A Method for Measuring Speech Intelligibility Using Connected Speech

Area of Science:

• Audiology
• Speech-Language Pathology
• Psychoacoustics

Background:

• Speech intelligibility testing is crucial for evaluating hearing aid efficacy and understanding speech perception.
• Existing methods often use isolated sentences, which may not reflect real-world listening conditions.
• Connected speech offers a more ecologically valid paradigm for assessing auditory comprehension.

Purpose of the Study:

• To present and evaluate an experimental method for measuring speech intelligibility using connected speech.
• To assess the performance of this method in both normal-hearing (NH) and hearing-impaired (HI) listeners.
• To compare speech intelligibility in connected speech versus non-connected speech contexts.

Main Methods:

• Short target sentences were embedded within connected speech from a same-sex talker.
• An interfering signal of connected speech from a different same-sex talker was used.
• The method was tested on NH and HI listeners, comparing intelligibility scores between embedded and non-embedded sentences.

Main Results:

• No significant average difference in speech intelligibility was found between the two presentation modes for NH or HI listeners.
• A large degree of inter-individual variability in performance was observed, particularly among HI listeners.
• This variability suggests potential for individualized benefits depending on the presentation mode.

Conclusions:

• The proposed method provides a viable tool for speech intelligibility testing with connected speech.
• This approach allows for the assessment of signal characteristics and processing algorithms over longer time windows.
• It offers a novel experimental approach to studying speech perception in complex auditory environments.