Influence of social and semantic contexts in processing speech in noise

Abstract

Social interactions occupy a substantial part of our life. Not only interacting in first person, but also listening to others' interactions serve critical functions in understanding our social world. Yet, these auditory social signals are very often mixed with surrounding noise, such as in a restaurant, which degrades the information and forces our brain to use strategies to compensate for this loss. Although the role of semantics in speech comprehension under noisy conditions has been extensively studied as one of these strategies, the role of social context, and how it may interact with semantics, is unknown. Here, we conducted a series of four perceptual experiments to better understand the processing of multiple-speaker conversations from a third-person viewpoint, by manipulating the social and semantic contexts of a conversation. To do so, we used an assortment of artificial intelligence (AI) tools to generate a set of auditory stimuli, that consisted of two-speaker dialogues or one-speaker monologues (social context factor) arranged in intact or sentence-scrambled order (semantic context factor). Each stimulus comprised five sentences, with the fifth sentence embedded in multi-talker babbling noise. This fifth sentence was subsequently repeated without noise, with a single word altered or unchanged. Stimuli were presented over headphones to healthy young adult listeners, who were asked whether the repeated sentence was same as or different from the previous in-noise sentence. As results, we found overall that both the manipulation of social and semantic contexts had significant effects on performance in this task. We also found a correlation between a measure of autistic traits and individual performances when processing dialogues, but not monologues. Taken together, our results suggest that both semantic and social features of a conversation can modulate speech processing and that individual performance in this task could be related to certain social traits. These results raise new questions regarding the predictive or other mechanisms that may be at play when perceiving speech in social contexts.