Speech perception, especially in noise, may be maximized if the perceiver observes the naturally occurring visual-plus-auditory cues inherent in the production of spoken language. Evidence is conflicting, however, about which aspects of visual information mediate enhanced speech perception in noise. For this reason, we investigated the relative contributions of audibility and the type of visual cue in three experiments in young adults with normal hearing and vision. Relative to static visual cues, access to the talker's phonetic gestures in speech production, especially in noise, was associated with (a) faster response times and sensitivity for speech understanding in noise, and (b) shorter latencies and reduced amplitudes of auditory N1 eventrelated potentials. Dynamic chewing facial motion also decreased the N1 latency, but only meaningful linguistic motions reduced the N1 amplitude. The hypothesis that auditory-visual facilitation is distinct to properties of natural, dynamic speech gestures was partially supported.
Keywords Evoked potentials . Multisensory processing . Speech perceptionEveryone experiences difficulty understanding speech at some point, especially in noisy environments or in the presence of a hearing loss. The natural setting for speech communication is in a face-to-face environment where people can both hear the speech and see the face of the talker. Considerable evidence has shown that performance on speech perception tasks improves when information from both auditory and visual modalities is available