Early Career Research Award PresentationAudiovisual Speech Enhancement: Effects of Development and Hearing Loss OBJECTIVES: Visual speech can help compensate for acoustic input that is degraded by background noise or hearing loss. The increased accuracy and efficiency of speech processing in the presence of visual speech is called audiovisual speech enhancement. This talk will review recent findings from experiments examining whether there are age-related changes in children’s sensitivity to the visual cues that contribute to audiovisual speech enhancement, whether permanent pediatric hearing loss impacts audiovisual speech enhancement, and the extent to which speechreading acuity and acoustic-phonetic access govern individual differences in audiovisual speech enhancement. DESIGN: In a series of experiments, typically-developing children and hard-of-hearing children between 5 and 15 years of age completed auditory-only, visual-only, and audiovisual speech recognition tests. Across experiments, target stimuli included syllables, words, and sentences. Acoustic speech was degraded by background noise and/or filtering. Accuracy was measured under conditions of fixed acoustic degradation. Speech-in-noise recognition thresholds were obtained using adaptive methods. Consonant confusion error patterns and speech-feature transmission were also evaluated. Sample sizes ranged from 16 to 63 children. Small samples of young adults with normal hearing served as comparison groups in some experiments. RESULTS: Some key findings emerged. School-age children were less accurate than adults at speechreading consonant-vowel syllables. However, their speechreading error patterns demonstrated that children as young as 5 years of age have some knowledge of the visual features associated with speech sounds. Children’s speechreading of consonants with frontal place of articulation showed faster development than their speechreading of consonants with dorsal place of articulation. Comparison between auditory-only, visual-only, and audiovisual error patterns indicated that school-age children accessed and used their visual speech representations during audiovisual speech recognition. Across multiple measures, typically-developing children demonstrated smaller audiovisual speech enhancement than adults with normal hearing. However, age did not predict individual differences in degree of audiovisual speech enhancement among school-age children. Instead, acoustic-phonetic access (signal-to-noise ratio or auditory-only accuracy) was the strongest predictor of individual differences typically-developing school-age children’s audiovisual speech enhancement. Finally, although children who are hard of hearing had similar speechreading accuracy as typically-developing children, their audiovisual enhancement for speech recognition in noise was significantly larger than in typically-developing children. Information transmission analysis suggested that there may be less redundancy (and thus greater synergy) between the visual and auditory cues accessible to children who are hard of hearing than between the visual cues and auditory cues accessible to children with typical development. CONCLUSIONS: The magnitude of audiovisual enhancement of speech perception in noise is constant over the school-age years in children with typical development. Yet, children who are hard of hearing exhibit greater audiovisual enhancement of speech perception in noise than children who are typically developing. In typically-developing children, acoustic-phonetic access predicts individual differences in audiovisual speech enhancement. Ongoing research is examining the extent to which acoustic-phonetic access—as determined by configuration of hearing loss—governs audiovisual enhancement in children who are hard of hearing.
|
