Prosodic Focus Interpretation in Spectrotemporally Degraded Speech by Non-Native Listeners

Sep 13, 2023·
Marita K Everhardt
,
Anastasios Sarampalis
Matt Coler
Matt Coler
,
Deniz Başkent
,
Wander Lowie
· 2 min read
Abstract
Purpose: This study assesses how spectrotemporal degradations that can occur in the sound transmission of a cochlear implant (CI) may influence the ability of non-native listeners to recognize the intended meaning of utterances based on the position of the prosodically focused word. Previous research suggests that perceptual accuracy and listening effort are negatively affected by CI processing (or CI simulations) or when the speech is presented in a non-native language, in a number of tasks and circumstances. How these two factors interact to affect prosodic focus interpretation, however, remains unclear. Method: In an online experiment, normal-hearing (NH) adolescent and adult native Dutch learners of English and a small control group of NH native English adolescents listened to CI-simulated (eight-channel noise-band vocoded) and non–CI-simulated English sentences differing in prosodically marked focus position. Participants indicated which alternative was being corrected in each sentence. Results: Both native English and Dutch groups performed worse with CI simulations. While the Dutch group performed comparably to the English group on non–CI-simulated speech, they performed markedly worse on CI-simulated speech. Self-reported English proficiency was a significant predictor of performance for the Dutch group. Conclusions: The results suggest that the challenges of spectrotemporally degraded speech and non-native listening are not simply additive but interact in ways that disproportionately affect prosodic focus interpretation in non-native listeners, especially those with lower language proficiency.
Type
Publication
Journal of Speech, Language, and Hearing Research, 66(9), 3649-3664

This study investigates the combined effects of spectrotemporal degradation (similar to cochlear implant processing) and non-native language processing on the ability to interpret prosodic focus in speech. Prosodic focus—the emphasis placed on certain words to signal their importance in an utterance—is critical for understanding a speaker’s intended meaning, especially in contexts where focus changes the communicative intent of otherwise identical sentences.

Through an online experiment comparing native English listeners with Dutch learners of English, we found that:

  1. Both groups showed decreased performance when listening to CI-simulated speech compared to unprocessed speech
  2. While Dutch listeners performed similarly to native English listeners with unprocessed speech, they showed disproportionately poorer performance with CI-simulated speech
  3. Language proficiency was a significant predictor of performance for non-native listeners

These findings suggest that the challenges of processing degraded speech and non-native language are not simply additive but interact in complex ways. This has important implications for cochlear implant users who must communicate in non-native languages, suggesting they may face unique challenges in interpreting prosodic cues that go beyond the difficulties experienced by either CI users listening in their native language or non-native listeners with normal hearing.

The research contributes to our understanding of the factors affecting speech perception in challenging conditions and may inform the development of more effective rehabilitation strategies for cochlear implant users who operate in multilingual environments.

Cochlear Implants Speech Perception Prosody Non-Native Listening Focus Interpretation Clinical Applications
Matt Coler
Authors
Matt Coler
Associate Professor of Speech Technology