Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
The Effect of Voice Familiarity on Attention to Speech in a Cocktail Party Scenario
Poster D60 in Poster Session D, Wednesday, October 25, 4:45 - 6:30 pm CEST, Espace Vieux-Port
Paz Har-shai Yahav1, Aviya Sharaabi1, Elana Zion Golumbic1; 1Bar-Ilan University
Processing speech in multi-talker environments is challenging due to the competition for attentional resources. In such scenarios, top-down attention operates to selectively amplify one task-relevant voice (‘target speech’) and suppress irrelevant competing voices. Although attention-based prioritization is primarily task driven, it may also be affected by acoustic and semantic properties of the voices themselves. Here we focus on voice-familiarity, an ecologically important feature of human voices, and ask whether it affects the ability to attend-to or ignore speech in multi-talker situations. We tested two hypotheses: (1) is it easier to selectively-amplify and attend to a familiar voice vs. an unfamiliar one? (2) Do familiar voices that are supposedly outside the focus of attention serve as ‘attention grabbers’, leading to interference? To test this, we measured Magnetoencephalography (MEG) from N=33 volunteers in a dichotic listening paradigm. Participants were presented with two narratives, spoken by different voices, one to each ear. They were instructed to attend to the content from one ear (‘target speech’) and ignore the other (‘non-target speech’). Critically, participants were familiarized with one particular voice during the week prior to the MEG experiment, rendering this voice ‘familiar’ to them. The familiar voice was either designated as the ‘target speech’ or the ‘non-target speech’, allowing us to test the interaction between voice familiarity and attention/task-relevance on the neural response to speech. Using multivariate encoding TRF analysis we estimated the neural responses to target and non-target speech, and we applied source-localization tools to study how attention and voice-familiarity affected the speech tracking response across different brain-regions. We replicated the well-established effect of selective-attention, showing more robust speech tracking response for target speech vs. non-target speech. Interestingly, this attentional effect was modulated by voice familiarity such that neural tracking of both target speech and non-target speech was enhanced when presented in a familiar voice, in a moderately early time window. For target speech this effect was observed in frontal and posterior opercular cortex and insula, and right lateral-temporal and inferior-parietal cortex. For non-target speech this effect was limited to auditory and temporal regions in the hemisphere ipsi-lateral to the non-target speech (contra-lateral to the target-speech), followed by a familiar voice suppression in a later time window. These findings indicate that voice familiarity, and by extension, auditory-semantics, can interact with goal-driven selective attention, and affects both the ability to pay attention to target speech and to ignore non-target speech.
Topic Areas: Speech Perception,