discussed. [Work supported by the NIH.]