Event-related brain activity in response to partners' speech in natural conversation.

Abstract

How listeners cognitively process speech during natural conversation remains poorly understood, particularly in terms of the role of the speaker's and listener's subjective mental states and empathic traits. This study examined relationships between these psychological factors and listener's cognitive processing of speech. We simultaneously recorded electroencephalograms from 20 romantic couples during natural face-to-face conversations. We identified the onset times of content words using morphological analysis of the speech data. Using multivariate temporal response functions (a method to estimate stimulus-related neural responses), we analyzed each listener's event-related brain activity in response to their partner's speech. We found positive associations between the speaker's interest levels and the listener's early attentional processing, as reflected in the P2 amplitude. Higher levels of personal distress, an empathic trait, corresponded with greater sustained attention among listeners, as indexed by the late positive potential. Moreover, by using brain activities along with behavioral measures of turn-taking, a support vector machine successfully distinguished between mutually satisfying and not mutually satisfying conversations. The observed associations between the cognitive processing of speech and both the speaker's mental states and the listener's empathic traits demonstrate that understanding speech processing in natural conversation requires consideration of factors from both participants not just from either the speaker or the listener.