Doctoral Dissertation Research: Speech perception in younger and older adults: Cue-weighting and contextual information

University Of Kansas Center For Research Inc, Lawrence KS

Investigators

Abstract

The way listeners process speech sounds changes across the lifespan. A better understanding of the aging effect in its early stage (55-65 years old) contributes to the advancement of speech technology (e.g., hearing aids, voice user interfaces) to better fulfill the needs of an aging population. This dissertation investigates the influence of aging on speech perception by comparing speech perception in older vs. younger adults. The experiments investigate listeners’ use of the acoustic signal and contextual information, and test how listeners adapt to ambiguity in speech sounds. Listeners integrate multiple acoustic cues when identifying speech sounds. When the primary acoustic cue is ambiguous, listeners may rely more heavily on a secondary cue to ensure speech understanding. As older listeners experience reduced temporal processing abilities, it is possible that they compensate by weighing non-temporal acoustic cues more heavily in speech perception. Experiment 1 tests whether the age-related cue-weighting shift found previously in the perception of the difference between /b/ and /p/ extends to the /d/ vs. /t/ distinction. Listeners may also adjust their cue weighting depending on the nature of the input. Experiment 2 tests whether younger and older listeners use different cue-weighting adjustment strategies. Additionally, listeners may access stored information about words to help resolve ambiguities in the acoustic signal. Older listeners were previously found to use this lexical information to a greater degree than younger listeners. Experiment 3 explores whether younger and older adults differ in their use of lexical information when learning to interpret utterances that are ambiguous. A final goal of the dissertation is to test whether younger and older adults differ in how they handle changes in speaking rate. In Experiment 4, the effect of speaking rate is examined by embedding /d/ vs. /t/ contrasts in slow and fast speech contexts and presenting these syllables to younger and older listeners to compare their performance in identifying the sounds. The results of the four experiments contribute to our understanding of the human aging process. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.