Predicting Audiovisual Word Recognition in Noisy Situations: Toward Precision Audiology

Area of Science:

• Audiology
• Speech Perception
• Human Communication

Background:

• Spoken communication benefits from both auditory and visual cues.
• Previous research suggested age-related deficits in audiovisual (AV) speech perception, often attributed to integration issues.
• Conventional wisdom posits age impacts AV speech perception due to integration deficits.

Purpose of the Study:

• To test if previous findings on AV speech perception prediction generalize to open-set, real-life scenarios.
• To evaluate a new predictive model for AV speech perception.
• To investigate the role of age in AV speech perception prediction.

Main Methods:

• 109 participants (ages 22-93) were tested using an open-set Lex-List test for auditory-only (A-only), visual-only (V-only), and AV speech perception.
• Testing occurred in a six-talker babble noise at 62 dB SPL.
• Multiple linear regression and a new parabolic model were used to predict AV performance from unimodal scores and age.

Main Results:

• Auditory-only and visual-only performance explained 87.9% of the variance in AV speech perception.
• Age did not significantly contribute to the prediction of AV speech perception.
• The new parabolic model explained 92.8% of AV performance variance, outperforming linear models and not including age as a significant factor.

Conclusions:

• AV speech perception is accurately predictable from unimodal performance, supporting previous findings.
• Age is not a significant predictor of AV speech perception accuracy in this open-set paradigm.
• The findings support a precision audiology approach, using unimodal and multimodal speech perception to guide rehabilitation for older adults.