Development and validation of a procedure to test perception of speech prosody in listeners with cochlear implants.

Abstract

Understanding vocal prosody is essential to successful communication. However, evaluations of speech recognition have relied heavily on word repetition-type tasks where success does not hinge on prosody perception, or where stimuli do not have enough prosodic variation to even test for this ability. Individuals who use cochlear implants (CIs) are at risk for poorer perception of prosody because of their limited access to pitch perception. This study used a multi-slider visual analog interface to measure perception of contrastive focus prosody in sentence-length stimuli by participants with CIs or with typical hearing (TH). Compared to TH listeners, CI users were more likely to misidentify which word had prosodic focus, as well as having weaker perception of prosodic focus, on average. Whereas TH listeners scaled their perceived strength of prosodic focus based on F0 and vowel intensity features, CI users scaled ratings in accordance with vowel intensity and vowel duration, with no relationship to F0. These results suggest that CI users are at risk of complete misperception of a talker's intended message, even in instances where there was no uncertainty about the words that were spoken.