| Summary: | Attending to a conversation in a crowded scene requires selection of relevant information, while ignoring other distracting sensory input, such as speech signals from surrounding people. The neural mechanisms of how distracting stimuli influence the processing of attended speech are not well understood. In this high-density electroencephalography (EEG) study, we investigated how different types of speech and non-speech stimuli influence the processing of attended audiovisual speech. Participants were presented with three horizontally aligned speakers who produced syllables. The faces of the three speakers flickered at specific frequencies (19 Hz for flanking speakers and 25 Hz for the center speaker), which induced steady-state visual evoked potentials (SSVEP) in the EEG that served as a measure of visual attention. The participant’s task was to detect an occasional audiovisual target syllable produced by the center speaker, while ignoring distracting signals originating from the two flanking speakers. In all experimental conditions the center speaker produced a bimodal audiovisual syllable. In three distraction conditions, which were contrasted with a no-distraction control condition, the flanking speakers produced either audiovisual speech, moved their lips and produced acoustic noise, or moved their lips without producing an auditory signal. We observed behavioral interference in particular when the flanking speakers produced naturalistic audiovisual speech. These effects were paralleled by enhanced 19 Hz SSVEP induced by the flanking speakers, indicative of a stimulus-driven capture of attention towards these speakers. The other distraction conditions did not reveal significant interference effects. Thus, our study provides evidence that non-relevant audiovisual speech signals serve as highly salient distractors, which capture attention in a stimulus-driven fashion.
|
|---|
